{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.7098,
  "global_step": 17745,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 2e-06,
      "loss": 3.1659,
      "theoretical_loss": 3.484796891144814,
      "tokens_seen": 1661665280
    },
    {
      "epoch": 0.0,
      "learning_rate": 4e-06,
      "loss": 3.0486,
      "theoretical_loss": 3.484785010063602,
      "tokens_seen": 1661730816
    },
    {
      "epoch": 0.0,
      "learning_rate": 6e-06,
      "loss": 2.5634,
      "theoretical_loss": 3.4847731295821447,
      "tokens_seen": 1661796352
    },
    {
      "epoch": 0.0,
      "learning_rate": 8e-06,
      "loss": 3.0242,
      "theoretical_loss": 3.48476124970039,
      "tokens_seen": 1661861888
    },
    {
      "epoch": 0.0,
      "learning_rate": 1e-05,
      "loss": 2.685,
      "theoretical_loss": 3.4847493704182817,
      "tokens_seen": 1661927424
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.2e-05,
      "loss": 2.7646,
      "theoretical_loss": 3.4847374917357676,
      "tokens_seen": 1661992960
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.4e-05,
      "loss": 2.8002,
      "theoretical_loss": 3.4847256136527935,
      "tokens_seen": 1662058496
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.6e-05,
      "loss": 3.0059,
      "theoretical_loss": 3.4847137361693052,
      "tokens_seen": 1662124032
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.8e-05,
      "loss": 3.0361,
      "theoretical_loss": 3.4847018592852494,
      "tokens_seen": 1662189568
    },
    {
      "epoch": 0.0,
      "learning_rate": 2e-05,
      "loss": 2.79,
      "theoretical_loss": 3.484689983000571,
      "tokens_seen": 1662255104
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.2e-05,
      "loss": 2.7581,
      "theoretical_loss": 3.4846781073152173,
      "tokens_seen": 1662320640
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.4e-05,
      "loss": 2.6647,
      "theoretical_loss": 3.484666232229134,
      "tokens_seen": 1662386176
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.6e-05,
      "loss": 2.8627,
      "theoretical_loss": 3.484654357742267,
      "tokens_seen": 1662451712
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.8e-05,
      "loss": 2.8585,
      "theoretical_loss": 3.4846424838545627,
      "tokens_seen": 1662517248
    },
    {
      "epoch": 0.0,
      "learning_rate": 3e-05,
      "loss": 2.7554,
      "theoretical_loss": 3.4846306105659677,
      "tokens_seen": 1662582784
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.2e-05,
      "loss": 2.5777,
      "theoretical_loss": 3.484618737876427,
      "tokens_seen": 1662648320
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.4000000000000007e-05,
      "loss": 2.5762,
      "theoretical_loss": 3.4846068657858877,
      "tokens_seen": 1662713856
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.6e-05,
      "loss": 2.6419,
      "theoretical_loss": 3.484594994294296,
      "tokens_seen": 1662779392
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.8e-05,
      "loss": 2.6613,
      "theoretical_loss": 3.4845831234015976,
      "tokens_seen": 1662844928
    },
    {
      "epoch": 0.0,
      "learning_rate": 4e-05,
      "loss": 2.6489,
      "theoretical_loss": 3.484571253107739,
      "tokens_seen": 1662910464
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.2000000000000004e-05,
      "loss": 2.8326,
      "theoretical_loss": 3.484559383412666,
      "tokens_seen": 1662976000
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.4e-05,
      "loss": 2.771,
      "theoretical_loss": 3.4845475143163256,
      "tokens_seen": 1663041536
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.6e-05,
      "loss": 2.8005,
      "theoretical_loss": 3.4845356458186627,
      "tokens_seen": 1663107072
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.8e-05,
      "loss": 2.6649,
      "theoretical_loss": 3.484523777919625,
      "tokens_seen": 1663172608
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 918485,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4342758655548096,
      "objective/train/theoretical_loss": 3.4845119106191578,
      "objective/train/tokens_used": 22097376,
      "theoretical_loss": 3.4845119106191578,
      "tokens_seen": 1663238144
    },
    {
      "epoch": 0.0,
      "learning_rate": 5e-05,
      "loss": 2.6636,
      "theoretical_loss": 3.4845119106191578,
      "tokens_seen": 1663238144
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.2e-05,
      "loss": 2.8955,
      "theoretical_loss": 3.4845000439172074,
      "tokens_seen": 1663303680
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.4e-05,
      "loss": 2.5046,
      "theoretical_loss": 3.48448817781372,
      "tokens_seen": 1663369216
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.6e-05,
      "loss": 2.709,
      "theoretical_loss": 3.484476312308643,
      "tokens_seen": 1663434752
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.800000000000001e-05,
      "loss": 2.4005,
      "theoretical_loss": 3.484464447401921,
      "tokens_seen": 1663500288
    },
    {
      "epoch": 0.0,
      "learning_rate": 6e-05,
      "loss": 2.5362,
      "theoretical_loss": 3.4844525830935007,
      "tokens_seen": 1663565824
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.2e-05,
      "loss": 2.5606,
      "theoretical_loss": 3.4844407193833282,
      "tokens_seen": 1663631360
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.4e-05,
      "loss": 2.6669,
      "theoretical_loss": 3.484428856271351,
      "tokens_seen": 1663696896
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.6e-05,
      "loss": 2.626,
      "theoretical_loss": 3.484416993757514,
      "tokens_seen": 1663762432
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.800000000000001e-05,
      "loss": 2.5635,
      "theoretical_loss": 3.484405131841765,
      "tokens_seen": 1663827968
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.000000000000001e-05,
      "loss": 2.454,
      "theoretical_loss": 3.4843932705240483,
      "tokens_seen": 1663893504
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.2e-05,
      "loss": 2.5559,
      "theoretical_loss": 3.4843814098043113,
      "tokens_seen": 1663959040
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.4e-05,
      "loss": 2.1608,
      "theoretical_loss": 3.4843695496825005,
      "tokens_seen": 1664024576
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.6e-05,
      "loss": 2.6816,
      "theoretical_loss": 3.4843576901585616,
      "tokens_seen": 1664090112
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.8e-05,
      "loss": 2.5955,
      "theoretical_loss": 3.4843458312324413,
      "tokens_seen": 1664155648
    },
    {
      "epoch": 0.0,
      "learning_rate": 8e-05,
      "loss": 2.7015,
      "theoretical_loss": 3.484333972904086,
      "tokens_seen": 1664221184
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.2e-05,
      "loss": 2.5895,
      "theoretical_loss": 3.484322115173442,
      "tokens_seen": 1664286720
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.400000000000001e-05,
      "loss": 2.6653,
      "theoretical_loss": 3.4843102580404555,
      "tokens_seen": 1664352256
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.599999999999999e-05,
      "loss": 2.4374,
      "theoretical_loss": 3.4842984015050726,
      "tokens_seen": 1664417792
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.8e-05,
      "loss": 2.6655,
      "theoretical_loss": 3.4842865455672403,
      "tokens_seen": 1664483328
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.999999999999999e-05,
      "loss": 2.1725,
      "theoretical_loss": 3.484274690226904,
      "tokens_seen": 1664548864
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.2e-05,
      "loss": 2.6074,
      "theoretical_loss": 3.484262835484011,
      "tokens_seen": 1664614400
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.400000000000001e-05,
      "loss": 2.2801,
      "theoretical_loss": 3.484250981338507,
      "tokens_seen": 1664679936
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.6e-05,
      "loss": 2.6196,
      "theoretical_loss": 3.484239127790339,
      "tokens_seen": 1664745472
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.800000000000001e-05,
      "loss": 2.349,
      "theoretical_loss": 3.484227274839453,
      "tokens_seen": 1664811008
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 919051,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5782582759857178,
      "objective/train/theoretical_loss": 3.4842154224857955,
      "objective/train/tokens_used": 23735776,
      "theoretical_loss": 3.4842154224857955,
      "tokens_seen": 1664876544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001,
      "loss": 2.5304,
      "theoretical_loss": 3.4842154224857955,
      "tokens_seen": 1664876544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000102,
      "loss": 2.5036,
      "theoretical_loss": 3.484203570729313,
      "tokens_seen": 1664942080
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000104,
      "loss": 2.4537,
      "theoretical_loss": 3.4841917195699517,
      "tokens_seen": 1665007616
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000106,
      "loss": 2.2072,
      "theoretical_loss": 3.484179869007658,
      "tokens_seen": 1665073152
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000108,
      "loss": 2.5087,
      "theoretical_loss": 3.4841680190423787,
      "tokens_seen": 1665138688
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011,
      "loss": 2.2714,
      "theoretical_loss": 3.4841561696740597,
      "tokens_seen": 1665204224
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000112,
      "loss": 2.4838,
      "theoretical_loss": 3.4841443209026477,
      "tokens_seen": 1665269760
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000114,
      "loss": 2.6179,
      "theoretical_loss": 3.4841324727280893,
      "tokens_seen": 1665335296
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011600000000000001,
      "loss": 2.5005,
      "theoretical_loss": 3.4841206251503305,
      "tokens_seen": 1665400832
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000118,
      "loss": 2.562,
      "theoretical_loss": 3.484108778169318,
      "tokens_seen": 1665466368
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012,
      "loss": 2.5914,
      "theoretical_loss": 3.4840969317849986,
      "tokens_seen": 1665531904
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000122,
      "loss": 2.3786,
      "theoretical_loss": 3.484085085997318,
      "tokens_seen": 1665597440
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000124,
      "loss": 2.3953,
      "theoretical_loss": 3.4840732408062234,
      "tokens_seen": 1665662976
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000126,
      "loss": 2.5667,
      "theoretical_loss": 3.484061396211661,
      "tokens_seen": 1665728512
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000128,
      "loss": 2.5867,
      "theoretical_loss": 3.4840495522135777,
      "tokens_seen": 1665794048
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013000000000000002,
      "loss": 2.4648,
      "theoretical_loss": 3.4840377088119188,
      "tokens_seen": 1665859584
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000132,
      "loss": 2.3746,
      "theoretical_loss": 3.484025866006632,
      "tokens_seen": 1665925120
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000134,
      "loss": 2.4735,
      "theoretical_loss": 3.4840140237976636,
      "tokens_seen": 1665990656
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013600000000000003,
      "loss": 2.3041,
      "theoretical_loss": 3.48400218218496,
      "tokens_seen": 1666056192
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013800000000000002,
      "loss": 2.5027,
      "theoretical_loss": 3.483990341168467,
      "tokens_seen": 1666121728
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014000000000000001,
      "loss": 2.3187,
      "theoretical_loss": 3.4839785007481323,
      "tokens_seen": 1666187264
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014199999999999998,
      "loss": 2.6289,
      "theoretical_loss": 3.483966660923902,
      "tokens_seen": 1666252800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000144,
      "loss": 2.5713,
      "theoretical_loss": 3.483954821695723,
      "tokens_seen": 1666318336
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000146,
      "loss": 2.5538,
      "theoretical_loss": 3.4839429830635407,
      "tokens_seen": 1666383872
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000148,
      "loss": 2.4974,
      "theoretical_loss": 3.4839311450273023,
      "tokens_seen": 1666449408
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 920018,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4347381591796875,
      "objective/train/theoretical_loss": 3.4839193075869543,
      "objective/train/tokens_used": 25374176,
      "theoretical_loss": 3.4839193075869543,
      "tokens_seen": 1666514944
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015,
      "loss": 2.4437,
      "theoretical_loss": 3.4839193075869543,
      "tokens_seen": 1666514944
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000152,
      "loss": 2.3852,
      "theoretical_loss": 3.4839074707424444,
      "tokens_seen": 1666580480
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000154,
      "loss": 2.5793,
      "theoretical_loss": 3.4838956344937175,
      "tokens_seen": 1666646016
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000156,
      "loss": 2.4539,
      "theoretical_loss": 3.483883798840721,
      "tokens_seen": 1666711552
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000158,
      "loss": 2.466,
      "theoretical_loss": 3.4838719637834012,
      "tokens_seen": 1666777088
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016,
      "loss": 2.4158,
      "theoretical_loss": 3.4838601293217053,
      "tokens_seen": 1666842624
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000162,
      "loss": 2.3978,
      "theoretical_loss": 3.483848295455579,
      "tokens_seen": 1666908160
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000164,
      "loss": 2.3159,
      "theoretical_loss": 3.4838364621849705,
      "tokens_seen": 1666973696
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016600000000000002,
      "loss": 2.5254,
      "theoretical_loss": 3.483824629509824,
      "tokens_seen": 1667039232
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016800000000000002,
      "loss": 2.2786,
      "theoretical_loss": 3.483812797430088,
      "tokens_seen": 1667104768
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017,
      "loss": 2.4052,
      "theoretical_loss": 3.4838009659457088,
      "tokens_seen": 1667170304
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017199999999999998,
      "loss": 2.5102,
      "theoretical_loss": 3.4837891350566323,
      "tokens_seen": 1667235840
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000174,
      "loss": 2.4416,
      "theoretical_loss": 3.4837773047628064,
      "tokens_seen": 1667301376
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000176,
      "loss": 2.2819,
      "theoretical_loss": 3.4837654750641764,
      "tokens_seen": 1667366912
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000178,
      "loss": 2.4608,
      "theoretical_loss": 3.4837536459606904,
      "tokens_seen": 1667432448
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017999999999999998,
      "loss": 2.367,
      "theoretical_loss": 3.4837418174522936,
      "tokens_seen": 1667497984
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000182,
      "loss": 2.4133,
      "theoretical_loss": 3.4837299895389333,
      "tokens_seen": 1667563520
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000184,
      "loss": 2.515,
      "theoretical_loss": 3.4837181622205566,
      "tokens_seen": 1667629056
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000186,
      "loss": 2.4365,
      "theoretical_loss": 3.48370633549711,
      "tokens_seen": 1667694592
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018800000000000002,
      "loss": 2.4476,
      "theoretical_loss": 3.483694509368539,
      "tokens_seen": 1667760128
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019,
      "loss": 2.6584,
      "theoretical_loss": 3.4836826838347923,
      "tokens_seen": 1667825664
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000192,
      "loss": 2.2917,
      "theoretical_loss": 3.483670858895815,
      "tokens_seen": 1667891200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000194,
      "loss": 2.5954,
      "theoretical_loss": 3.483659034551555,
      "tokens_seen": 1667956736
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019600000000000002,
      "loss": 2.4191,
      "theoretical_loss": 3.4836472108019585,
      "tokens_seen": 1668022272
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019800000000000002,
      "loss": 2.4538,
      "theoretical_loss": 3.483635387646972,
      "tokens_seen": 1668087808
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 921237,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6976654529571533,
      "objective/train/theoretical_loss": 3.483623565086542,
      "objective/train/tokens_used": 27012576,
      "theoretical_loss": 3.483623565086542,
      "tokens_seen": 1668153344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 2.6155,
      "theoretical_loss": 3.483623565086542,
      "tokens_seen": 1668153344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000202,
      "loss": 2.5577,
      "theoretical_loss": 3.4836117431206164,
      "tokens_seen": 1668218880
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000204,
      "loss": 2.397,
      "theoretical_loss": 3.483599921749141,
      "tokens_seen": 1668284416
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000206,
      "loss": 2.5303,
      "theoretical_loss": 3.483588100972063,
      "tokens_seen": 1668349952
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000208,
      "loss": 2.3655,
      "theoretical_loss": 3.483576280789328,
      "tokens_seen": 1668415488
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021,
      "loss": 2.3679,
      "theoretical_loss": 3.483564461200885,
      "tokens_seen": 1668481024
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000212,
      "loss": 2.3781,
      "theoretical_loss": 3.4835526422066785,
      "tokens_seen": 1668546560
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000214,
      "loss": 2.444,
      "theoretical_loss": 3.4835408238066567,
      "tokens_seen": 1668612096
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000216,
      "loss": 2.3319,
      "theoretical_loss": 3.483529006000766,
      "tokens_seen": 1668677632
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000218,
      "loss": 2.2344,
      "theoretical_loss": 3.4835171887889533,
      "tokens_seen": 1668743168
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022,
      "loss": 2.4793,
      "theoretical_loss": 3.483505372171165,
      "tokens_seen": 1668808704
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000222,
      "loss": 2.4368,
      "theoretical_loss": 3.483493556147349,
      "tokens_seen": 1668874240
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000224,
      "loss": 2.4653,
      "theoretical_loss": 3.4834817407174503,
      "tokens_seen": 1668939776
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022600000000000002,
      "loss": 2.2769,
      "theoretical_loss": 3.4834699258814172,
      "tokens_seen": 1669005312
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000228,
      "loss": 2.5012,
      "theoretical_loss": 3.4834581116391963,
      "tokens_seen": 1669070848
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023,
      "loss": 2.7566,
      "theoretical_loss": 3.483446297990734,
      "tokens_seen": 1669136384
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023200000000000003,
      "loss": 2.427,
      "theoretical_loss": 3.4834344849359775,
      "tokens_seen": 1669201920
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023400000000000002,
      "loss": 2.5868,
      "theoretical_loss": 3.4834226724748736,
      "tokens_seen": 1669267456
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000236,
      "loss": 2.5531,
      "theoretical_loss": 3.483410860607369,
      "tokens_seen": 1669332992
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023799999999999998,
      "loss": 2.5605,
      "theoretical_loss": 3.4833990493334106,
      "tokens_seen": 1669398528
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024,
      "loss": 2.2757,
      "theoretical_loss": 3.4833872386529454,
      "tokens_seen": 1669464064
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000242,
      "loss": 2.5149,
      "theoretical_loss": 3.48337542856592,
      "tokens_seen": 1669529600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000244,
      "loss": 2.2792,
      "theoretical_loss": 3.483363619072282,
      "tokens_seen": 1669595136
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000246,
      "loss": 2.4745,
      "theoretical_loss": 3.4833518101719774,
      "tokens_seen": 1669660672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000248,
      "loss": 2.37,
      "theoretical_loss": 3.4833400018649536,
      "tokens_seen": 1669726208
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 921959,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.675856590270996,
      "objective/train/theoretical_loss": 3.4833281941511576,
      "objective/train/tokens_used": 28650976,
      "theoretical_loss": 3.4833281941511576,
      "tokens_seen": 1669791744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025,
      "loss": 2.8538,
      "theoretical_loss": 3.4833281941511576,
      "tokens_seen": 1669791744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000252,
      "loss": 2.3873,
      "theoretical_loss": 3.483316387030536,
      "tokens_seen": 1669857280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000254,
      "loss": 2.3762,
      "theoretical_loss": 3.4833045805030363,
      "tokens_seen": 1669922816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000256,
      "loss": 2.5253,
      "theoretical_loss": 3.4832927745686044,
      "tokens_seen": 1669988352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025800000000000004,
      "loss": 2.489,
      "theoretical_loss": 3.483280969227188,
      "tokens_seen": 1670053888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026000000000000003,
      "loss": 2.3014,
      "theoretical_loss": 3.4832691644787337,
      "tokens_seen": 1670119424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000262,
      "loss": 2.2745,
      "theoretical_loss": 3.4832573603231887,
      "tokens_seen": 1670184960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000264,
      "loss": 2.6538,
      "theoretical_loss": 3.4832455567605,
      "tokens_seen": 1670250496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000266,
      "loss": 2.3405,
      "theoretical_loss": 3.4832337537906146,
      "tokens_seen": 1670316032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000268,
      "loss": 2.4211,
      "theoretical_loss": 3.483221951413479,
      "tokens_seen": 1670381568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027,
      "loss": 2.5642,
      "theoretical_loss": 3.4832101496290404,
      "tokens_seen": 1670447104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027200000000000005,
      "loss": 2.3653,
      "theoretical_loss": 3.483198348437246,
      "tokens_seen": 1670512640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027400000000000005,
      "loss": 2.4734,
      "theoretical_loss": 3.4831865478380433,
      "tokens_seen": 1670578176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027600000000000004,
      "loss": 2.41,
      "theoretical_loss": 3.483174747831378,
      "tokens_seen": 1670643712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027800000000000004,
      "loss": 2.2828,
      "theoretical_loss": 3.4831629484171978,
      "tokens_seen": 1670709248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028000000000000003,
      "loss": 2.5743,
      "theoretical_loss": 3.4831511495954497,
      "tokens_seen": 1670774784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028199999999999997,
      "loss": 2.4478,
      "theoretical_loss": 3.4831393513660807,
      "tokens_seen": 1670840320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028399999999999996,
      "loss": 2.5524,
      "theoretical_loss": 3.483127553729038,
      "tokens_seen": 1670905856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028599999999999996,
      "loss": 2.6412,
      "theoretical_loss": 3.483115756684268,
      "tokens_seen": 1670971392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000288,
      "loss": 2.3564,
      "theoretical_loss": 3.483103960231719,
      "tokens_seen": 1671036928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029,
      "loss": 2.5918,
      "theoretical_loss": 3.4830921643713366,
      "tokens_seen": 1671102464
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000292,
      "loss": 2.2021,
      "theoretical_loss": 3.4830803691030683,
      "tokens_seen": 1671168000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000294,
      "loss": 2.3097,
      "theoretical_loss": 3.483068574426862,
      "tokens_seen": 1671233536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000296,
      "loss": 2.4126,
      "theoretical_loss": 3.4830567803426637,
      "tokens_seen": 1671299072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000298,
      "loss": 2.4376,
      "theoretical_loss": 3.483044986850421,
      "tokens_seen": 1671364608
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 922598,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7050137519836426,
      "objective/train/theoretical_loss": 3.4830331939500807,
      "objective/train/tokens_used": 30289376,
      "theoretical_loss": 3.4830331939500807,
      "tokens_seen": 1671430144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003,
      "loss": 2.5272,
      "theoretical_loss": 3.4830331939500807,
      "tokens_seen": 1671430144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000302,
      "loss": 2.4159,
      "theoretical_loss": 3.48302140164159,
      "tokens_seen": 1671495680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000304,
      "loss": 2.3413,
      "theoretical_loss": 3.483009609924897,
      "tokens_seen": 1671561216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000306,
      "loss": 2.7255,
      "theoretical_loss": 3.482997818799947,
      "tokens_seen": 1671626752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000308,
      "loss": 2.4114,
      "theoretical_loss": 3.482986028266688,
      "tokens_seen": 1671692288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031,
      "loss": 2.3985,
      "theoretical_loss": 3.4829742383250673,
      "tokens_seen": 1671757824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000312,
      "loss": 2.4447,
      "theoretical_loss": 3.4829624489750315,
      "tokens_seen": 1671823360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000314,
      "loss": 2.1292,
      "theoretical_loss": 3.4829506602165283,
      "tokens_seen": 1671888896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000316,
      "loss": 2.4829,
      "theoretical_loss": 3.4829388720495045,
      "tokens_seen": 1671954432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031800000000000003,
      "loss": 2.3397,
      "theoretical_loss": 3.482927084473907,
      "tokens_seen": 1672019968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032,
      "loss": 2.5797,
      "theoretical_loss": 3.4829152974896838,
      "tokens_seen": 1672085504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000322,
      "loss": 2.6122,
      "theoretical_loss": 3.482903511096781,
      "tokens_seen": 1672151040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000324,
      "loss": 2.3373,
      "theoretical_loss": 3.482891725295146,
      "tokens_seen": 1672216576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000326,
      "loss": 2.2181,
      "theoretical_loss": 3.482879940084727,
      "tokens_seen": 1672282112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000328,
      "loss": 2.5586,
      "theoretical_loss": 3.48286815546547,
      "tokens_seen": 1672347648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033,
      "loss": 2.4485,
      "theoretical_loss": 3.4828563714373226,
      "tokens_seen": 1672413184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033200000000000005,
      "loss": 2.6228,
      "theoretical_loss": 3.482844588000232,
      "tokens_seen": 1672478720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033400000000000004,
      "loss": 2.2578,
      "theoretical_loss": 3.4828328051541453,
      "tokens_seen": 1672544256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033600000000000004,
      "loss": 2.5647,
      "theoretical_loss": 3.48282102289901,
      "tokens_seen": 1672609792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033800000000000003,
      "loss": 2.27,
      "theoretical_loss": 3.482809241234773,
      "tokens_seen": 1672675328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034,
      "loss": 2.8122,
      "theoretical_loss": 3.4827974601613816,
      "tokens_seen": 1672740864
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000342,
      "loss": 2.6542,
      "theoretical_loss": 3.482785679678783,
      "tokens_seen": 1672806400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034399999999999996,
      "loss": 2.4875,
      "theoretical_loss": 3.482773899786924,
      "tokens_seen": 1672871936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000346,
      "loss": 2.5134,
      "theoretical_loss": 3.482762120485753,
      "tokens_seen": 1672937472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000348,
      "loss": 2.5356,
      "theoretical_loss": 3.4827503417752164,
      "tokens_seen": 1673003008
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 923228,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5057578086853027,
      "objective/train/theoretical_loss": 3.482738563655261,
      "objective/train/tokens_used": 31927776,
      "theoretical_loss": 3.482738563655261,
      "tokens_seen": 1673068544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035,
      "loss": 2.7017,
      "theoretical_loss": 3.482738563655261,
      "tokens_seen": 1673068544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000352,
      "loss": 2.3692,
      "theoretical_loss": 3.482726786125835,
      "tokens_seen": 1673134080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000354,
      "loss": 2.4674,
      "theoretical_loss": 3.4827150091868853,
      "tokens_seen": 1673199616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000356,
      "loss": 2.3015,
      "theoretical_loss": 3.482703232838359,
      "tokens_seen": 1673265152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000358,
      "loss": 2.4719,
      "theoretical_loss": 3.482691457080204,
      "tokens_seen": 1673330688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035999999999999997,
      "loss": 2.3731,
      "theoretical_loss": 3.4826796819123667,
      "tokens_seen": 1673396224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000362,
      "loss": 2.3984,
      "theoretical_loss": 3.482667907334795,
      "tokens_seen": 1673461760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000364,
      "loss": 2.5499,
      "theoretical_loss": 3.4826561333474357,
      "tokens_seen": 1673527296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000366,
      "loss": 2.4005,
      "theoretical_loss": 3.482644359950237,
      "tokens_seen": 1673592832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000368,
      "loss": 2.5822,
      "theoretical_loss": 3.482632587143145,
      "tokens_seen": 1673658368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037,
      "loss": 2.5947,
      "theoretical_loss": 3.4826208149261078,
      "tokens_seen": 1673723904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000372,
      "loss": 2.4242,
      "theoretical_loss": 3.482609043299073,
      "tokens_seen": 1673789440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000374,
      "loss": 2.5893,
      "theoretical_loss": 3.482597272261987,
      "tokens_seen": 1673854976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037600000000000003,
      "loss": 2.7811,
      "theoretical_loss": 3.482585501814798,
      "tokens_seen": 1673920512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000378,
      "loss": 2.6416,
      "theoretical_loss": 3.4825737319574523,
      "tokens_seen": 1673986048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038,
      "loss": 2.3689,
      "theoretical_loss": 3.482561962689899,
      "tokens_seen": 1674051584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000382,
      "loss": 2.4532,
      "theoretical_loss": 3.4825501940120835,
      "tokens_seen": 1674117120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000384,
      "loss": 2.2893,
      "theoretical_loss": 3.4825384259239547,
      "tokens_seen": 1674182656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000386,
      "loss": 2.4148,
      "theoretical_loss": 3.4825266584254586,
      "tokens_seen": 1674248192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000388,
      "loss": 2.4807,
      "theoretical_loss": 3.482514891516544,
      "tokens_seen": 1674313728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039000000000000005,
      "loss": 2.5216,
      "theoretical_loss": 3.482503125197157,
      "tokens_seen": 1674379264
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039200000000000004,
      "loss": 2.2569,
      "theoretical_loss": 3.4824913594672458,
      "tokens_seen": 1674444800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039400000000000004,
      "loss": 2.2831,
      "theoretical_loss": 3.4824795943267577,
      "tokens_seen": 1674510336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039600000000000003,
      "loss": 2.5188,
      "theoretical_loss": 3.48246782977564,
      "tokens_seen": 1674575872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000398,
      "loss": 2.576,
      "theoretical_loss": 3.4824560658138397,
      "tokens_seen": 1674641408
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 924425,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.702085256576538,
      "objective/train/theoretical_loss": 3.482444302441305,
      "objective/train/tokens_used": 33566176,
      "theoretical_loss": 3.482444302441305,
      "tokens_seen": 1674706944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004,
      "loss": 2.6435,
      "theoretical_loss": 3.482444302441305,
      "tokens_seen": 1674706944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000402,
      "loss": 2.4602,
      "theoretical_loss": 3.482432539657983,
      "tokens_seen": 1674772480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000404,
      "loss": 2.3951,
      "theoretical_loss": 3.4824207774638207,
      "tokens_seen": 1674838016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040600000000000006,
      "loss": 2.4707,
      "theoretical_loss": 3.4824090158587664,
      "tokens_seen": 1674903552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000408,
      "loss": 2.4597,
      "theoretical_loss": 3.482397254842767,
      "tokens_seen": 1674969088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041,
      "loss": 2.5413,
      "theoretical_loss": 3.4823854944157695,
      "tokens_seen": 1675034624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000412,
      "loss": 2.4902,
      "theoretical_loss": 3.482373734577722,
      "tokens_seen": 1675100160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000414,
      "loss": 2.3291,
      "theoretical_loss": 3.482361975328572,
      "tokens_seen": 1675165696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000416,
      "loss": 2.4068,
      "theoretical_loss": 3.4823502166682667,
      "tokens_seen": 1675231232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041799999999999997,
      "loss": 2.4366,
      "theoretical_loss": 3.482338458596754,
      "tokens_seen": 1675296768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042,
      "loss": 2.5455,
      "theoretical_loss": 3.4823267011139807,
      "tokens_seen": 1675362304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000422,
      "loss": 2.6574,
      "theoretical_loss": 3.482314944219895,
      "tokens_seen": 1675427840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000424,
      "loss": 2.3464,
      "theoretical_loss": 3.482303187914444,
      "tokens_seen": 1675493376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000426,
      "loss": 2.5067,
      "theoretical_loss": 3.482291432197575,
      "tokens_seen": 1675558912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000428,
      "loss": 2.4157,
      "theoretical_loss": 3.4822796770692364,
      "tokens_seen": 1675624448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043,
      "loss": 2.5832,
      "theoretical_loss": 3.482267922529374,
      "tokens_seen": 1675689984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000432,
      "loss": 2.8137,
      "theoretical_loss": 3.4822561685779374,
      "tokens_seen": 1675755520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043400000000000003,
      "loss": 2.3953,
      "theoretical_loss": 3.482244415214873,
      "tokens_seen": 1675821056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000436,
      "loss": 2.6011,
      "theoretical_loss": 3.482232662440128,
      "tokens_seen": 1675886592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000438,
      "loss": 2.2128,
      "theoretical_loss": 3.482220910253651,
      "tokens_seen": 1675952128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044,
      "loss": 2.4778,
      "theoretical_loss": 3.4822091586553885,
      "tokens_seen": 1676017664
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000442,
      "loss": 2.2702,
      "theoretical_loss": 3.482197407645289,
      "tokens_seen": 1676083200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000444,
      "loss": 2.5679,
      "theoretical_loss": 3.482185657223299,
      "tokens_seen": 1676148736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000446,
      "loss": 2.4653,
      "theoretical_loss": 3.4821739073893676,
      "tokens_seen": 1676214272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000448,
      "loss": 2.27,
      "theoretical_loss": 3.482162158143441,
      "tokens_seen": 1676279808
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 924948,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.733900785446167,
      "objective/train/theoretical_loss": 3.482150409485467,
      "objective/train/tokens_used": 35204576,
      "theoretical_loss": 3.482150409485467,
      "tokens_seen": 1676345344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045000000000000004,
      "loss": 2.5402,
      "theoretical_loss": 3.482150409485467,
      "tokens_seen": 1676345344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045200000000000004,
      "loss": 2.5286,
      "theoretical_loss": 3.4821386614153944,
      "tokens_seen": 1676410880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045400000000000003,
      "loss": 2.6377,
      "theoretical_loss": 3.482126913933169,
      "tokens_seen": 1676476416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000456,
      "loss": 2.5819,
      "theoretical_loss": 3.4821151670387396,
      "tokens_seen": 1676541952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000458,
      "loss": 2.5565,
      "theoretical_loss": 3.482103420732053,
      "tokens_seen": 1676607488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046,
      "loss": 2.5833,
      "theoretical_loss": 3.482091675013058,
      "tokens_seen": 1676673024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000462,
      "loss": 2.6621,
      "theoretical_loss": 3.482079929881701,
      "tokens_seen": 1676738560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046400000000000006,
      "loss": 2.3564,
      "theoretical_loss": 3.4820681853379307,
      "tokens_seen": 1676804096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046600000000000005,
      "loss": 2.4111,
      "theoretical_loss": 3.482056441381694,
      "tokens_seen": 1676869632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046800000000000005,
      "loss": 2.5978,
      "theoretical_loss": 3.4820446980129383,
      "tokens_seen": 1676935168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047,
      "loss": 2.5111,
      "theoretical_loss": 3.4820329552316123,
      "tokens_seen": 1677000704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000472,
      "loss": 2.5309,
      "theoretical_loss": 3.4820212130376627,
      "tokens_seen": 1677066240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000474,
      "loss": 2.4254,
      "theoretical_loss": 3.482009471431038,
      "tokens_seen": 1677131776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047599999999999997,
      "loss": 2.3422,
      "theoretical_loss": 3.4819977304116847,
      "tokens_seen": 1677197312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047799999999999996,
      "loss": 2.5434,
      "theoretical_loss": 3.4819859899795516,
      "tokens_seen": 1677262848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048,
      "loss": 2.5123,
      "theoretical_loss": 3.481974250134586,
      "tokens_seen": 1677328384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000482,
      "loss": 2.6815,
      "theoretical_loss": 3.481962510876736,
      "tokens_seen": 1677393920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000484,
      "loss": 2.3318,
      "theoretical_loss": 3.481950772205948,
      "tokens_seen": 1677459456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000486,
      "loss": 2.4658,
      "theoretical_loss": 3.481939034122171,
      "tokens_seen": 1677524992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000488,
      "loss": 2.3931,
      "theoretical_loss": 3.4819272966253525,
      "tokens_seen": 1677590528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049,
      "loss": 2.5888,
      "theoretical_loss": 3.4819155597154396,
      "tokens_seen": 1677656064
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000492,
      "loss": 2.5776,
      "theoretical_loss": 3.481903823392381,
      "tokens_seen": 1677721600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000494,
      "loss": 2.3611,
      "theoretical_loss": 3.4818920876561235,
      "tokens_seen": 1677787136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000496,
      "loss": 2.501,
      "theoretical_loss": 3.4818803525066153,
      "tokens_seen": 1677852672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498,
      "loss": 2.6304,
      "theoretical_loss": 3.481868617943804,
      "tokens_seen": 1677918208
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 926151,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.59950590133667,
      "objective/train/theoretical_loss": 3.4818568839676374,
      "objective/train/tokens_used": 36842976,
      "theoretical_loss": 3.4818568839676374,
      "tokens_seen": 1677983744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005,
      "loss": 2.582,
      "theoretical_loss": 3.4818568839676374,
      "tokens_seen": 1677983744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499979797979798,
      "loss": 2.3636,
      "theoretical_loss": 3.481845150578063,
      "tokens_seen": 1678049280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999595959595959,
      "loss": 2.4696,
      "theoretical_loss": 3.4818334177750296,
      "tokens_seen": 1678114816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999393939393939,
      "loss": 2.5003,
      "theoretical_loss": 3.481821685558484,
      "tokens_seen": 1678180352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499919191919192,
      "loss": 2.6498,
      "theoretical_loss": 3.481809953928374,
      "tokens_seen": 1678245888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049989898989899,
      "loss": 2.5553,
      "theoretical_loss": 3.4817982228846476,
      "tokens_seen": 1678311424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998787878787879,
      "loss": 2.5759,
      "theoretical_loss": 3.4817864924272524,
      "tokens_seen": 1678376960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998585858585859,
      "loss": 2.7333,
      "theoretical_loss": 3.481774762556137,
      "tokens_seen": 1678442496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998383838383839,
      "loss": 2.6417,
      "theoretical_loss": 3.4817630332712484,
      "tokens_seen": 1678508032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998181818181818,
      "loss": 2.8269,
      "theoretical_loss": 3.4817513045725343,
      "tokens_seen": 1678573568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997979797979798,
      "loss": 2.4396,
      "theoretical_loss": 3.481739576459943,
      "tokens_seen": 1678639104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997777777777778,
      "loss": 2.6778,
      "theoretical_loss": 3.4817278489334225,
      "tokens_seen": 1678704640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997575757575757,
      "loss": 2.4876,
      "theoretical_loss": 3.48171612199292,
      "tokens_seen": 1678770176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997373737373737,
      "loss": 2.5085,
      "theoretical_loss": 3.481704395638383,
      "tokens_seen": 1678835712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997171717171717,
      "loss": 2.523,
      "theoretical_loss": 3.4816926698697612,
      "tokens_seen": 1678901248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996969696969697,
      "loss": 2.5983,
      "theoretical_loss": 3.4816809446870005,
      "tokens_seen": 1678966784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996767676767676,
      "loss": 2.5985,
      "theoretical_loss": 3.48166922009005,
      "tokens_seen": 1679032320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996565656565657,
      "loss": 2.6887,
      "theoretical_loss": 3.4816574960788564,
      "tokens_seen": 1679097856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996363636363637,
      "loss": 2.6201,
      "theoretical_loss": 3.481645772653369,
      "tokens_seen": 1679163392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996161616161617,
      "loss": 2.6373,
      "theoretical_loss": 3.4816340498135343,
      "tokens_seen": 1679228928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995959595959596,
      "loss": 2.5179,
      "theoretical_loss": 3.4816223275593012,
      "tokens_seen": 1679294464
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995757575757576,
      "loss": 2.514,
      "theoretical_loss": 3.4816106058906175,
      "tokens_seen": 1679360000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995555555555555,
      "loss": 2.5432,
      "theoretical_loss": 3.48159888480743,
      "tokens_seen": 1679425536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995353535353536,
      "loss": 2.5299,
      "theoretical_loss": 3.481587164309688,
      "tokens_seen": 1679491072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995151515151515,
      "loss": 2.3886,
      "theoretical_loss": 3.481575444397339,
      "tokens_seen": 1679556608
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 927219,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.202238082885742,
      "objective/train/theoretical_loss": 3.4815637250703304,
      "objective/train/tokens_used": 38481376,
      "theoretical_loss": 3.4815637250703304,
      "tokens_seen": 1679622144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994949494949495,
      "loss": 2.5746,
      "theoretical_loss": 3.4815637250703304,
      "tokens_seen": 1679622144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994747474747474,
      "loss": 2.6998,
      "theoretical_loss": 3.4815520063286103,
      "tokens_seen": 1679687680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994545454545454,
      "loss": 2.4526,
      "theoretical_loss": 3.4815402881721274,
      "tokens_seen": 1679753216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994343434343435,
      "loss": 2.6013,
      "theoretical_loss": 3.4815285706008288,
      "tokens_seen": 1679818752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994141414141415,
      "loss": 2.2639,
      "theoretical_loss": 3.4815168536146626,
      "tokens_seen": 1679884288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993939393939394,
      "loss": 2.5337,
      "theoretical_loss": 3.481505137213577,
      "tokens_seen": 1679949824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993737373737374,
      "loss": 2.6064,
      "theoretical_loss": 3.48149342139752,
      "tokens_seen": 1680015360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993535353535354,
      "loss": 2.4297,
      "theoretical_loss": 3.4814817061664396,
      "tokens_seen": 1680080896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993333333333334,
      "loss": 2.548,
      "theoretical_loss": 3.481469991520283,
      "tokens_seen": 1680146432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993131313131313,
      "loss": 2.5833,
      "theoretical_loss": 3.481458277458999,
      "tokens_seen": 1680211968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992929292929293,
      "loss": 2.6257,
      "theoretical_loss": 3.4814465639825354,
      "tokens_seen": 1680277504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992727272727272,
      "loss": 2.2464,
      "theoretical_loss": 3.4814348510908397,
      "tokens_seen": 1680343040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992525252525252,
      "loss": 2.526,
      "theoretical_loss": 3.481423138783861,
      "tokens_seen": 1680408576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992323232323232,
      "loss": 2.5034,
      "theoretical_loss": 3.4814114270615466,
      "tokens_seen": 1680474112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992121212121212,
      "loss": 2.5144,
      "theoretical_loss": 3.4813997159238443,
      "tokens_seen": 1680539648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991919191919192,
      "loss": 2.7348,
      "theoretical_loss": 3.481388005370702,
      "tokens_seen": 1680605184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991717171717172,
      "loss": 2.3679,
      "theoretical_loss": 3.481376295402069,
      "tokens_seen": 1680670720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991515151515152,
      "loss": 2.5985,
      "theoretical_loss": 3.481364586017892,
      "tokens_seen": 1680736256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991313131313132,
      "loss": 2.6316,
      "theoretical_loss": 3.4813528772181193,
      "tokens_seen": 1680801792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991111111111111,
      "loss": 2.601,
      "theoretical_loss": 3.4813411690026994,
      "tokens_seen": 1680867328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990909090909091,
      "loss": 2.57,
      "theoretical_loss": 3.4813294613715797,
      "tokens_seen": 1680932864
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990707070707071,
      "loss": 2.4923,
      "theoretical_loss": 3.4813177543247096,
      "tokens_seen": 1680998400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990505050505051,
      "loss": 2.5084,
      "theoretical_loss": 3.4813060478620352,
      "tokens_seen": 1681063936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499030303030303,
      "loss": 2.6574,
      "theoretical_loss": 3.4812943419835065,
      "tokens_seen": 1681129472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499010101010101,
      "loss": 2.7915,
      "theoretical_loss": 3.48128263668907,
      "tokens_seen": 1681195008
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 927968,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5613365173339844,
      "objective/train/theoretical_loss": 3.481270931978675,
      "objective/train/tokens_used": 40119776,
      "theoretical_loss": 3.481270931978675,
      "tokens_seen": 1681260544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498989898989899,
      "loss": 2.4737,
      "theoretical_loss": 3.481270931978675,
      "tokens_seen": 1681260544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498969696969697,
      "loss": 2.6981,
      "theoretical_loss": 3.4812592278522687,
      "tokens_seen": 1681326080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498949494949495,
      "loss": 2.6027,
      "theoretical_loss": 3.4812475243098,
      "tokens_seen": 1681391616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498929292929293,
      "loss": 2.6348,
      "theoretical_loss": 3.4812358213512162,
      "tokens_seen": 1681457152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989090909090909,
      "loss": 2.6082,
      "theoretical_loss": 3.4812241189764657,
      "tokens_seen": 1681522688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988888888888889,
      "loss": 2.6419,
      "theoretical_loss": 3.4812124171854966,
      "tokens_seen": 1681588224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988686868686869,
      "loss": 2.6168,
      "theoretical_loss": 3.4812007159782574,
      "tokens_seen": 1681653760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988484848484849,
      "loss": 2.5625,
      "theoretical_loss": 3.481189015354696,
      "tokens_seen": 1681719296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988282828282828,
      "loss": 2.6334,
      "theoretical_loss": 3.4811773153147607,
      "tokens_seen": 1681784832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988080808080808,
      "loss": 2.6787,
      "theoretical_loss": 3.481165615858399,
      "tokens_seen": 1681850368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987878787878788,
      "loss": 2.8212,
      "theoretical_loss": 3.48115391698556,
      "tokens_seen": 1681915904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987676767676767,
      "loss": 2.5745,
      "theoretical_loss": 3.481142218696191,
      "tokens_seen": 1681981440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987474747474747,
      "loss": 2.848,
      "theoretical_loss": 3.4811305209902406,
      "tokens_seen": 1682046976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987272727272728,
      "loss": 2.4886,
      "theoretical_loss": 3.481118823867657,
      "tokens_seen": 1682112512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987070707070708,
      "loss": 2.4485,
      "theoretical_loss": 3.4811071273283885,
      "tokens_seen": 1682178048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986868686868687,
      "loss": 2.7483,
      "theoretical_loss": 3.481095431372383,
      "tokens_seen": 1682243584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986666666666667,
      "loss": 2.472,
      "theoretical_loss": 3.4810837359995883,
      "tokens_seen": 1682309120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986464646464647,
      "loss": 2.6229,
      "theoretical_loss": 3.481072041209954,
      "tokens_seen": 1682374656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986262626262627,
      "loss": 2.679,
      "theoretical_loss": 3.4810603470034263,
      "tokens_seen": 1682440192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986060606060606,
      "loss": 2.5162,
      "theoretical_loss": 3.481048653379955,
      "tokens_seen": 1682505728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985858585858586,
      "loss": 2.6945,
      "theoretical_loss": 3.4810369603394884,
      "tokens_seen": 1682571264
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985656565656565,
      "loss": 2.6558,
      "theoretical_loss": 3.4810252678819733,
      "tokens_seen": 1682636800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985454545454545,
      "loss": 2.6452,
      "theoretical_loss": 3.481013576007359,
      "tokens_seen": 1682702336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985252525252525,
      "loss": 2.6641,
      "theoretical_loss": 3.481001884715594,
      "tokens_seen": 1682767872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985050505050506,
      "loss": 2.5645,
      "theoretical_loss": 3.4809901940066257,
      "tokens_seen": 1682833408
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 929249,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0171878337860107,
      "objective/train/theoretical_loss": 3.4809785038804026,
      "objective/train/tokens_used": 41758176,
      "theoretical_loss": 3.4809785038804026,
      "tokens_seen": 1682898944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984848484848485,
      "loss": 2.6663,
      "theoretical_loss": 3.4809785038804026,
      "tokens_seen": 1682898944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984646464646465,
      "loss": 2.733,
      "theoretical_loss": 3.480966814336873,
      "tokens_seen": 1682964480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984444444444445,
      "loss": 2.5084,
      "theoretical_loss": 3.480955125375986,
      "tokens_seen": 1683030016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984242424242425,
      "loss": 2.7494,
      "theoretical_loss": 3.4809434369976886,
      "tokens_seen": 1683095552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984040404040404,
      "loss": 2.3954,
      "theoretical_loss": 3.48093174920193,
      "tokens_seen": 1683161088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983838383838384,
      "loss": 2.5031,
      "theoretical_loss": 3.480920061988657,
      "tokens_seen": 1683226624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983636363636363,
      "loss": 2.5551,
      "theoretical_loss": 3.48090837535782,
      "tokens_seen": 1683292160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983434343434344,
      "loss": 2.5852,
      "theoretical_loss": 3.480896689309366,
      "tokens_seen": 1683357696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983232323232323,
      "loss": 2.7372,
      "theoretical_loss": 3.4808850038432437,
      "tokens_seen": 1683423232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983030303030303,
      "loss": 2.6248,
      "theoretical_loss": 3.480873318959401,
      "tokens_seen": 1683488768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982828282828282,
      "loss": 2.3415,
      "theoretical_loss": 3.4808616346577868,
      "tokens_seen": 1683554304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982626262626263,
      "loss": 2.6582,
      "theoretical_loss": 3.480849950938349,
      "tokens_seen": 1683619840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982424242424243,
      "loss": 2.5619,
      "theoretical_loss": 3.4808382678010363,
      "tokens_seen": 1683685376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982222222222223,
      "loss": 2.7625,
      "theoretical_loss": 3.480826585245797,
      "tokens_seen": 1683750912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982020202020202,
      "loss": 2.5896,
      "theoretical_loss": 3.480814903272579,
      "tokens_seen": 1683816448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981818181818182,
      "loss": 2.7364,
      "theoretical_loss": 3.4808032218813305,
      "tokens_seen": 1683881984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981616161616162,
      "loss": 2.7586,
      "theoretical_loss": 3.4807915410720005,
      "tokens_seen": 1683947520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981414141414142,
      "loss": 2.5097,
      "theoretical_loss": 3.480779860844538,
      "tokens_seen": 1684013056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981212121212121,
      "loss": 2.7616,
      "theoretical_loss": 3.4807681811988895,
      "tokens_seen": 1684078592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981010101010101,
      "loss": 2.5229,
      "theoretical_loss": 3.480756502135005,
      "tokens_seen": 1684144128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498080808080808,
      "loss": 2.5937,
      "theoretical_loss": 3.4807448236528318,
      "tokens_seen": 1684209664
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498060606060606,
      "loss": 2.705,
      "theoretical_loss": 3.480733145752319,
      "tokens_seen": 1684275200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498040404040404,
      "loss": 2.6938,
      "theoretical_loss": 3.4807214684334147,
      "tokens_seen": 1684340736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980202020202021,
      "loss": 2.6702,
      "theoretical_loss": 3.480709791696068,
      "tokens_seen": 1684406272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498,
      "loss": 2.466,
      "theoretical_loss": 3.4806981155402257,
      "tokens_seen": 1684471808
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 929812,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.769881010055542,
      "objective/train/theoretical_loss": 3.4806864399658375,
      "objective/train/tokens_used": 43396576,
      "theoretical_loss": 3.4806864399658375,
      "tokens_seen": 1684537344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497979797979798,
      "loss": 2.5807,
      "theoretical_loss": 3.4806864399658375,
      "tokens_seen": 1684537344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497959595959596,
      "loss": 2.532,
      "theoretical_loss": 3.4806747649728518,
      "tokens_seen": 1684602880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497939393939394,
      "loss": 2.4939,
      "theoretical_loss": 3.480663090561217,
      "tokens_seen": 1684668416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979191919191919,
      "loss": 2.6286,
      "theoretical_loss": 3.480651416730881,
      "tokens_seen": 1684733952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978989898989899,
      "loss": 2.5486,
      "theoretical_loss": 3.4806397434817926,
      "tokens_seen": 1684799488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978787878787879,
      "loss": 2.4135,
      "theoretical_loss": 3.4806280708139,
      "tokens_seen": 1684865024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978585858585858,
      "loss": 2.487,
      "theoretical_loss": 3.480616398727152,
      "tokens_seen": 1684930560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978383838383838,
      "loss": 2.5598,
      "theoretical_loss": 3.480604727221497,
      "tokens_seen": 1684996096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978181818181818,
      "loss": 2.6736,
      "theoretical_loss": 3.4805930562968834,
      "tokens_seen": 1685061632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977979797979799,
      "loss": 2.4844,
      "theoretical_loss": 3.4805813859532595,
      "tokens_seen": 1685127168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977777777777778,
      "loss": 2.7091,
      "theoretical_loss": 3.4805697161905744,
      "tokens_seen": 1685192704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977575757575758,
      "loss": 2.5024,
      "theoretical_loss": 3.480558047008776,
      "tokens_seen": 1685258240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977373737373738,
      "loss": 2.5198,
      "theoretical_loss": 3.480546378407812,
      "tokens_seen": 1685323776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977171717171717,
      "loss": 2.44,
      "theoretical_loss": 3.4805347103876327,
      "tokens_seen": 1685389312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976969696969697,
      "loss": 2.524,
      "theoretical_loss": 3.480523042948186,
      "tokens_seen": 1685454848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976767676767677,
      "loss": 2.6416,
      "theoretical_loss": 3.48051137608942,
      "tokens_seen": 1685520384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976565656565656,
      "loss": 2.5126,
      "theoretical_loss": 3.480499709811283,
      "tokens_seen": 1685585920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976363636363636,
      "loss": 2.6729,
      "theoretical_loss": 3.4804880441137245,
      "tokens_seen": 1685651456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976161616161616,
      "loss": 2.4778,
      "theoretical_loss": 3.480476378996692,
      "tokens_seen": 1685716992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975959595959596,
      "loss": 2.5328,
      "theoretical_loss": 3.4804647144601346,
      "tokens_seen": 1685782528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975757575757575,
      "loss": 2.6505,
      "theoretical_loss": 3.480453050504001,
      "tokens_seen": 1685848064
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975555555555556,
      "loss": 2.5126,
      "theoretical_loss": 3.4804413871282396,
      "tokens_seen": 1685913600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975353535353536,
      "loss": 2.6655,
      "theoretical_loss": 3.480429724332798,
      "tokens_seen": 1685979136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975151515151516,
      "loss": 2.5183,
      "theoretical_loss": 3.480418062117627,
      "tokens_seen": 1686044672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974949494949495,
      "loss": 2.2442,
      "theoretical_loss": 3.480406400482673,
      "tokens_seen": 1686110208
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 931049,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4905221462249756,
      "objective/train/theoretical_loss": 3.4803947394278856,
      "objective/train/tokens_used": 45034976,
      "theoretical_loss": 3.4803947394278856,
      "tokens_seen": 1686175744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974747474747475,
      "loss": 2.4943,
      "theoretical_loss": 3.4803947394278856,
      "tokens_seen": 1686175744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974545454545454,
      "loss": 2.2533,
      "theoretical_loss": 3.4803830789532126,
      "tokens_seen": 1686241280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974343434343435,
      "loss": 2.7416,
      "theoretical_loss": 3.4803714190586037,
      "tokens_seen": 1686306816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974141414141414,
      "loss": 2.4504,
      "theoretical_loss": 3.4803597597440072,
      "tokens_seen": 1686372352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973939393939394,
      "loss": 2.7452,
      "theoretical_loss": 3.4803481010093718,
      "tokens_seen": 1686437888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973737373737373,
      "loss": 2.6507,
      "theoretical_loss": 3.4803364428546453,
      "tokens_seen": 1686503424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973535353535353,
      "loss": 2.2988,
      "theoretical_loss": 3.4803247852797767,
      "tokens_seen": 1686568960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973333333333334,
      "loss": 2.5663,
      "theoretical_loss": 3.480313128284715,
      "tokens_seen": 1686634496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973131313131314,
      "loss": 2.518,
      "theoretical_loss": 3.4803014718694087,
      "tokens_seen": 1686700032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972929292929293,
      "loss": 2.4251,
      "theoretical_loss": 3.4802898160338063,
      "tokens_seen": 1686765568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972727272727273,
      "loss": 2.5484,
      "theoretical_loss": 3.480278160777856,
      "tokens_seen": 1686831104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972525252525253,
      "loss": 2.3676,
      "theoretical_loss": 3.4802665061015077,
      "tokens_seen": 1686896640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972323232323233,
      "loss": 2.3593,
      "theoretical_loss": 3.480254852004709,
      "tokens_seen": 1686962176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972121212121212,
      "loss": 2.5334,
      "theoretical_loss": 3.4802431984874094,
      "tokens_seen": 1687027712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971919191919192,
      "loss": 2.5443,
      "theoretical_loss": 3.4802315455495565,
      "tokens_seen": 1687093248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971717171717171,
      "loss": 2.473,
      "theoretical_loss": 3.4802198931910993,
      "tokens_seen": 1687158784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971515151515152,
      "loss": 2.5458,
      "theoretical_loss": 3.480208241411987,
      "tokens_seen": 1687224320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971313131313131,
      "loss": 2.5911,
      "theoretical_loss": 3.4801965902121683,
      "tokens_seen": 1687289856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971111111111111,
      "loss": 2.5036,
      "theoretical_loss": 3.480184939591591,
      "tokens_seen": 1687355392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497090909090909,
      "loss": 2.452,
      "theoretical_loss": 3.480173289550205,
      "tokens_seen": 1687420928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970707070707071,
      "loss": 2.3563,
      "theoretical_loss": 3.4801616400879585,
      "tokens_seen": 1687486464
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970505050505051,
      "loss": 2.6628,
      "theoretical_loss": 3.4801499912048,
      "tokens_seen": 1687552000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970303030303031,
      "loss": 2.4553,
      "theoretical_loss": 3.480138342900678,
      "tokens_seen": 1687617536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497010101010101,
      "loss": 2.6518,
      "theoretical_loss": 3.480126695175542,
      "tokens_seen": 1687683072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496989898989899,
      "loss": 2.6693,
      "theoretical_loss": 3.48011504802934,
      "tokens_seen": 1687748608
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 931336,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.531233072280884,
      "objective/train/theoretical_loss": 3.4801034014620216,
      "objective/train/tokens_used": 46673376,
      "theoretical_loss": 3.4801034014620216,
      "tokens_seen": 1687814144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496969696969697,
      "loss": 2.5694,
      "theoretical_loss": 3.4801034014620216,
      "tokens_seen": 1687814144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496949494949495,
      "loss": 2.4312,
      "theoretical_loss": 3.4800917554735347,
      "tokens_seen": 1687879680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969292929292929,
      "loss": 2.6901,
      "theoretical_loss": 3.4800801100638283,
      "tokens_seen": 1687945216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969090909090909,
      "loss": 2.7315,
      "theoretical_loss": 3.480068465232852,
      "tokens_seen": 1688010752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968888888888889,
      "loss": 2.5066,
      "theoretical_loss": 3.480056820980553,
      "tokens_seen": 1688076288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968686868686868,
      "loss": 2.4754,
      "theoretical_loss": 3.4800451773068817,
      "tokens_seen": 1688141824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968484848484849,
      "loss": 2.8056,
      "theoretical_loss": 3.4800335342117856,
      "tokens_seen": 1688207360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968282828282829,
      "loss": 2.6707,
      "theoretical_loss": 3.4800218916952144,
      "tokens_seen": 1688272896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968080808080808,
      "loss": 2.6028,
      "theoretical_loss": 3.4800102497571164,
      "tokens_seen": 1688338432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967878787878788,
      "loss": 2.447,
      "theoretical_loss": 3.47999860839744,
      "tokens_seen": 1688403968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967676767676768,
      "loss": 2.5797,
      "theoretical_loss": 3.479986967616135,
      "tokens_seen": 1688469504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967474747474748,
      "loss": 2.6184,
      "theoretical_loss": 3.4799753274131495,
      "tokens_seen": 1688535040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967272727272727,
      "loss": 2.651,
      "theoretical_loss": 3.479963687788433,
      "tokens_seen": 1688600576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967070707070707,
      "loss": 2.6058,
      "theoretical_loss": 3.4799520487419335,
      "tokens_seen": 1688666112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966868686868687,
      "loss": 2.5568,
      "theoretical_loss": 3.4799404102736005,
      "tokens_seen": 1688731648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966666666666666,
      "loss": 2.6325,
      "theoretical_loss": 3.4799287723833823,
      "tokens_seen": 1688797184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966464646464646,
      "loss": 2.4631,
      "theoretical_loss": 3.4799171350712284,
      "tokens_seen": 1688862720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966262626262627,
      "loss": 2.8322,
      "theoretical_loss": 3.4799054983370867,
      "tokens_seen": 1688928256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966060606060607,
      "loss": 2.5475,
      "theoretical_loss": 3.4798938621809077,
      "tokens_seen": 1688993792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965858585858586,
      "loss": 2.3884,
      "theoretical_loss": 3.4798822266026384,
      "tokens_seen": 1689059328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965656565656566,
      "loss": 2.6883,
      "theoretical_loss": 3.4798705916022286,
      "tokens_seen": 1689124864
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965454545454546,
      "loss": 2.4336,
      "theoretical_loss": 3.4798589571796272,
      "tokens_seen": 1689190400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965252525252525,
      "loss": 2.6315,
      "theoretical_loss": 3.4798473233347833,
      "tokens_seen": 1689255936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965050505050505,
      "loss": 2.5791,
      "theoretical_loss": 3.4798356900676444,
      "tokens_seen": 1689321472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964848484848485,
      "loss": 2.4727,
      "theoretical_loss": 3.4798240573781616,
      "tokens_seen": 1689387008
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 932508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.615687847137451,
      "objective/train/theoretical_loss": 3.479812425266282,
      "objective/train/tokens_used": 48311776,
      "theoretical_loss": 3.479812425266282,
      "tokens_seen": 1689452544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964646464646464,
      "loss": 2.4249,
      "theoretical_loss": 3.479812425266282,
      "tokens_seen": 1689452544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964444444444444,
      "loss": 2.4928,
      "theoretical_loss": 3.4798007937319557,
      "tokens_seen": 1689518080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964242424242424,
      "loss": 2.7017,
      "theoretical_loss": 3.479789162775131,
      "tokens_seen": 1689583616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964040404040405,
      "loss": 2.6564,
      "theoretical_loss": 3.4797775323957567,
      "tokens_seen": 1689649152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963838383838384,
      "loss": 2.4665,
      "theoretical_loss": 3.479765902593782,
      "tokens_seen": 1689714688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963636363636364,
      "loss": 2.7239,
      "theoretical_loss": 3.479754273369156,
      "tokens_seen": 1689780224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963434343434344,
      "loss": 2.3918,
      "theoretical_loss": 3.4797426447218274,
      "tokens_seen": 1689845760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963232323232324,
      "loss": 2.6125,
      "theoretical_loss": 3.479731016651745,
      "tokens_seen": 1689911296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963030303030303,
      "loss": 2.9569,
      "theoretical_loss": 3.4797193891588583,
      "tokens_seen": 1689976832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962828282828283,
      "loss": 2.43,
      "theoretical_loss": 3.479707762243116,
      "tokens_seen": 1690042368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962626262626262,
      "loss": 2.442,
      "theoretical_loss": 3.479696135904467,
      "tokens_seen": 1690107904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962424242424243,
      "loss": 2.9013,
      "theoretical_loss": 3.47968451014286,
      "tokens_seen": 1690173440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962222222222222,
      "loss": 2.6167,
      "theoretical_loss": 3.479672884958245,
      "tokens_seen": 1690238976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962020202020202,
      "loss": 2.7755,
      "theoretical_loss": 3.4796612603505697,
      "tokens_seen": 1690304512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961818181818181,
      "loss": 2.6197,
      "theoretical_loss": 3.4796496363197837,
      "tokens_seen": 1690370048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961616161616161,
      "loss": 2.6195,
      "theoretical_loss": 3.479638012865836,
      "tokens_seen": 1690435584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961414141414142,
      "loss": 2.691,
      "theoretical_loss": 3.4796263899886757,
      "tokens_seen": 1690501120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961212121212122,
      "loss": 2.7274,
      "theoretical_loss": 3.479614767688252,
      "tokens_seen": 1690566656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961010101010101,
      "loss": 2.7084,
      "theoretical_loss": 3.4796031459645134,
      "tokens_seen": 1690632192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960808080808081,
      "loss": 2.6796,
      "theoretical_loss": 3.479591524817409,
      "tokens_seen": 1690697728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960606060606061,
      "loss": 2.3627,
      "theoretical_loss": 3.4795799042468882,
      "tokens_seen": 1690763264
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960404040404041,
      "loss": 2.5495,
      "theoretical_loss": 3.4795682842529,
      "tokens_seen": 1690828800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496020202020202,
      "loss": 2.6268,
      "theoretical_loss": 3.4795566648353935,
      "tokens_seen": 1690894336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496,
      "loss": 2.614,
      "theoretical_loss": 3.4795450459943167,
      "tokens_seen": 1690959872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959797979797979,
      "loss": 2.4627,
      "theoretical_loss": 3.4795334277296206,
      "tokens_seen": 1691025408
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 933011,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.424670696258545,
      "objective/train/theoretical_loss": 3.479521810041253,
      "objective/train/tokens_used": 49950176,
      "theoretical_loss": 3.479521810041253,
      "tokens_seen": 1691090944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959595959595959,
      "loss": 2.6558,
      "theoretical_loss": 3.479521810041253,
      "tokens_seen": 1691090944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495939393939394,
      "loss": 2.5806,
      "theoretical_loss": 3.4795101929291627,
      "tokens_seen": 1691156480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495919191919192,
      "loss": 2.6571,
      "theoretical_loss": 3.4794985763932997,
      "tokens_seen": 1691222016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958989898989899,
      "loss": 2.6959,
      "theoretical_loss": 3.4794869604336123,
      "tokens_seen": 1691287552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958787878787879,
      "loss": 2.4706,
      "theoretical_loss": 3.47947534505005,
      "tokens_seen": 1691353088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958585858585859,
      "loss": 2.572,
      "theoretical_loss": 3.4794637302425624,
      "tokens_seen": 1691418624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958383838383839,
      "loss": 2.7139,
      "theoretical_loss": 3.479452116011098,
      "tokens_seen": 1691484160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958181818181818,
      "loss": 2.6099,
      "theoretical_loss": 3.4794405023556054,
      "tokens_seen": 1691549696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957979797979798,
      "loss": 2.4053,
      "theoretical_loss": 3.4794288892760346,
      "tokens_seen": 1691615232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957777777777778,
      "loss": 2.8233,
      "theoretical_loss": 3.4794172767723346,
      "tokens_seen": 1691680768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957575757575757,
      "loss": 2.7725,
      "theoretical_loss": 3.479405664844454,
      "tokens_seen": 1691746304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957373737373737,
      "loss": 2.6491,
      "theoretical_loss": 3.479394053492343,
      "tokens_seen": 1691811840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957171717171717,
      "loss": 2.516,
      "theoretical_loss": 3.4793824427159494,
      "tokens_seen": 1691877376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956969696969698,
      "loss": 2.474,
      "theoretical_loss": 3.4793708325152237,
      "tokens_seen": 1691942912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956767676767677,
      "loss": 2.3708,
      "theoretical_loss": 3.479359222890114,
      "tokens_seen": 1692008448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956565656565657,
      "loss": 2.6737,
      "theoretical_loss": 3.47934761384057,
      "tokens_seen": 1692073984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956363636363637,
      "loss": 2.5308,
      "theoretical_loss": 3.479336005366541,
      "tokens_seen": 1692139520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956161616161616,
      "loss": 2.4276,
      "theoretical_loss": 3.4793243974679755,
      "tokens_seen": 1692205056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955959595959596,
      "loss": 2.6885,
      "theoretical_loss": 3.479312790144823,
      "tokens_seen": 1692270592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955757575757576,
      "loss": 2.2714,
      "theoretical_loss": 3.479301183397033,
      "tokens_seen": 1692336128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955555555555556,
      "loss": 2.4882,
      "theoretical_loss": 3.479289577224555,
      "tokens_seen": 1692401664
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955353535353535,
      "loss": 2.4553,
      "theoretical_loss": 3.4792779716273365,
      "tokens_seen": 1692467200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955151515151515,
      "loss": 2.6182,
      "theoretical_loss": 3.479266366605329,
      "tokens_seen": 1692532736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954949494949495,
      "loss": 2.5212,
      "theoretical_loss": 3.4792547621584804,
      "tokens_seen": 1692598272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954747474747474,
      "loss": 2.4225,
      "theoretical_loss": 3.47924315828674,
      "tokens_seen": 1692663808
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 934010,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.852196455001831,
      "objective/train/theoretical_loss": 3.479231554990057,
      "objective/train/tokens_used": 51588576,
      "theoretical_loss": 3.479231554990057,
      "tokens_seen": 1692729344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954545454545455,
      "loss": 2.6898,
      "theoretical_loss": 3.479231554990057,
      "tokens_seen": 1692729344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954343434343435,
      "loss": 2.6057,
      "theoretical_loss": 3.479219952268381,
      "tokens_seen": 1692794880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954141414141415,
      "loss": 2.5534,
      "theoretical_loss": 3.4792083501216613,
      "tokens_seen": 1692860416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953939393939394,
      "loss": 2.602,
      "theoretical_loss": 3.479196748549847,
      "tokens_seen": 1692925952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953737373737374,
      "loss": 2.6322,
      "theoretical_loss": 3.4791851475528874,
      "tokens_seen": 1692991488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953535353535354,
      "loss": 2.5945,
      "theoretical_loss": 3.4791735471307312,
      "tokens_seen": 1693057024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953333333333334,
      "loss": 2.6971,
      "theoretical_loss": 3.4791619472833286,
      "tokens_seen": 1693122560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953131313131313,
      "loss": 2.6478,
      "theoretical_loss": 3.479150348010628,
      "tokens_seen": 1693188096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952929292929293,
      "loss": 2.5666,
      "theoretical_loss": 3.479138749312579,
      "tokens_seen": 1693253632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952727272727272,
      "loss": 2.5181,
      "theoretical_loss": 3.4791271511891315,
      "tokens_seen": 1693319168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952525252525252,
      "loss": 2.566,
      "theoretical_loss": 3.4791155536402343,
      "tokens_seen": 1693384704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952323232323232,
      "loss": 2.7651,
      "theoretical_loss": 3.479103956665836,
      "tokens_seen": 1693450240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952121212121213,
      "loss": 2.5736,
      "theoretical_loss": 3.479092360265887,
      "tokens_seen": 1693515776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951919191919192,
      "loss": 2.4578,
      "theoretical_loss": 3.4790807644403365,
      "tokens_seen": 1693581312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951717171717172,
      "loss": 2.6557,
      "theoretical_loss": 3.479069169189133,
      "tokens_seen": 1693646848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951515151515152,
      "loss": 2.6888,
      "theoretical_loss": 3.479057574512227,
      "tokens_seen": 1693712384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951313131313132,
      "loss": 2.5443,
      "theoretical_loss": 3.4790459804095666,
      "tokens_seen": 1693777920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951111111111111,
      "loss": 2.7016,
      "theoretical_loss": 3.479034386881102,
      "tokens_seen": 1693843456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950909090909091,
      "loss": 2.4926,
      "theoretical_loss": 3.4790227939267826,
      "tokens_seen": 1693908992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495070707070707,
      "loss": 2.7287,
      "theoretical_loss": 3.479011201546557,
      "tokens_seen": 1693974528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950505050505051,
      "loss": 2.71,
      "theoretical_loss": 3.478999609740375,
      "tokens_seen": 1694040064
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495030303030303,
      "loss": 2.4208,
      "theoretical_loss": 3.4789880185081863,
      "tokens_seen": 1694105600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495010101010101,
      "loss": 2.4963,
      "theoretical_loss": 3.4789764278499398,
      "tokens_seen": 1694171136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494989898989899,
      "loss": 2.7918,
      "theoretical_loss": 3.478964837765585,
      "tokens_seen": 1694236672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494969696969697,
      "loss": 2.3404,
      "theoretical_loss": 3.4789532482550714,
      "tokens_seen": 1694302208
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 934404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7420923709869385,
      "objective/train/theoretical_loss": 3.478941659318348,
      "objective/train/tokens_used": 53226976,
      "theoretical_loss": 3.478941659318348,
      "tokens_seen": 1694367744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494949494949495,
      "loss": 2.6966,
      "theoretical_loss": 3.478941659318348,
      "tokens_seen": 1694367744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494929292929293,
      "loss": 2.652,
      "theoretical_loss": 3.478930070955365,
      "tokens_seen": 1694433280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949090909090909,
      "loss": 2.5731,
      "theoretical_loss": 3.478918483166071,
      "tokens_seen": 1694498816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948888888888889,
      "loss": 2.5032,
      "theoretical_loss": 3.478906895950416,
      "tokens_seen": 1694564352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948686868686869,
      "loss": 2.7563,
      "theoretical_loss": 3.4788953093083492,
      "tokens_seen": 1694629888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948484848484849,
      "loss": 2.5857,
      "theoretical_loss": 3.4788837232398198,
      "tokens_seen": 1694695424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948282828282828,
      "loss": 2.5759,
      "theoretical_loss": 3.4788721377447773,
      "tokens_seen": 1694760960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948080808080808,
      "loss": 2.4183,
      "theoretical_loss": 3.4788605528231713,
      "tokens_seen": 1694826496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947878787878787,
      "loss": 2.4682,
      "theoretical_loss": 3.478848968474951,
      "tokens_seen": 1694892032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947676767676767,
      "loss": 2.7972,
      "theoretical_loss": 3.4788373847000664,
      "tokens_seen": 1694957568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947474747474748,
      "loss": 2.5319,
      "theoretical_loss": 3.4788258014984663,
      "tokens_seen": 1695023104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947272727272728,
      "loss": 2.4632,
      "theoretical_loss": 3.478814218870101,
      "tokens_seen": 1695088640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947070707070707,
      "loss": 2.5513,
      "theoretical_loss": 3.478802636814919,
      "tokens_seen": 1695154176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946868686868687,
      "loss": 2.5416,
      "theoretical_loss": 3.4787910553328705,
      "tokens_seen": 1695219712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946666666666667,
      "loss": 2.6542,
      "theoretical_loss": 3.4787794744239044,
      "tokens_seen": 1695285248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946464646464647,
      "loss": 2.5943,
      "theoretical_loss": 3.4787678940879707,
      "tokens_seen": 1695350784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946262626262626,
      "loss": 2.5594,
      "theoretical_loss": 3.4787563143250186,
      "tokens_seen": 1695416320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946060606060606,
      "loss": 2.4973,
      "theoretical_loss": 3.478744735134998,
      "tokens_seen": 1695481856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945858585858586,
      "loss": 2.5056,
      "theoretical_loss": 3.4787331565178574,
      "tokens_seen": 1695547392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945656565656565,
      "loss": 2.5805,
      "theoretical_loss": 3.4787215784735475,
      "tokens_seen": 1695612928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945454545454545,
      "loss": 2.5839,
      "theoretical_loss": 3.4787100010020175,
      "tokens_seen": 1695678464
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945252525252526,
      "loss": 2.551,
      "theoretical_loss": 3.478698424103216,
      "tokens_seen": 1695744000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945050505050506,
      "loss": 2.7043,
      "theoretical_loss": 3.4786868477770945,
      "tokens_seen": 1695809536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944848484848485,
      "loss": 2.6309,
      "theoretical_loss": 3.4786752720236005,
      "tokens_seen": 1695875072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944646464646465,
      "loss": 2.7457,
      "theoretical_loss": 3.4786636968426845,
      "tokens_seen": 1695940608
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 935560,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.167574882507324,
      "objective/train/theoretical_loss": 3.478652122234296,
      "objective/train/tokens_used": 54865376,
      "theoretical_loss": 3.478652122234296,
      "tokens_seen": 1696006144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944444444444445,
      "loss": 2.4537,
      "theoretical_loss": 3.478652122234296,
      "tokens_seen": 1696006144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944242424242424,
      "loss": 2.5082,
      "theoretical_loss": 3.4786405481983844,
      "tokens_seen": 1696071680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944040404040404,
      "loss": 2.716,
      "theoretical_loss": 3.4786289747348995,
      "tokens_seen": 1696137216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943838383838384,
      "loss": 2.5441,
      "theoretical_loss": 3.478617401843791,
      "tokens_seen": 1696202752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943636363636363,
      "loss": 2.5488,
      "theoretical_loss": 3.4786058295250077,
      "tokens_seen": 1696268288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943434343434343,
      "loss": 2.5682,
      "theoretical_loss": 3.4785942577785,
      "tokens_seen": 1696333824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943232323232323,
      "loss": 2.4028,
      "theoretical_loss": 3.4785826866042173,
      "tokens_seen": 1696399360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943030303030303,
      "loss": 2.5064,
      "theoretical_loss": 3.478571116002109,
      "tokens_seen": 1696464896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942828282828283,
      "loss": 2.4192,
      "theoretical_loss": 3.4785595459721246,
      "tokens_seen": 1696530432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942626262626263,
      "loss": 2.6105,
      "theoretical_loss": 3.478547976514214,
      "tokens_seen": 1696595968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942424242424243,
      "loss": 2.6893,
      "theoretical_loss": 3.4785364076283267,
      "tokens_seen": 1696661504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942222222222223,
      "loss": 2.4376,
      "theoretical_loss": 3.478524839314413,
      "tokens_seen": 1696727040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942020202020202,
      "loss": 2.7309,
      "theoretical_loss": 3.478513271572421,
      "tokens_seen": 1696792576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941818181818182,
      "loss": 2.8496,
      "theoretical_loss": 3.4785017044023014,
      "tokens_seen": 1696858112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941616161616161,
      "loss": 2.423,
      "theoretical_loss": 3.4784901378040036,
      "tokens_seen": 1696923648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941414141414142,
      "loss": 2.494,
      "theoretical_loss": 3.4784785717774778,
      "tokens_seen": 1696989184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941212121212121,
      "loss": 2.53,
      "theoretical_loss": 3.4784670063226732,
      "tokens_seen": 1697054720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941010101010101,
      "loss": 2.749,
      "theoretical_loss": 3.478455441439539,
      "tokens_seen": 1697120256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494080808080808,
      "loss": 2.3697,
      "theoretical_loss": 3.4784438771280257,
      "tokens_seen": 1697185792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494060606060606,
      "loss": 2.6826,
      "theoretical_loss": 3.4784323133880823,
      "tokens_seen": 1697251328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940404040404041,
      "loss": 2.5308,
      "theoretical_loss": 3.4784207502196587,
      "tokens_seen": 1697316864
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940202020202021,
      "loss": 2.4568,
      "theoretical_loss": 3.4784091876227046,
      "tokens_seen": 1697382400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494,
      "loss": 2.7751,
      "theoretical_loss": 3.47839762559717,
      "tokens_seen": 1697447936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493979797979798,
      "loss": 2.5151,
      "theoretical_loss": 3.478386064143004,
      "tokens_seen": 1697513472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493959595959596,
      "loss": 2.8479,
      "theoretical_loss": 3.4783745032601567,
      "tokens_seen": 1697579008
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 936682,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.305478572845459,
      "objective/train/theoretical_loss": 3.4783629429485785,
      "objective/train/tokens_used": 56503776,
      "theoretical_loss": 3.4783629429485785,
      "tokens_seen": 1697644544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493939393939394,
      "loss": 2.5086,
      "theoretical_loss": 3.4783629429485785,
      "tokens_seen": 1697644544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939191919191919,
      "loss": 2.7043,
      "theoretical_loss": 3.4783513832082176,
      "tokens_seen": 1697710080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938989898989899,
      "loss": 2.3848,
      "theoretical_loss": 3.4783398240390246,
      "tokens_seen": 1697775616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938787878787878,
      "loss": 2.7298,
      "theoretical_loss": 3.4783282654409495,
      "tokens_seen": 1697841152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938585858585859,
      "loss": 2.5754,
      "theoretical_loss": 3.478316707413941,
      "tokens_seen": 1697906688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938383838383838,
      "loss": 2.7063,
      "theoretical_loss": 3.4783051499579507,
      "tokens_seen": 1697972224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938181818181819,
      "loss": 2.6402,
      "theoretical_loss": 3.478293593072926,
      "tokens_seen": 1698037760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937979797979798,
      "loss": 2.4452,
      "theoretical_loss": 3.4782820367588183,
      "tokens_seen": 1698103296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937777777777778,
      "loss": 2.6044,
      "theoretical_loss": 3.478270481015577,
      "tokens_seen": 1698168832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937575757575758,
      "loss": 2.6482,
      "theoretical_loss": 3.4782589258431518,
      "tokens_seen": 1698234368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937373737373738,
      "loss": 2.7206,
      "theoretical_loss": 3.4782473712414923,
      "tokens_seen": 1698299904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937171717171717,
      "loss": 2.4687,
      "theoretical_loss": 3.4782358172105483,
      "tokens_seen": 1698365440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936969696969697,
      "loss": 2.7044,
      "theoretical_loss": 3.47822426375027,
      "tokens_seen": 1698430976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936767676767677,
      "loss": 2.5647,
      "theoretical_loss": 3.478212710860607,
      "tokens_seen": 1698496512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936565656565657,
      "loss": 2.6277,
      "theoretical_loss": 3.4782011585415087,
      "tokens_seen": 1698562048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936363636363636,
      "loss": 2.7133,
      "theoretical_loss": 3.4781896067929257,
      "tokens_seen": 1698627584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936161616161616,
      "loss": 2.5221,
      "theoretical_loss": 3.4781780556148068,
      "tokens_seen": 1698693120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935959595959597,
      "loss": 2.6434,
      "theoretical_loss": 3.4781665050071027,
      "tokens_seen": 1698758656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935757575757576,
      "loss": 2.7506,
      "theoretical_loss": 3.478154954969763,
      "tokens_seen": 1698824192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935555555555556,
      "loss": 2.578,
      "theoretical_loss": 3.4781434055027365,
      "tokens_seen": 1698889728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935353535353536,
      "loss": 2.5255,
      "theoretical_loss": 3.478131856605975,
      "tokens_seen": 1698955264
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935151515151515,
      "loss": 2.4671,
      "theoretical_loss": 3.4781203082794265,
      "tokens_seen": 1699020800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934949494949495,
      "loss": 2.5736,
      "theoretical_loss": 3.4781087605230425,
      "tokens_seen": 1699086336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934747474747475,
      "loss": 2.5738,
      "theoretical_loss": 3.4780972133367714,
      "tokens_seen": 1699151872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934545454545455,
      "loss": 2.6036,
      "theoretical_loss": 3.478085666720564,
      "tokens_seen": 1699217408
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 937279,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8765950202941895,
      "objective/train/theoretical_loss": 3.47807412067437,
      "objective/train/tokens_used": 58142176,
      "theoretical_loss": 3.47807412067437,
      "tokens_seen": 1699282944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934343434343434,
      "loss": 2.6055,
      "theoretical_loss": 3.47807412067437,
      "tokens_seen": 1699282944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934141414141414,
      "loss": 2.5349,
      "theoretical_loss": 3.4780625751981384,
      "tokens_seen": 1699348480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933939393939394,
      "loss": 2.7046,
      "theoretical_loss": 3.478051030291821,
      "tokens_seen": 1699414016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933737373737373,
      "loss": 2.6394,
      "theoretical_loss": 3.4780394859553656,
      "tokens_seen": 1699479552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933535353535354,
      "loss": 2.4959,
      "theoretical_loss": 3.4780279421887235,
      "tokens_seen": 1699545088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933333333333334,
      "loss": 2.7816,
      "theoretical_loss": 3.478016398991844,
      "tokens_seen": 1699610624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933131313131314,
      "loss": 2.622,
      "theoretical_loss": 3.478004856364677,
      "tokens_seen": 1699676160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932929292929293,
      "loss": 2.6067,
      "theoretical_loss": 3.477993314307173,
      "tokens_seen": 1699741696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932727272727273,
      "loss": 2.6708,
      "theoretical_loss": 3.4779817728192803,
      "tokens_seen": 1699807232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932525252525253,
      "loss": 2.7044,
      "theoretical_loss": 3.477970231900951,
      "tokens_seen": 1699872768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932323232323232,
      "loss": 2.5886,
      "theoretical_loss": 3.477958691552134,
      "tokens_seen": 1699938304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932121212121212,
      "loss": 2.4632,
      "theoretical_loss": 3.4779471517727787,
      "tokens_seen": 1700003840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931919191919192,
      "loss": 2.6896,
      "theoretical_loss": 3.4779356125628365,
      "tokens_seen": 1700069376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931717171717171,
      "loss": 2.655,
      "theoretical_loss": 3.4779240739222557,
      "tokens_seen": 1700134912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931515151515151,
      "loss": 2.6488,
      "theoretical_loss": 3.4779125358509875,
      "tokens_seen": 1700200448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931313131313131,
      "loss": 2.6737,
      "theoretical_loss": 3.4779009983489813,
      "tokens_seen": 1700265984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931111111111112,
      "loss": 2.5862,
      "theoretical_loss": 3.4778894614161873,
      "tokens_seen": 1700331520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930909090909091,
      "loss": 2.5534,
      "theoretical_loss": 3.477877925052555,
      "tokens_seen": 1700397056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930707070707071,
      "loss": 2.6173,
      "theoretical_loss": 3.4778663892580353,
      "tokens_seen": 1700462592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930505050505051,
      "loss": 2.7138,
      "theoretical_loss": 3.477854854032577,
      "tokens_seen": 1700528128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930303030303031,
      "loss": 2.5778,
      "theoretical_loss": 3.477843319376131,
      "tokens_seen": 1700593664
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493010101010101,
      "loss": 2.5768,
      "theoretical_loss": 3.4778317852886476,
      "tokens_seen": 1700659200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492989898989899,
      "loss": 2.8706,
      "theoretical_loss": 3.4778202517700754,
      "tokens_seen": 1700724736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929696969696969,
      "loss": 2.4533,
      "theoretical_loss": 3.4778087188203655,
      "tokens_seen": 1700790272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492949494949495,
      "loss": 2.6694,
      "theoretical_loss": 3.477797186439468,
      "tokens_seen": 1700855808
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 938365,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9837400913238525,
      "objective/train/theoretical_loss": 3.4777856546273327,
      "objective/train/tokens_used": 59780576,
      "theoretical_loss": 3.4777856546273327,
      "tokens_seen": 1700921344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929292929292929,
      "loss": 2.7473,
      "theoretical_loss": 3.4777856546273327,
      "tokens_seen": 1700921344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929090909090909,
      "loss": 2.4495,
      "theoretical_loss": 3.477774123383909,
      "tokens_seen": 1700986880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928888888888888,
      "loss": 2.609,
      "theoretical_loss": 3.477762592709148,
      "tokens_seen": 1701052416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928686868686869,
      "loss": 2.5167,
      "theoretical_loss": 3.477751062602999,
      "tokens_seen": 1701117952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928484848484849,
      "loss": 2.4444,
      "theoretical_loss": 3.477739533065412,
      "tokens_seen": 1701183488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928282828282829,
      "loss": 2.5933,
      "theoretical_loss": 3.4777280040963383,
      "tokens_seen": 1701249024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928080808080808,
      "loss": 2.707,
      "theoretical_loss": 3.4777164756957264,
      "tokens_seen": 1701314560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927878787878788,
      "loss": 2.5917,
      "theoretical_loss": 3.477704947863527,
      "tokens_seen": 1701380096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927676767676768,
      "loss": 2.5673,
      "theoretical_loss": 3.4776934205996906,
      "tokens_seen": 1701445632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927474747474748,
      "loss": 2.8966,
      "theoretical_loss": 3.4776818939041663,
      "tokens_seen": 1701511168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927272727272727,
      "loss": 2.6156,
      "theoretical_loss": 3.477670367776905,
      "tokens_seen": 1701576704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927070707070707,
      "loss": 2.6147,
      "theoretical_loss": 3.477658842217857,
      "tokens_seen": 1701642240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926868686868686,
      "loss": 2.613,
      "theoretical_loss": 3.4776473172269715,
      "tokens_seen": 1701707776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926666666666666,
      "loss": 2.7808,
      "theoretical_loss": 3.4776357928041994,
      "tokens_seen": 1701773312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926464646464647,
      "loss": 2.6945,
      "theoretical_loss": 3.4776242689494903,
      "tokens_seen": 1701838848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926262626262627,
      "loss": 2.7847,
      "theoretical_loss": 3.4776127456627948,
      "tokens_seen": 1701904384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926060606060606,
      "loss": 2.7205,
      "theoretical_loss": 3.4776012229440623,
      "tokens_seen": 1701969920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925858585858586,
      "loss": 2.6992,
      "theoretical_loss": 3.477589700793244,
      "tokens_seen": 1702035456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925656565656566,
      "loss": 2.6906,
      "theoretical_loss": 3.4775781792102887,
      "tokens_seen": 1702100992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925454545454546,
      "loss": 2.659,
      "theoretical_loss": 3.477566658195148,
      "tokens_seen": 1702166528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925252525252525,
      "loss": 2.727,
      "theoretical_loss": 3.477555137747771,
      "tokens_seen": 1702232064
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925050505050505,
      "loss": 2.7682,
      "theoretical_loss": 3.4775436178681085,
      "tokens_seen": 1702297600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924848484848485,
      "loss": 2.5353,
      "theoretical_loss": 3.4775320985561105,
      "tokens_seen": 1702363136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924646464646464,
      "loss": 2.6144,
      "theoretical_loss": 3.477520579811727,
      "tokens_seen": 1702428672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924444444444444,
      "loss": 2.7138,
      "theoretical_loss": 3.477509061634908,
      "tokens_seen": 1702494208
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 938945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2912395000457764,
      "objective/train/theoretical_loss": 3.4774975440256037,
      "objective/train/tokens_used": 61418976,
      "theoretical_loss": 3.4774975440256037,
      "tokens_seen": 1702559744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924242424242425,
      "loss": 2.7635,
      "theoretical_loss": 3.4774975440256037,
      "tokens_seen": 1702559744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924040404040405,
      "loss": 2.761,
      "theoretical_loss": 3.4774860269837653,
      "tokens_seen": 1702625280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923838383838384,
      "loss": 2.6293,
      "theoretical_loss": 3.4774745105093414,
      "tokens_seen": 1702690816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923636363636364,
      "loss": 2.6983,
      "theoretical_loss": 3.4774629946022837,
      "tokens_seen": 1702756352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923434343434344,
      "loss": 2.6871,
      "theoretical_loss": 3.477451479262541,
      "tokens_seen": 1702821888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923232323232323,
      "loss": 2.7305,
      "theoretical_loss": 3.477439964490065,
      "tokens_seen": 1702887424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923030303030303,
      "loss": 2.6126,
      "theoretical_loss": 3.477428450284805,
      "tokens_seen": 1702952960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922828282828283,
      "loss": 2.365,
      "theoretical_loss": 3.4774169366467116,
      "tokens_seen": 1703018496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922626262626262,
      "loss": 2.6455,
      "theoretical_loss": 3.4774054235757346,
      "tokens_seen": 1703084032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922424242424242,
      "loss": 2.5675,
      "theoretical_loss": 3.4773939110718244,
      "tokens_seen": 1703149568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922222222222222,
      "loss": 2.5027,
      "theoretical_loss": 3.4773823991349317,
      "tokens_seen": 1703215104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922020202020202,
      "loss": 2.7077,
      "theoretical_loss": 3.4773708877650065,
      "tokens_seen": 1703280640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921818181818182,
      "loss": 2.5908,
      "theoretical_loss": 3.477359376961999,
      "tokens_seen": 1703346176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921616161616162,
      "loss": 2.6174,
      "theoretical_loss": 3.477347866725859,
      "tokens_seen": 1703411712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921414141414142,
      "loss": 2.6228,
      "theoretical_loss": 3.4773363570565374,
      "tokens_seen": 1703477248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921212121212122,
      "loss": 2.8639,
      "theoretical_loss": 3.4773248479539847,
      "tokens_seen": 1703542784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921010101010101,
      "loss": 2.75,
      "theoretical_loss": 3.47731333941815,
      "tokens_seen": 1703608320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920808080808081,
      "loss": 2.4292,
      "theoretical_loss": 3.477301831448985,
      "tokens_seen": 1703673856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920606060606061,
      "loss": 2.6518,
      "theoretical_loss": 3.4772903240464395,
      "tokens_seen": 1703739392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920404040404041,
      "loss": 2.7381,
      "theoretical_loss": 3.477278817210463,
      "tokens_seen": 1703804928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492020202020202,
      "loss": 2.7286,
      "theoretical_loss": 3.4772673109410075,
      "tokens_seen": 1703870464
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492,
      "loss": 2.7131,
      "theoretical_loss": 3.4772558052380216,
      "tokens_seen": 1703936000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919797979797979,
      "loss": 2.8491,
      "theoretical_loss": 3.4772443001014564,
      "tokens_seen": 1704001536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491959595959596,
      "loss": 2.6865,
      "theoretical_loss": 3.4772327955312625,
      "tokens_seen": 1704067072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491939393939394,
      "loss": 2.4784,
      "theoretical_loss": 3.47722129152739,
      "tokens_seen": 1704132608
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 940087,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.313035726547241,
      "objective/train/theoretical_loss": 3.477209788089789,
      "objective/train/tokens_used": 63057376,
      "theoretical_loss": 3.477209788089789,
      "tokens_seen": 1704198144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491919191919192,
      "loss": 2.5144,
      "theoretical_loss": 3.477209788089789,
      "tokens_seen": 1704198144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918989898989899,
      "loss": 2.4958,
      "theoretical_loss": 3.4771982852184102,
      "tokens_seen": 1704263680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918787878787879,
      "loss": 2.6901,
      "theoretical_loss": 3.4771867829132033,
      "tokens_seen": 1704329216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918585858585859,
      "loss": 2.7283,
      "theoretical_loss": 3.4771752811741194,
      "tokens_seen": 1704394752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918383838383839,
      "loss": 2.5867,
      "theoretical_loss": 3.477163780001109,
      "tokens_seen": 1704460288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918181818181818,
      "loss": 2.6477,
      "theoretical_loss": 3.477152279394122,
      "tokens_seen": 1704525824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917979797979798,
      "loss": 2.6709,
      "theoretical_loss": 3.4771407793531086,
      "tokens_seen": 1704591360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917777777777777,
      "loss": 2.6038,
      "theoretical_loss": 3.4771292798780196,
      "tokens_seen": 1704656896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917575757575758,
      "loss": 2.9226,
      "theoretical_loss": 3.477117780968805,
      "tokens_seen": 1704722432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917373737373737,
      "loss": 2.5123,
      "theoretical_loss": 3.4771062826254155,
      "tokens_seen": 1704787968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917171717171718,
      "loss": 2.5392,
      "theoretical_loss": 3.477094784847802,
      "tokens_seen": 1704853504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916969696969697,
      "loss": 2.8546,
      "theoretical_loss": 3.477083287635914,
      "tokens_seen": 1704919040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916767676767677,
      "loss": 2.5932,
      "theoretical_loss": 3.4770717909897026,
      "tokens_seen": 1704984576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916565656565657,
      "loss": 2.6186,
      "theoretical_loss": 3.4770602949091174,
      "tokens_seen": 1705050112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916363636363637,
      "loss": 2.6519,
      "theoretical_loss": 3.47704879939411,
      "tokens_seen": 1705115648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916161616161616,
      "loss": 2.6116,
      "theoretical_loss": 3.4770373044446297,
      "tokens_seen": 1705181184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915959595959596,
      "loss": 2.7875,
      "theoretical_loss": 3.4770258100606277,
      "tokens_seen": 1705246720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915757575757576,
      "loss": 2.621,
      "theoretical_loss": 3.4770143162420544,
      "tokens_seen": 1705312256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915555555555556,
      "loss": 2.7483,
      "theoretical_loss": 3.47700282298886,
      "tokens_seen": 1705377792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915353535353535,
      "loss": 2.6616,
      "theoretical_loss": 3.4769913303009945,
      "tokens_seen": 1705443328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915151515151515,
      "loss": 2.8028,
      "theoretical_loss": 3.4769798381784094,
      "tokens_seen": 1705508864
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914949494949494,
      "loss": 2.3576,
      "theoretical_loss": 3.4769683466210544,
      "tokens_seen": 1705574400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914747474747475,
      "loss": 2.6636,
      "theoretical_loss": 3.47695685562888,
      "tokens_seen": 1705639936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914545454545455,
      "loss": 2.7262,
      "theoretical_loss": 3.4769453652018374,
      "tokens_seen": 1705705472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914343434343435,
      "loss": 2.6114,
      "theoretical_loss": 3.476933875339877,
      "tokens_seen": 1705771008
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 941279,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1284117698669434,
      "objective/train/theoretical_loss": 3.4769223860429483,
      "objective/train/tokens_used": 64695776,
      "theoretical_loss": 3.4769223860429483,
      "tokens_seen": 1705836544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914141414141414,
      "loss": 2.6038,
      "theoretical_loss": 3.4769223860429483,
      "tokens_seen": 1705836544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913939393939394,
      "loss": 2.5564,
      "theoretical_loss": 3.4769108973110026,
      "tokens_seen": 1705902080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913737373737374,
      "loss": 2.6878,
      "theoretical_loss": 3.47689940914399,
      "tokens_seen": 1705967616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913535353535354,
      "loss": 2.7768,
      "theoretical_loss": 3.476887921541862,
      "tokens_seen": 1706033152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913333333333333,
      "loss": 2.5874,
      "theoretical_loss": 3.4768764345045677,
      "tokens_seen": 1706098688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913131313131313,
      "loss": 2.4851,
      "theoretical_loss": 3.476864948032059,
      "tokens_seen": 1706164224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912929292929293,
      "loss": 2.3652,
      "theoretical_loss": 3.4768534621242853,
      "tokens_seen": 1706229760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912727272727272,
      "loss": 2.3602,
      "theoretical_loss": 3.4768419767811976,
      "tokens_seen": 1706295296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912525252525252,
      "loss": 2.6307,
      "theoretical_loss": 3.4768304920027466,
      "tokens_seen": 1706360832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912323232323233,
      "loss": 2.6657,
      "theoretical_loss": 3.4768190077888823,
      "tokens_seen": 1706426368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912121212121213,
      "loss": 2.5756,
      "theoretical_loss": 3.4768075241395566,
      "tokens_seen": 1706491904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911919191919192,
      "loss": 2.7722,
      "theoretical_loss": 3.4767960410547185,
      "tokens_seen": 1706557440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911717171717172,
      "loss": 2.288,
      "theoretical_loss": 3.4767845585343196,
      "tokens_seen": 1706622976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911515151515152,
      "loss": 2.4128,
      "theoretical_loss": 3.4767730765783096,
      "tokens_seen": 1706688512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911313131313131,
      "loss": 2.4081,
      "theoretical_loss": 3.47676159518664,
      "tokens_seen": 1706754048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911111111111111,
      "loss": 2.6869,
      "theoretical_loss": 3.4767501143592607,
      "tokens_seen": 1706819584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910909090909091,
      "loss": 2.5965,
      "theoretical_loss": 3.4767386340961233,
      "tokens_seen": 1706885120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491070707070707,
      "loss": 2.5827,
      "theoretical_loss": 3.476727154397177,
      "tokens_seen": 1706950656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491050505050505,
      "loss": 2.7387,
      "theoretical_loss": 3.4767156752623736,
      "tokens_seen": 1707016192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491030303030303,
      "loss": 2.7383,
      "theoretical_loss": 3.4767041966916628,
      "tokens_seen": 1707081728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910101010101011,
      "loss": 2.7158,
      "theoretical_loss": 3.476692718684996,
      "tokens_seen": 1707147264
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490989898989899,
      "loss": 2.6993,
      "theoretical_loss": 3.476681241242323,
      "tokens_seen": 1707212800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490969696969697,
      "loss": 2.6435,
      "theoretical_loss": 3.4766697643635958,
      "tokens_seen": 1707278336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490949494949495,
      "loss": 2.6064,
      "theoretical_loss": 3.4766582880487635,
      "tokens_seen": 1707343872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490929292929293,
      "loss": 2.5374,
      "theoretical_loss": 3.4766468122977776,
      "tokens_seen": 1707409408
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 941668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3641209602355957,
      "objective/train/theoretical_loss": 3.4766353371105887,
      "objective/train/tokens_used": 66334176,
      "theoretical_loss": 3.4766353371105887,
      "tokens_seen": 1707474944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909090909090909,
      "loss": 2.6239,
      "theoretical_loss": 3.4766353371105887,
      "tokens_seen": 1707474944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908888888888889,
      "loss": 2.5148,
      "theoretical_loss": 3.4766238624871475,
      "tokens_seen": 1707540480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908686868686868,
      "loss": 2.5122,
      "theoretical_loss": 3.4766123884274043,
      "tokens_seen": 1707606016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908484848484849,
      "loss": 2.9696,
      "theoretical_loss": 3.4766009149313097,
      "tokens_seen": 1707671552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908282828282828,
      "loss": 2.4315,
      "theoretical_loss": 3.476589441998815,
      "tokens_seen": 1707737088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908080808080808,
      "loss": 2.6819,
      "theoretical_loss": 3.4765779696298704,
      "tokens_seen": 1707802624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907878787878787,
      "loss": 2.6707,
      "theoretical_loss": 3.476566497824427,
      "tokens_seen": 1707868160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907676767676768,
      "loss": 2.6873,
      "theoretical_loss": 3.476555026582435,
      "tokens_seen": 1707933696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907474747474748,
      "loss": 2.6117,
      "theoretical_loss": 3.476543555903846,
      "tokens_seen": 1707999232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907272727272728,
      "loss": 2.7386,
      "theoretical_loss": 3.4765320857886097,
      "tokens_seen": 1708064768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907070707070707,
      "loss": 2.7186,
      "theoretical_loss": 3.476520616236677,
      "tokens_seen": 1708130304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906868686868687,
      "loss": 2.6979,
      "theoretical_loss": 3.476509147247999,
      "tokens_seen": 1708195840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906666666666667,
      "loss": 2.8907,
      "theoretical_loss": 3.4764976788225264,
      "tokens_seen": 1708261376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906464646464647,
      "loss": 2.6496,
      "theoretical_loss": 3.4764862109602097,
      "tokens_seen": 1708326912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906262626262626,
      "loss": 2.7096,
      "theoretical_loss": 3.4764747436609995,
      "tokens_seen": 1708392448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906060606060606,
      "loss": 2.7149,
      "theoretical_loss": 3.476463276924847,
      "tokens_seen": 1708457984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905858585858585,
      "loss": 2.8988,
      "theoretical_loss": 3.476451810751703,
      "tokens_seen": 1708523520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905656565656565,
      "loss": 2.5287,
      "theoretical_loss": 3.4764403451415173,
      "tokens_seen": 1708589056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905454545454546,
      "loss": 2.4729,
      "theoretical_loss": 3.476428880094242,
      "tokens_seen": 1708654592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905252525252526,
      "loss": 2.7736,
      "theoretical_loss": 3.476417415609827,
      "tokens_seen": 1708720128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905050505050505,
      "loss": 2.732,
      "theoretical_loss": 3.476405951688223,
      "tokens_seen": 1708785664
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904848484848485,
      "loss": 2.6687,
      "theoretical_loss": 3.476394488329382,
      "tokens_seen": 1708851200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904646464646465,
      "loss": 2.7529,
      "theoretical_loss": 3.476383025533253,
      "tokens_seen": 1708916736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904444444444445,
      "loss": 2.5937,
      "theoretical_loss": 3.4763715632997885,
      "tokens_seen": 1708982272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904242424242424,
      "loss": 2.7642,
      "theoretical_loss": 3.476360101628938,
      "tokens_seen": 1709047808
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 943105,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5795464515686035,
      "objective/train/theoretical_loss": 3.476348640520653,
      "objective/train/tokens_used": 67972576,
      "theoretical_loss": 3.476348640520653,
      "tokens_seen": 1709113344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904040404040404,
      "loss": 2.6005,
      "theoretical_loss": 3.476348640520653,
      "tokens_seen": 1709113344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903838383838384,
      "loss": 2.4409,
      "theoretical_loss": 3.476337179974884,
      "tokens_seen": 1709178880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903636363636364,
      "loss": 2.6039,
      "theoretical_loss": 3.476325719991582,
      "tokens_seen": 1709244416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903434343434343,
      "loss": 2.5914,
      "theoretical_loss": 3.476314260570698,
      "tokens_seen": 1709309952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903232323232323,
      "loss": 2.4368,
      "theoretical_loss": 3.4763028017121824,
      "tokens_seen": 1709375488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903030303030304,
      "loss": 2.7762,
      "theoretical_loss": 3.4762913434159866,
      "tokens_seen": 1709441024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902828282828283,
      "loss": 2.6016,
      "theoretical_loss": 3.476279885682061,
      "tokens_seen": 1709506560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902626262626263,
      "loss": 2.6904,
      "theoretical_loss": 3.4762684285103562,
      "tokens_seen": 1709572096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902424242424243,
      "loss": 2.6499,
      "theoretical_loss": 3.476256971900824,
      "tokens_seen": 1709637632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902222222222222,
      "loss": 2.5542,
      "theoretical_loss": 3.4762455158534147,
      "tokens_seen": 1709703168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902020202020202,
      "loss": 2.8664,
      "theoretical_loss": 3.476234060368079,
      "tokens_seen": 1709768704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901818181818182,
      "loss": 2.7259,
      "theoretical_loss": 3.4762226054447676,
      "tokens_seen": 1709834240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901616161616162,
      "loss": 2.7595,
      "theoretical_loss": 3.476211151083432,
      "tokens_seen": 1709899776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901414141414141,
      "loss": 2.7021,
      "theoretical_loss": 3.4761996972840232,
      "tokens_seen": 1709965312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901212121212121,
      "loss": 2.5646,
      "theoretical_loss": 3.476188244046491,
      "tokens_seen": 1710030848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901010101010101,
      "loss": 2.7032,
      "theoretical_loss": 3.4761767913707877,
      "tokens_seen": 1710096384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490080808080808,
      "loss": 2.5891,
      "theoretical_loss": 3.4761653392568634,
      "tokens_seen": 1710161920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900606060606061,
      "loss": 2.6411,
      "theoretical_loss": 3.4761538877046694,
      "tokens_seen": 1710227456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900404040404041,
      "loss": 2.5804,
      "theoretical_loss": 3.476142436714156,
      "tokens_seen": 1710292992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900202020202021,
      "loss": 2.6421,
      "theoretical_loss": 3.4761309862852747,
      "tokens_seen": 1710358528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049,
      "loss": 2.891,
      "theoretical_loss": 3.4761195364179764,
      "tokens_seen": 1710424064
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489979797979798,
      "loss": 2.6098,
      "theoretical_loss": 3.4761080871122116,
      "tokens_seen": 1710489600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489959595959596,
      "loss": 2.8071,
      "theoretical_loss": 3.476096638367932,
      "tokens_seen": 1710555136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899393939393939,
      "loss": 2.5849,
      "theoretical_loss": 3.4760851901850875,
      "tokens_seen": 1710620672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899191919191919,
      "loss": 2.8232,
      "theoretical_loss": 3.4760737425636297,
      "tokens_seen": 1710686208
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 943708,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.789628744125366,
      "objective/train/theoretical_loss": 3.4760622955035103,
      "objective/train/tokens_used": 69610976,
      "theoretical_loss": 3.4760622955035103,
      "tokens_seen": 1710751744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898989898989899,
      "loss": 2.7948,
      "theoretical_loss": 3.4760622955035103,
      "tokens_seen": 1710751744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898787878787878,
      "loss": 2.7579,
      "theoretical_loss": 3.476050849004679,
      "tokens_seen": 1710817280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898585858585858,
      "loss": 2.8273,
      "theoretical_loss": 3.4760394030670874,
      "tokens_seen": 1710882816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898383838383839,
      "loss": 2.5987,
      "theoretical_loss": 3.476027957690686,
      "tokens_seen": 1710948352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898181818181819,
      "loss": 2.6575,
      "theoretical_loss": 3.476016512875426,
      "tokens_seen": 1711013888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897979797979798,
      "loss": 2.792,
      "theoretical_loss": 3.4760050686212596,
      "tokens_seen": 1711079424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897777777777778,
      "loss": 2.703,
      "theoretical_loss": 3.475993624928136,
      "tokens_seen": 1711144960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897575757575758,
      "loss": 2.6028,
      "theoretical_loss": 3.475982181796007,
      "tokens_seen": 1711210496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897373737373738,
      "loss": 2.8526,
      "theoretical_loss": 3.475970739224824,
      "tokens_seen": 1711276032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897171717171717,
      "loss": 2.5888,
      "theoretical_loss": 3.475959297214537,
      "tokens_seen": 1711341568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896969696969697,
      "loss": 2.7614,
      "theoretical_loss": 3.475947855765098,
      "tokens_seen": 1711407104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896767676767676,
      "loss": 2.723,
      "theoretical_loss": 3.4759364148764575,
      "tokens_seen": 1711472640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896565656565657,
      "loss": 2.5174,
      "theoretical_loss": 3.4759249745485667,
      "tokens_seen": 1711538176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896363636363636,
      "loss": 2.6374,
      "theoretical_loss": 3.4759135347813768,
      "tokens_seen": 1711603712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896161616161617,
      "loss": 2.7387,
      "theoretical_loss": 3.4759020955748388,
      "tokens_seen": 1711669248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895959595959596,
      "loss": 2.6615,
      "theoretical_loss": 3.4758906569289034,
      "tokens_seen": 1711734784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895757575757576,
      "loss": 2.6527,
      "theoretical_loss": 3.475879218843522,
      "tokens_seen": 1711800320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895555555555556,
      "loss": 2.6785,
      "theoretical_loss": 3.475867781318646,
      "tokens_seen": 1711865856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895353535353536,
      "loss": 2.4673,
      "theoretical_loss": 3.4758563443542254,
      "tokens_seen": 1711931392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895151515151515,
      "loss": 2.6067,
      "theoretical_loss": 3.4758449079502123,
      "tokens_seen": 1711996928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894949494949495,
      "loss": 2.6138,
      "theoretical_loss": 3.475833472106557,
      "tokens_seen": 1712062464
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894747474747475,
      "loss": 2.6808,
      "theoretical_loss": 3.475822036823212,
      "tokens_seen": 1712128000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894545454545455,
      "loss": 2.6603,
      "theoretical_loss": 3.4758106021001267,
      "tokens_seen": 1712193536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894343434343434,
      "loss": 2.4896,
      "theoretical_loss": 3.4757991679372533,
      "tokens_seen": 1712259072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894141414141414,
      "loss": 2.5848,
      "theoretical_loss": 3.4757877343345425,
      "tokens_seen": 1712324608
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 944906,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.562922239303589,
      "objective/train/theoretical_loss": 3.4757763012919454,
      "objective/train/tokens_used": 71249376,
      "theoretical_loss": 3.4757763012919454,
      "tokens_seen": 1712390144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893939393939393,
      "loss": 2.6692,
      "theoretical_loss": 3.4757763012919454,
      "tokens_seen": 1712390144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893737373737374,
      "loss": 2.848,
      "theoretical_loss": 3.475764868809413,
      "tokens_seen": 1712455680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893535353535354,
      "loss": 2.437,
      "theoretical_loss": 3.4757534368868974,
      "tokens_seen": 1712521216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893333333333334,
      "loss": 2.7263,
      "theoretical_loss": 3.4757420055243484,
      "tokens_seen": 1712586752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893131313131313,
      "loss": 2.6201,
      "theoretical_loss": 3.4757305747217178,
      "tokens_seen": 1712652288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892929292929293,
      "loss": 2.774,
      "theoretical_loss": 3.4757191444789566,
      "tokens_seen": 1712717824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892727272727273,
      "loss": 2.8183,
      "theoretical_loss": 3.475707714796016,
      "tokens_seen": 1712783360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892525252525253,
      "loss": 2.6004,
      "theoretical_loss": 3.4756962856728473,
      "tokens_seen": 1712848896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892323232323232,
      "loss": 2.8133,
      "theoretical_loss": 3.4756848571094014,
      "tokens_seen": 1712914432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892121212121212,
      "loss": 2.8219,
      "theoretical_loss": 3.4756734291056297,
      "tokens_seen": 1712979968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891919191919192,
      "loss": 2.826,
      "theoretical_loss": 3.475662001661483,
      "tokens_seen": 1713045504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891717171717171,
      "loss": 2.6803,
      "theoretical_loss": 3.475650574776914,
      "tokens_seen": 1713111040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891515151515151,
      "loss": 2.7393,
      "theoretical_loss": 3.4756391484518714,
      "tokens_seen": 1713176576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891313131313132,
      "loss": 2.5757,
      "theoretical_loss": 3.4756277226863084,
      "tokens_seen": 1713242112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891111111111112,
      "loss": 2.7433,
      "theoretical_loss": 3.475616297480175,
      "tokens_seen": 1713307648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890909090909091,
      "loss": 2.5726,
      "theoretical_loss": 3.475604872833424,
      "tokens_seen": 1713373184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890707070707071,
      "loss": 2.7851,
      "theoretical_loss": 3.4755934487460047,
      "tokens_seen": 1713438720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890505050505051,
      "loss": 2.6237,
      "theoretical_loss": 3.475582025217869,
      "tokens_seen": 1713504256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489030303030303,
      "loss": 2.8517,
      "theoretical_loss": 3.4755706022489687,
      "tokens_seen": 1713569792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489010101010101,
      "loss": 2.4496,
      "theoretical_loss": 3.475559179839254,
      "tokens_seen": 1713635328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488989898989899,
      "loss": 2.4707,
      "theoretical_loss": 3.4755477579886778,
      "tokens_seen": 1713700864
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889696969696969,
      "loss": 2.7956,
      "theoretical_loss": 3.4755363366971896,
      "tokens_seen": 1713766400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889494949494949,
      "loss": 2.6349,
      "theoretical_loss": 3.4755249159647414,
      "tokens_seen": 1713831936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889292929292929,
      "loss": 2.6285,
      "theoretical_loss": 3.4755134957912848,
      "tokens_seen": 1713897472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488909090909091,
      "loss": 2.7714,
      "theoretical_loss": 3.4755020761767703,
      "tokens_seen": 1713963008
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 945396,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1092090606689453,
      "objective/train/theoretical_loss": 3.47549065712115,
      "objective/train/tokens_used": 72887776,
      "theoretical_loss": 3.47549065712115,
      "tokens_seen": 1714028544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888888888888889,
      "loss": 2.4456,
      "theoretical_loss": 3.47549065712115,
      "tokens_seen": 1714028544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888686868686869,
      "loss": 2.5141,
      "theoretical_loss": 3.4754792386243745,
      "tokens_seen": 1714094080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888484848484849,
      "loss": 2.7276,
      "theoretical_loss": 3.4754678206863954,
      "tokens_seen": 1714159616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888282828282829,
      "loss": 2.5177,
      "theoretical_loss": 3.475456403307164,
      "tokens_seen": 1714225152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888080808080808,
      "loss": 2.7962,
      "theoretical_loss": 3.4754449864866315,
      "tokens_seen": 1714290688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887878787878788,
      "loss": 2.5457,
      "theoretical_loss": 3.475433570224749,
      "tokens_seen": 1714356224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887676767676768,
      "loss": 2.6354,
      "theoretical_loss": 3.475422154521468,
      "tokens_seen": 1714421760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887474747474748,
      "loss": 2.6071,
      "theoretical_loss": 3.47541073937674,
      "tokens_seen": 1714487296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887272727272727,
      "loss": 2.5188,
      "theoretical_loss": 3.475399324790516,
      "tokens_seen": 1714552832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887070707070707,
      "loss": 2.874,
      "theoretical_loss": 3.4753879107627483,
      "tokens_seen": 1714618368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886868686868686,
      "loss": 2.7343,
      "theoretical_loss": 3.4753764972933867,
      "tokens_seen": 1714683904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886666666666667,
      "loss": 2.8105,
      "theoretical_loss": 3.4753650843823833,
      "tokens_seen": 1714749440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886464646464647,
      "loss": 2.8456,
      "theoretical_loss": 3.47535367202969,
      "tokens_seen": 1714814976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886262626262627,
      "loss": 2.5544,
      "theoretical_loss": 3.475342260235257,
      "tokens_seen": 1714880512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886060606060606,
      "loss": 2.7038,
      "theoretical_loss": 3.475330848999036,
      "tokens_seen": 1714946048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885858585858586,
      "loss": 2.4342,
      "theoretical_loss": 3.4753194383209793,
      "tokens_seen": 1715011584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885656565656566,
      "loss": 2.8637,
      "theoretical_loss": 3.475308028201037,
      "tokens_seen": 1715077120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885454545454546,
      "loss": 2.8579,
      "theoretical_loss": 3.4752966186391614,
      "tokens_seen": 1715142656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885252525252525,
      "loss": 2.7374,
      "theoretical_loss": 3.4752852096353033,
      "tokens_seen": 1715208192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885050505050505,
      "loss": 2.6466,
      "theoretical_loss": 3.4752738011894144,
      "tokens_seen": 1715273728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884848484848484,
      "loss": 2.7546,
      "theoretical_loss": 3.4752623933014464,
      "tokens_seen": 1715339264
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884646464646465,
      "loss": 2.5621,
      "theoretical_loss": 3.47525098597135,
      "tokens_seen": 1715404800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884444444444445,
      "loss": 2.5583,
      "theoretical_loss": 3.475239579199077,
      "tokens_seen": 1715470336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884242424242425,
      "loss": 2.6457,
      "theoretical_loss": 3.4752281729845786,
      "tokens_seen": 1715535872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884040404040404,
      "loss": 2.7425,
      "theoretical_loss": 3.4752167673278063,
      "tokens_seen": 1715601408
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 946450,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.578782796859741,
      "objective/train/theoretical_loss": 3.475205362228712,
      "objective/train/tokens_used": 74526176,
      "theoretical_loss": 3.475205362228712,
      "tokens_seen": 1715666944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883838383838384,
      "loss": 2.544,
      "theoretical_loss": 3.475205362228712,
      "tokens_seen": 1715666944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883636363636364,
      "loss": 2.5762,
      "theoretical_loss": 3.475193957687246,
      "tokens_seen": 1715732480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883434343434344,
      "loss": 2.6404,
      "theoretical_loss": 3.4751825537033607,
      "tokens_seen": 1715798016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883232323232323,
      "loss": 2.5441,
      "theoretical_loss": 3.4751711502770077,
      "tokens_seen": 1715863552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883030303030303,
      "loss": 2.6666,
      "theoretical_loss": 3.4751597474081377,
      "tokens_seen": 1715929088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882828282828283,
      "loss": 2.4691,
      "theoretical_loss": 3.4751483450967027,
      "tokens_seen": 1715994624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882626262626263,
      "loss": 2.8491,
      "theoretical_loss": 3.4751369433426538,
      "tokens_seen": 1716060160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882424242424243,
      "loss": 2.6214,
      "theoretical_loss": 3.4751255421459426,
      "tokens_seen": 1716125696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048822222222222225,
      "loss": 2.7773,
      "theoretical_loss": 3.4751141415065208,
      "tokens_seen": 1716191232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882020202020202,
      "loss": 2.4918,
      "theoretical_loss": 3.4751027414243394,
      "tokens_seen": 1716256768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048818181818181816,
      "loss": 2.5795,
      "theoretical_loss": 3.4750913418993505,
      "tokens_seen": 1716322304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048816161616161617,
      "loss": 2.5246,
      "theoretical_loss": 3.4750799429315054,
      "tokens_seen": 1716387840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881414141414142,
      "loss": 2.514,
      "theoretical_loss": 3.475068544520755,
      "tokens_seen": 1716453376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048812121212121214,
      "loss": 2.639,
      "theoretical_loss": 3.4750571466670515,
      "tokens_seen": 1716518912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881010101010101,
      "loss": 2.6768,
      "theoretical_loss": 3.475045749370346,
      "tokens_seen": 1716584448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048808080808080805,
      "loss": 2.7541,
      "theoretical_loss": 3.4750343526305905,
      "tokens_seen": 1716649984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880606060606061,
      "loss": 2.5554,
      "theoretical_loss": 3.475022956447736,
      "tokens_seen": 1716715520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880404040404041,
      "loss": 2.7061,
      "theoretical_loss": 3.4750115608217342,
      "tokens_seen": 1716781056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048802020202020203,
      "loss": 2.5479,
      "theoretical_loss": 3.4750001657525367,
      "tokens_seen": 1716846592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488,
      "loss": 2.5462,
      "theoretical_loss": 3.4749887712400955,
      "tokens_seen": 1716912128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048797979797979795,
      "loss": 2.4832,
      "theoretical_loss": 3.474977377284361,
      "tokens_seen": 1716977664
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000487959595959596,
      "loss": 2.7024,
      "theoretical_loss": 3.4749659838852858,
      "tokens_seen": 1717043200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048793939393939397,
      "loss": 2.841,
      "theoretical_loss": 3.4749545910428212,
      "tokens_seen": 1717108736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879191919191919,
      "loss": 2.589,
      "theoretical_loss": 3.474943198756918,
      "tokens_seen": 1717174272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878989898989899,
      "loss": 2.5626,
      "theoretical_loss": 3.474931807027529,
      "tokens_seen": 1717239808
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 947071,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.492351531982422,
      "objective/train/theoretical_loss": 3.474920415854605,
      "objective/train/tokens_used": 76164576,
      "theoretical_loss": 3.474920415854605,
      "tokens_seen": 1717305344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048787878787878784,
      "loss": 2.7781,
      "theoretical_loss": 3.474920415854605,
      "tokens_seen": 1717305344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878585858585859,
      "loss": 2.6613,
      "theoretical_loss": 3.474909025238098,
      "tokens_seen": 1717370880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048783838383838386,
      "loss": 2.6513,
      "theoretical_loss": 3.474897635177959,
      "tokens_seen": 1717436416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878181818181818,
      "loss": 2.5682,
      "theoretical_loss": 3.4748862456741403,
      "tokens_seen": 1717501952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877979797979798,
      "loss": 2.7969,
      "theoretical_loss": 3.4748748567265926,
      "tokens_seen": 1717567488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877777777777778,
      "loss": 2.7713,
      "theoretical_loss": 3.474863468335269,
      "tokens_seen": 1717633024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877575757575758,
      "loss": 2.7294,
      "theoretical_loss": 3.4748520805001193,
      "tokens_seen": 1717698560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048773737373737375,
      "loss": 2.4451,
      "theoretical_loss": 3.474840693221096,
      "tokens_seen": 1717764096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877171717171717,
      "loss": 2.5473,
      "theoretical_loss": 3.4748293064981515,
      "tokens_seen": 1717829632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048769696969696967,
      "loss": 2.6743,
      "theoretical_loss": 3.474817920331236,
      "tokens_seen": 1717895168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876767676767677,
      "loss": 2.4018,
      "theoretical_loss": 3.4748065347203023,
      "tokens_seen": 1717960704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876565656565657,
      "loss": 2.5568,
      "theoretical_loss": 3.474795149665301,
      "tokens_seen": 1718026240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048763636363636365,
      "loss": 2.4997,
      "theoretical_loss": 3.474783765166185,
      "tokens_seen": 1718091776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876161616161616,
      "loss": 2.5956,
      "theoretical_loss": 3.4747723812229045,
      "tokens_seen": 1718157312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875959595959596,
      "loss": 2.4705,
      "theoretical_loss": 3.4747609978354124,
      "tokens_seen": 1718222848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048757575757575757,
      "loss": 2.8133,
      "theoretical_loss": 3.4747496150036596,
      "tokens_seen": 1718288384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875555555555556,
      "loss": 2.6255,
      "theoretical_loss": 3.4747382327275984,
      "tokens_seen": 1718353920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048753535353535354,
      "loss": 2.3969,
      "theoretical_loss": 3.47472685100718,
      "tokens_seen": 1718419456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875151515151515,
      "loss": 2.8356,
      "theoretical_loss": 3.474715469842356,
      "tokens_seen": 1718484992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874949494949495,
      "loss": 2.4337,
      "theoretical_loss": 3.4747040892330787,
      "tokens_seen": 1718550528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048747474747474746,
      "loss": 2.5269,
      "theoretical_loss": 3.474692709179299,
      "tokens_seen": 1718616064
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874545454545455,
      "loss": 2.7374,
      "theoretical_loss": 3.474681329680969,
      "tokens_seen": 1718681600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048743434343434343,
      "loss": 2.7276,
      "theoretical_loss": 3.4746699507380407,
      "tokens_seen": 1718747136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048741414141414144,
      "loss": 2.5347,
      "theoretical_loss": 3.4746585723504655,
      "tokens_seen": 1718812672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004873939393939394,
      "loss": 2.6824,
      "theoretical_loss": 3.4746471945181954,
      "tokens_seen": 1718878208
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 948336,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.850125789642334,
      "objective/train/theoretical_loss": 3.4746358172411815,
      "objective/train/tokens_used": 77802976,
      "theoretical_loss": 3.4746358172411815,
      "tokens_seen": 1718943744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048737373737373736,
      "loss": 2.6919,
      "theoretical_loss": 3.4746358172411815,
      "tokens_seen": 1718943744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048735353535353537,
      "loss": 2.537,
      "theoretical_loss": 3.474624440519376,
      "tokens_seen": 1719009280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873333333333333,
      "loss": 2.7071,
      "theoretical_loss": 3.4746130643527304,
      "tokens_seen": 1719074816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048731313131313134,
      "loss": 2.3608,
      "theoretical_loss": 3.474601688741197,
      "tokens_seen": 1719140352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872929292929293,
      "loss": 2.6354,
      "theoretical_loss": 3.474590313684727,
      "tokens_seen": 1719205888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048727272727272725,
      "loss": 2.7612,
      "theoretical_loss": 3.474578939183272,
      "tokens_seen": 1719271424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048725252525252526,
      "loss": 2.561,
      "theoretical_loss": 3.474567565236785,
      "tokens_seen": 1719336960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048723232323232327,
      "loss": 2.7526,
      "theoretical_loss": 3.474556191845216,
      "tokens_seen": 1719402496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048721212121212123,
      "loss": 2.8614,
      "theoretical_loss": 3.4745448190085177,
      "tokens_seen": 1719468032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871919191919192,
      "loss": 2.3397,
      "theoretical_loss": 3.4745334467266424,
      "tokens_seen": 1719533568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048717171717171714,
      "loss": 2.4128,
      "theoretical_loss": 3.4745220749995407,
      "tokens_seen": 1719599104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048715151515151515,
      "loss": 2.5001,
      "theoretical_loss": 3.474510703827165,
      "tokens_seen": 1719664640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048713131313131316,
      "loss": 2.8698,
      "theoretical_loss": 3.474499333209468,
      "tokens_seen": 1719730176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871111111111111,
      "loss": 2.8356,
      "theoretical_loss": 3.4744879631464,
      "tokens_seen": 1719795712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870909090909091,
      "loss": 2.5725,
      "theoretical_loss": 3.474476593637913,
      "tokens_seen": 1719861248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048707070707070704,
      "loss": 2.4696,
      "theoretical_loss": 3.47446522468396,
      "tokens_seen": 1719926784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870505050505051,
      "loss": 2.5063,
      "theoretical_loss": 3.4744538562844913,
      "tokens_seen": 1719992320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048703030303030306,
      "loss": 2.478,
      "theoretical_loss": 3.4744424884394602,
      "tokens_seen": 1720057856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487010101010101,
      "loss": 2.609,
      "theoretical_loss": 3.4744311211488172,
      "tokens_seen": 1720123392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048698989898989897,
      "loss": 2.8236,
      "theoretical_loss": 3.4744197544125153,
      "tokens_seen": 1720188928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486969696969697,
      "loss": 2.7664,
      "theoretical_loss": 3.4744083882305055,
      "tokens_seen": 1720254464
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486949494949495,
      "loss": 2.6229,
      "theoretical_loss": 3.4743970226027403,
      "tokens_seen": 1720320000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048692929292929295,
      "loss": 2.8784,
      "theoretical_loss": 3.474385657529171,
      "tokens_seen": 1720385536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869090909090909,
      "loss": 2.6972,
      "theoretical_loss": 3.47437429300975,
      "tokens_seen": 1720451072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048688888888888886,
      "loss": 2.9794,
      "theoretical_loss": 3.4743629290444282,
      "tokens_seen": 1720516608
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 949353,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0498993396759033,
      "objective/train/theoretical_loss": 3.4743515656331594,
      "objective/train/tokens_used": 79441376,
      "theoretical_loss": 3.4743515656331594,
      "tokens_seen": 1720582144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048686868686868693,
      "loss": 2.8742,
      "theoretical_loss": 3.4743515656331594,
      "tokens_seen": 1720582144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868484848484849,
      "loss": 2.6368,
      "theoretical_loss": 3.474340202775893,
      "tokens_seen": 1720647680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048682828282828284,
      "loss": 2.6623,
      "theoretical_loss": 3.474328840472583,
      "tokens_seen": 1720713216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868080808080808,
      "loss": 2.5302,
      "theoretical_loss": 3.47431747872318,
      "tokens_seen": 1720778752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048678787878787876,
      "loss": 2.827,
      "theoretical_loss": 3.4743061175276364,
      "tokens_seen": 1720844288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867676767676768,
      "loss": 2.6887,
      "theoretical_loss": 3.474294756885904,
      "tokens_seen": 1720909824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867474747474748,
      "loss": 2.6388,
      "theoretical_loss": 3.4742833967979347,
      "tokens_seen": 1720975360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048672727272727274,
      "loss": 2.5768,
      "theoretical_loss": 3.474272037263681,
      "tokens_seen": 1721040896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867070707070707,
      "loss": 2.5206,
      "theoretical_loss": 3.474260678283094,
      "tokens_seen": 1721106432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866868686868687,
      "loss": 2.7724,
      "theoretical_loss": 3.474249319856126,
      "tokens_seen": 1721171968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866666666666667,
      "loss": 2.8046,
      "theoretical_loss": 3.4742379619827286,
      "tokens_seen": 1721237504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048664646464646467,
      "loss": 2.6154,
      "theoretical_loss": 3.4742266046628543,
      "tokens_seen": 1721303040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048662626262626263,
      "loss": 2.7199,
      "theoretical_loss": 3.4742152478964545,
      "tokens_seen": 1721368576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866060606060606,
      "loss": 2.5633,
      "theoretical_loss": 3.474203891683482,
      "tokens_seen": 1721434112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865858585858586,
      "loss": 2.7631,
      "theoretical_loss": 3.474192536023888,
      "tokens_seen": 1721499648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865656565656566,
      "loss": 2.5908,
      "theoretical_loss": 3.474181180917624,
      "tokens_seen": 1721565184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048654545454545456,
      "loss": 2.601,
      "theoretical_loss": 3.474169826364643,
      "tokens_seen": 1721630720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865252525252525,
      "loss": 2.8114,
      "theoretical_loss": 3.4741584723648975,
      "tokens_seen": 1721696256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048650505050505053,
      "loss": 2.7601,
      "theoretical_loss": 3.4741471189183377,
      "tokens_seen": 1721761792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864848484848485,
      "loss": 2.546,
      "theoretical_loss": 3.474135766024917,
      "tokens_seen": 1721827328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864646464646465,
      "loss": 2.4717,
      "theoretical_loss": 3.4741244136845864,
      "tokens_seen": 1721892864
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048644444444444446,
      "loss": 2.68,
      "theoretical_loss": 3.474113061897299,
      "tokens_seen": 1721958400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864242424242424,
      "loss": 2.5694,
      "theoretical_loss": 3.474101710663006,
      "tokens_seen": 1722023936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864040404040404,
      "loss": 2.6604,
      "theoretical_loss": 3.474090359981659,
      "tokens_seen": 1722089472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863838383838384,
      "loss": 2.6332,
      "theoretical_loss": 3.4740790098532113,
      "tokens_seen": 1722155008
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 949989,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.493234634399414,
      "objective/train/theoretical_loss": 3.4740676602776146,
      "objective/train/tokens_used": 81079776,
      "theoretical_loss": 3.4740676602776146,
      "tokens_seen": 1722220544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863636363636364,
      "loss": 2.5725,
      "theoretical_loss": 3.4740676602776146,
      "tokens_seen": 1722220544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048634343434343435,
      "loss": 2.5757,
      "theoretical_loss": 3.47405631125482,
      "tokens_seen": 1722286080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048632323232323236,
      "loss": 2.6726,
      "theoretical_loss": 3.4740449627847805,
      "tokens_seen": 1722351616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863030303030303,
      "loss": 2.5449,
      "theoretical_loss": 3.4740336148674476,
      "tokens_seen": 1722417152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862828282828283,
      "loss": 2.7329,
      "theoretical_loss": 3.4740222675027734,
      "tokens_seen": 1722482688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862626262626263,
      "loss": 2.5414,
      "theoretical_loss": 3.474010920690711,
      "tokens_seen": 1722548224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048624242424242424,
      "loss": 2.5322,
      "theoretical_loss": 3.4739995744312107,
      "tokens_seen": 1722613760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048622222222222225,
      "loss": 2.6438,
      "theoretical_loss": 3.4739882287242256,
      "tokens_seen": 1722679296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862020202020202,
      "loss": 2.7798,
      "theoretical_loss": 3.473976883569708,
      "tokens_seen": 1722744832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048618181818181817,
      "loss": 2.6148,
      "theoretical_loss": 3.47396553896761,
      "tokens_seen": 1722810368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861616161616162,
      "loss": 2.7383,
      "theoretical_loss": 3.4739541949178827,
      "tokens_seen": 1722875904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861414141414142,
      "loss": 2.5262,
      "theoretical_loss": 3.4739428514204787,
      "tokens_seen": 1722941440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048612121212121215,
      "loss": 2.6558,
      "theoretical_loss": 3.47393150847535,
      "tokens_seen": 1723006976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861010101010101,
      "loss": 2.7145,
      "theoretical_loss": 3.4739201660824497,
      "tokens_seen": 1723072512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048608080808080806,
      "loss": 2.8424,
      "theoretical_loss": 3.473908824241729,
      "tokens_seen": 1723138048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048606060606060607,
      "loss": 2.4815,
      "theoretical_loss": 3.4738974829531397,
      "tokens_seen": 1723203584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860404040404041,
      "loss": 2.5685,
      "theoretical_loss": 3.4738861422166343,
      "tokens_seen": 1723269120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048602020202020204,
      "loss": 2.5099,
      "theoretical_loss": 3.4738748020321655,
      "tokens_seen": 1723334656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486,
      "loss": 2.5624,
      "theoretical_loss": 3.4738634623996845,
      "tokens_seen": 1723400192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048597979797979795,
      "loss": 2.8233,
      "theoretical_loss": 3.473852123319144,
      "tokens_seen": 1723465728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048595959595959596,
      "loss": 2.6669,
      "theoretical_loss": 3.4738407847904966,
      "tokens_seen": 1723531264
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485939393939394,
      "loss": 2.6042,
      "theoretical_loss": 3.4738294468136934,
      "tokens_seen": 1723596800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048591919191919193,
      "loss": 2.7076,
      "theoretical_loss": 3.473818109388687,
      "tokens_seen": 1723662336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858989898989899,
      "loss": 2.6665,
      "theoretical_loss": 3.47380677251543,
      "tokens_seen": 1723727872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048587878787878784,
      "loss": 2.7535,
      "theoretical_loss": 3.473795436193874,
      "tokens_seen": 1723793408
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 951217,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.653082847595215,
      "objective/train/theoretical_loss": 3.473784100423971,
      "objective/train/tokens_used": 82718176,
      "theoretical_loss": 3.473784100423971,
      "tokens_seen": 1723858944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858585858585859,
      "loss": 2.8862,
      "theoretical_loss": 3.473784100423971,
      "tokens_seen": 1723858944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048583838383838387,
      "loss": 2.6551,
      "theoretical_loss": 3.4737727652056742,
      "tokens_seen": 1723924480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858181818181818,
      "loss": 2.5733,
      "theoretical_loss": 3.4737614305389344,
      "tokens_seen": 1723990016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857979797979798,
      "loss": 2.7681,
      "theoretical_loss": 3.473750096423705,
      "tokens_seen": 1724055552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048577777777777774,
      "loss": 2.5725,
      "theoretical_loss": 3.473738762859938,
      "tokens_seen": 1724121088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857575757575758,
      "loss": 2.7263,
      "theoretical_loss": 3.4737274298475844,
      "tokens_seen": 1724186624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048573737373737376,
      "loss": 2.6403,
      "theoretical_loss": 3.4737160973865984,
      "tokens_seen": 1724252160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857171717171717,
      "loss": 2.5917,
      "theoretical_loss": 3.4737047654769304,
      "tokens_seen": 1724317696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856969696969697,
      "loss": 2.7386,
      "theoretical_loss": 3.473693434118534,
      "tokens_seen": 1724383232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856767676767677,
      "loss": 2.7532,
      "theoretical_loss": 3.4736821033113605,
      "tokens_seen": 1724448768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856565656565657,
      "loss": 2.6684,
      "theoretical_loss": 3.4736707730553626,
      "tokens_seen": 1724514304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048563636363636365,
      "loss": 2.4919,
      "theoretical_loss": 3.473659443350493,
      "tokens_seen": 1724579840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856161616161616,
      "loss": 2.739,
      "theoretical_loss": 3.4736481141967026,
      "tokens_seen": 1724645376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048559595959595957,
      "loss": 2.8205,
      "theoretical_loss": 3.4736367855939445,
      "tokens_seen": 1724710912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855757575757576,
      "loss": 2.8129,
      "theoretical_loss": 3.4736254575421706,
      "tokens_seen": 1724776448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855555555555556,
      "loss": 2.5104,
      "theoretical_loss": 3.4736141300413337,
      "tokens_seen": 1724841984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048553535353535355,
      "loss": 2.5611,
      "theoretical_loss": 3.4736028030913864,
      "tokens_seen": 1724907520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855151515151515,
      "loss": 2.564,
      "theoretical_loss": 3.47359147669228,
      "tokens_seen": 1724973056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854949494949495,
      "loss": 2.6007,
      "theoretical_loss": 3.473580150843967,
      "tokens_seen": 1725038592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048547474747474747,
      "loss": 2.7037,
      "theoretical_loss": 3.4735688255464003,
      "tokens_seen": 1725104128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854545454545455,
      "loss": 2.6524,
      "theoretical_loss": 3.4735575007995316,
      "tokens_seen": 1725169664
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048543434343434344,
      "loss": 2.6171,
      "theoretical_loss": 3.4735461766033136,
      "tokens_seen": 1725235200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854141414141414,
      "loss": 2.682,
      "theoretical_loss": 3.473534852957698,
      "tokens_seen": 1725300736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853939393939394,
      "loss": 2.7294,
      "theoretical_loss": 3.4735235298626375,
      "tokens_seen": 1725366272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048537373737373736,
      "loss": 2.5625,
      "theoretical_loss": 3.473512207318085,
      "tokens_seen": 1725431808
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 951846,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9040403366088867,
      "objective/train/theoretical_loss": 3.473500885323992,
      "objective/train/tokens_used": 84356576,
      "theoretical_loss": 3.473500885323992,
      "tokens_seen": 1725497344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853535353535354,
      "loss": 2.6504,
      "theoretical_loss": 3.473500885323992,
      "tokens_seen": 1725497344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048533333333333333,
      "loss": 2.8612,
      "theoretical_loss": 3.473489563880311,
      "tokens_seen": 1725562880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048531313131313134,
      "loss": 2.747,
      "theoretical_loss": 3.473478242986994,
      "tokens_seen": 1725628416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852929292929293,
      "loss": 2.6916,
      "theoretical_loss": 3.4734669226439943,
      "tokens_seen": 1725693952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852727272727273,
      "loss": 2.5991,
      "theoretical_loss": 3.473455602851264,
      "tokens_seen": 1725759488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048525252525252527,
      "loss": 2.5984,
      "theoretical_loss": 3.473444283608755,
      "tokens_seen": 1725825024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852323232323232,
      "loss": 2.772,
      "theoretical_loss": 3.4734329649164195,
      "tokens_seen": 1725890560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048521212121212123,
      "loss": 2.6357,
      "theoretical_loss": 3.4734216467742103,
      "tokens_seen": 1725956096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851919191919192,
      "loss": 2.6213,
      "theoretical_loss": 3.47341032918208,
      "tokens_seen": 1726021632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851717171717172,
      "loss": 2.8269,
      "theoretical_loss": 3.4733990121399807,
      "tokens_seen": 1726087168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048515151515151516,
      "loss": 2.6231,
      "theoretical_loss": 3.4733876956478644,
      "tokens_seen": 1726152704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048513131313131317,
      "loss": 2.6762,
      "theoretical_loss": 3.4733763797056842,
      "tokens_seen": 1726218240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851111111111111,
      "loss": 2.7135,
      "theoretical_loss": 3.473365064313392,
      "tokens_seen": 1726283776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850909090909091,
      "loss": 2.7963,
      "theoretical_loss": 3.4733537494709408,
      "tokens_seen": 1726349312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850707070707071,
      "loss": 2.7273,
      "theoretical_loss": 3.4733424351782816,
      "tokens_seen": 1726414848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048505050505050505,
      "loss": 2.9032,
      "theoretical_loss": 3.4733311214353684,
      "tokens_seen": 1726480384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048503030303030306,
      "loss": 2.7651,
      "theoretical_loss": 3.4733198082421533,
      "tokens_seen": 1726545920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485010101010101,
      "loss": 2.5911,
      "theoretical_loss": 3.473308495598588,
      "tokens_seen": 1726611456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484989898989899,
      "loss": 2.6404,
      "theoretical_loss": 3.4732971835046254,
      "tokens_seen": 1726676992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484969696969697,
      "loss": 2.6003,
      "theoretical_loss": 3.4732858719602175,
      "tokens_seen": 1726742528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484949494949495,
      "loss": 2.6377,
      "theoretical_loss": 3.4732745609653177,
      "tokens_seen": 1726808064
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048492929292929296,
      "loss": 2.6006,
      "theoretical_loss": 3.473263250519878,
      "tokens_seen": 1726873600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849090909090909,
      "loss": 2.6061,
      "theoretical_loss": 3.4732519406238507,
      "tokens_seen": 1726939136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048488888888888887,
      "loss": 2.6086,
      "theoretical_loss": 3.4732406312771884,
      "tokens_seen": 1727004672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848686868686869,
      "loss": 2.8007,
      "theoretical_loss": 3.473229322479843,
      "tokens_seen": 1727070208
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 953107,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.85353422164917,
      "objective/train/theoretical_loss": 3.473218014231768,
      "objective/train/tokens_used": 85994976,
      "theoretical_loss": 3.473218014231768,
      "tokens_seen": 1727135744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848484848484849,
      "loss": 2.6263,
      "theoretical_loss": 3.473218014231768,
      "tokens_seen": 1727135744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048482828282828285,
      "loss": 2.5669,
      "theoretical_loss": 3.473206706532915,
      "tokens_seen": 1727201280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848080808080808,
      "loss": 2.4935,
      "theoretical_loss": 3.473195399383237,
      "tokens_seen": 1727266816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048478787878787876,
      "loss": 2.5743,
      "theoretical_loss": 3.4731840927826867,
      "tokens_seen": 1727332352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048476767676767683,
      "loss": 2.5745,
      "theoretical_loss": 3.473172786731216,
      "tokens_seen": 1727397888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847474747474748,
      "loss": 2.8035,
      "theoretical_loss": 3.4731614812287774,
      "tokens_seen": 1727463424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048472727272727274,
      "loss": 2.7708,
      "theoretical_loss": 3.4731501762753236,
      "tokens_seen": 1727528960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847070707070707,
      "loss": 2.5373,
      "theoretical_loss": 3.4731388718708076,
      "tokens_seen": 1727594496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048468686868686865,
      "loss": 2.4428,
      "theoretical_loss": 3.4731275680151814,
      "tokens_seen": 1727660032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846666666666667,
      "loss": 2.7221,
      "theoretical_loss": 3.4731162647083975,
      "tokens_seen": 1727725568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846464646464647,
      "loss": 2.8533,
      "theoretical_loss": 3.4731049619504084,
      "tokens_seen": 1727791104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048462626262626263,
      "loss": 2.4632,
      "theoretical_loss": 3.473093659741167,
      "tokens_seen": 1727856640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846060606060606,
      "loss": 2.6879,
      "theoretical_loss": 3.4730823580806254,
      "tokens_seen": 1727922176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048458585858585855,
      "loss": 2.7259,
      "theoretical_loss": 3.473071056968737,
      "tokens_seen": 1727987712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845656565656566,
      "loss": 2.8179,
      "theoretical_loss": 3.473059756405453,
      "tokens_seen": 1728053248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048454545454545457,
      "loss": 2.7101,
      "theoretical_loss": 3.4730484563907273,
      "tokens_seen": 1728118784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845252525252525,
      "loss": 2.7504,
      "theoretical_loss": 3.4730371569245113,
      "tokens_seen": 1728184320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845050505050505,
      "loss": 2.3934,
      "theoretical_loss": 3.473025858006759,
      "tokens_seen": 1728249856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844848484848485,
      "loss": 2.4794,
      "theoretical_loss": 3.4730145596374213,
      "tokens_seen": 1728315392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844646464646465,
      "loss": 2.4167,
      "theoretical_loss": 3.4730032618164524,
      "tokens_seen": 1728380928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048444444444444446,
      "loss": 2.7904,
      "theoretical_loss": 3.4729919645438034,
      "tokens_seen": 1728446464
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844242424242424,
      "loss": 2.5223,
      "theoretical_loss": 3.472980667819428,
      "tokens_seen": 1728512000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844040404040404,
      "loss": 2.7771,
      "theoretical_loss": 3.4729693716432783,
      "tokens_seen": 1728577536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843838383838384,
      "loss": 2.6938,
      "theoretical_loss": 3.472958076015307,
      "tokens_seen": 1728643072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843636363636364,
      "loss": 2.6798,
      "theoretical_loss": 3.472946780935467,
      "tokens_seen": 1728708608
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 954102,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.302818536758423,
      "objective/train/theoretical_loss": 3.4729354864037107,
      "objective/train/tokens_used": 87633376,
      "theoretical_loss": 3.4729354864037107,
      "tokens_seen": 1728774144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048434343434343435,
      "loss": 2.6855,
      "theoretical_loss": 3.4729354864037107,
      "tokens_seen": 1728774144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843232323232323,
      "loss": 2.6738,
      "theoretical_loss": 3.4729241924199905,
      "tokens_seen": 1728839680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843030303030303,
      "loss": 2.552,
      "theoretical_loss": 3.4729128989842595,
      "tokens_seen": 1728905216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842828282828283,
      "loss": 2.682,
      "theoretical_loss": 3.4729016060964697,
      "tokens_seen": 1728970752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842626262626263,
      "loss": 2.4425,
      "theoretical_loss": 3.4728903137565745,
      "tokens_seen": 1729036288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048424242424242425,
      "loss": 2.7817,
      "theoretical_loss": 3.472879021964526,
      "tokens_seen": 1729101824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842222222222222,
      "loss": 2.825,
      "theoretical_loss": 3.472867730720277,
      "tokens_seen": 1729167360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842020202020202,
      "loss": 2.5866,
      "theoretical_loss": 3.4728564400237802,
      "tokens_seen": 1729232896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048418181818181817,
      "loss": 2.9395,
      "theoretical_loss": 3.4728451498749884,
      "tokens_seen": 1729298432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841616161616162,
      "loss": 2.7942,
      "theoretical_loss": 3.4728338602738544,
      "tokens_seen": 1729363968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048414141414141414,
      "loss": 2.5925,
      "theoretical_loss": 3.4728225712203304,
      "tokens_seen": 1729429504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048412121212121215,
      "loss": 2.5568,
      "theoretical_loss": 3.4728112827143693,
      "tokens_seen": 1729495040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841010101010101,
      "loss": 2.7105,
      "theoretical_loss": 3.472799994755924,
      "tokens_seen": 1729560576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048408080808080807,
      "loss": 2.7117,
      "theoretical_loss": 3.4727887073449466,
      "tokens_seen": 1729626112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840606060606061,
      "loss": 2.687,
      "theoretical_loss": 3.4727774204813904,
      "tokens_seen": 1729691648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048404040404040403,
      "loss": 2.7663,
      "theoretical_loss": 3.472766134165208,
      "tokens_seen": 1729757184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048402020202020204,
      "loss": 2.5287,
      "theoretical_loss": 3.472754848396352,
      "tokens_seen": 1729822720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484,
      "loss": 2.6449,
      "theoretical_loss": 3.4727435631747756,
      "tokens_seen": 1729888256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048397979797979796,
      "loss": 2.7483,
      "theoretical_loss": 3.4727322785004304,
      "tokens_seen": 1729953792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048395959595959597,
      "loss": 2.5295,
      "theoretical_loss": 3.4727209943732706,
      "tokens_seen": 1730019328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483939393939394,
      "loss": 2.597,
      "theoretical_loss": 3.4727097107932474,
      "tokens_seen": 1730084864
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048391919191919194,
      "loss": 2.5403,
      "theoretical_loss": 3.4726984277603146,
      "tokens_seen": 1730150400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838989898989899,
      "loss": 2.7501,
      "theoretical_loss": 3.472687145274425,
      "tokens_seen": 1730215936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048387878787878785,
      "loss": 2.6343,
      "theoretical_loss": 3.472675863335531,
      "tokens_seen": 1730281472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048385858585858586,
      "loss": 2.5461,
      "theoretical_loss": 3.472664581943585,
      "tokens_seen": 1730347008
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 954813,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2854347229003906,
      "objective/train/theoretical_loss": 3.47265330109854,
      "objective/train/tokens_used": 89271776,
      "theoretical_loss": 3.47265330109854,
      "tokens_seen": 1730412544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048383838383838387,
      "loss": 2.5293,
      "theoretical_loss": 3.47265330109854,
      "tokens_seen": 1730412544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048381818181818183,
      "loss": 2.7872,
      "theoretical_loss": 3.4726420208003494,
      "tokens_seen": 1730478080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837979797979798,
      "loss": 2.7759,
      "theoretical_loss": 3.472630741048966,
      "tokens_seen": 1730543616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048377777777777774,
      "loss": 2.4633,
      "theoretical_loss": 3.4726194618443413,
      "tokens_seen": 1730609152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837575757575758,
      "loss": 2.6828,
      "theoretical_loss": 3.4726081831864293,
      "tokens_seen": 1730674688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048373737373737377,
      "loss": 2.6382,
      "theoretical_loss": 3.472596905075182,
      "tokens_seen": 1730740224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837171717171717,
      "loss": 2.6293,
      "theoretical_loss": 3.4725856275105533,
      "tokens_seen": 1730805760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836969696969697,
      "loss": 2.421,
      "theoretical_loss": 3.472574350492495,
      "tokens_seen": 1730871296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048367676767676764,
      "loss": 2.5533,
      "theoretical_loss": 3.47256307402096,
      "tokens_seen": 1730936832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836565656565657,
      "loss": 2.8264,
      "theoretical_loss": 3.4725517980959015,
      "tokens_seen": 1731002368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048363636363636366,
      "loss": 2.4918,
      "theoretical_loss": 3.4725405227172725,
      "tokens_seen": 1731067904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836161616161616,
      "loss": 2.5235,
      "theoretical_loss": 3.472529247885025,
      "tokens_seen": 1731133440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048359595959595957,
      "loss": 2.5957,
      "theoretical_loss": 3.472517973599113,
      "tokens_seen": 1731198976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048357575757575764,
      "loss": 2.8658,
      "theoretical_loss": 3.4725066998594882,
      "tokens_seen": 1731264512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835555555555556,
      "loss": 2.5194,
      "theoretical_loss": 3.4724954266661046,
      "tokens_seen": 1731330048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048353535353535355,
      "loss": 2.8079,
      "theoretical_loss": 3.4724841540189137,
      "tokens_seen": 1731395584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835151515151515,
      "loss": 2.7591,
      "theoretical_loss": 3.472472881917869,
      "tokens_seen": 1731461120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048349494949494946,
      "loss": 2.6462,
      "theoretical_loss": 3.4724616103629242,
      "tokens_seen": 1731526656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048347474747474753,
      "loss": 2.4711,
      "theoretical_loss": 3.4724503393540314,
      "tokens_seen": 1731592192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834545454545455,
      "loss": 2.5492,
      "theoretical_loss": 3.472439068891143,
      "tokens_seen": 1731657728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048343434343434344,
      "loss": 2.604,
      "theoretical_loss": 3.4724277989742127,
      "tokens_seen": 1731723264
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834141414141414,
      "loss": 2.6603,
      "theoretical_loss": 3.4724165296031924,
      "tokens_seen": 1731788800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833939393939394,
      "loss": 2.5478,
      "theoretical_loss": 3.4724052607780367,
      "tokens_seen": 1731854336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833737373737374,
      "loss": 2.4989,
      "theoretical_loss": 3.472393992498697,
      "tokens_seen": 1731919872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833535353535354,
      "loss": 2.5639,
      "theoretical_loss": 3.4723827247651267,
      "tokens_seen": 1731985408
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 956095,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3975791931152344,
      "objective/train/theoretical_loss": 3.472371457577279,
      "objective/train/tokens_used": 90910176,
      "theoretical_loss": 3.472371457577279,
      "tokens_seen": 1732050944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048333333333333334,
      "loss": 2.5653,
      "theoretical_loss": 3.472371457577279,
      "tokens_seen": 1732050944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833131313131313,
      "loss": 2.6249,
      "theoretical_loss": 3.472360190935106,
      "tokens_seen": 1732116480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832929292929293,
      "loss": 2.7115,
      "theoretical_loss": 3.472348924838562,
      "tokens_seen": 1732182016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832727272727273,
      "loss": 2.7388,
      "theoretical_loss": 3.4723376592875983,
      "tokens_seen": 1732247552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048325252525252527,
      "loss": 2.572,
      "theoretical_loss": 3.4723263942821694,
      "tokens_seen": 1732313088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048323232323232323,
      "loss": 2.7735,
      "theoretical_loss": 3.472315129822227,
      "tokens_seen": 1732378624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048321212121212124,
      "loss": 2.4005,
      "theoretical_loss": 3.4723038659077248,
      "tokens_seen": 1732444160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831919191919192,
      "loss": 2.7573,
      "theoretical_loss": 3.4722926025386154,
      "tokens_seen": 1732509696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831717171717172,
      "loss": 2.6348,
      "theoretical_loss": 3.4722813397148515,
      "tokens_seen": 1732575232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048315151515151516,
      "loss": 2.5441,
      "theoretical_loss": 3.472270077436387,
      "tokens_seen": 1732640768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831313131313131,
      "loss": 2.6699,
      "theoretical_loss": 3.4722588157031744,
      "tokens_seen": 1732706304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048311111111111113,
      "loss": 2.6538,
      "theoretical_loss": 3.472247554515166,
      "tokens_seen": 1732771840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830909090909091,
      "loss": 2.4593,
      "theoretical_loss": 3.4722362938723155,
      "tokens_seen": 1732837376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830707070707071,
      "loss": 2.5461,
      "theoretical_loss": 3.4722250337745764,
      "tokens_seen": 1732902912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048305050505050506,
      "loss": 2.7509,
      "theoretical_loss": 3.472213774221901,
      "tokens_seen": 1732968448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048303030303030307,
      "loss": 2.5207,
      "theoretical_loss": 3.4722025152142413,
      "tokens_seen": 1733033984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483010101010101,
      "loss": 2.7144,
      "theoretical_loss": 3.472191256751552,
      "tokens_seen": 1733099520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482989898989899,
      "loss": 2.578,
      "theoretical_loss": 3.472179998833786,
      "tokens_seen": 1733165056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482969696969697,
      "loss": 2.6586,
      "theoretical_loss": 3.4721687414608953,
      "tokens_seen": 1733230592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048294949494949495,
      "loss": 2.826,
      "theoretical_loss": 3.4721574846328336,
      "tokens_seen": 1733296128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048292929292929296,
      "loss": 2.6236,
      "theoretical_loss": 3.472146228349554,
      "tokens_seen": 1733361664
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829090909090909,
      "loss": 2.5089,
      "theoretical_loss": 3.472134972611009,
      "tokens_seen": 1733427200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828888888888889,
      "loss": 2.6058,
      "theoretical_loss": 3.472123717417152,
      "tokens_seen": 1733492736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828686868686869,
      "loss": 2.4047,
      "theoretical_loss": 3.472112462767936,
      "tokens_seen": 1733558272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828484848484849,
      "loss": 2.6876,
      "theoretical_loss": 3.4721012086633145,
      "tokens_seen": 1733623808
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 956643,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.931276798248291,
      "objective/train/theoretical_loss": 3.4720899551032396,
      "objective/train/tokens_used": 92548576,
      "theoretical_loss": 3.4720899551032396,
      "tokens_seen": 1733689344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048282828282828285,
      "loss": 2.5012,
      "theoretical_loss": 3.4720899551032396,
      "tokens_seen": 1733689344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828080808080808,
      "loss": 2.5515,
      "theoretical_loss": 3.472078702087665,
      "tokens_seen": 1733754880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048278787878787877,
      "loss": 2.6655,
      "theoretical_loss": 3.472067449616544,
      "tokens_seen": 1733820416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827676767676768,
      "loss": 2.7347,
      "theoretical_loss": 3.4720561976898288,
      "tokens_seen": 1733885952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827474747474748,
      "loss": 2.7386,
      "theoretical_loss": 3.4720449463074736,
      "tokens_seen": 1733951488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048272727272727275,
      "loss": 2.4823,
      "theoretical_loss": 3.47203369546943,
      "tokens_seen": 1734017024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827070707070707,
      "loss": 2.6493,
      "theoretical_loss": 3.472022445175653,
      "tokens_seen": 1734082560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048268686868686866,
      "loss": 2.746,
      "theoretical_loss": 3.472011195426095,
      "tokens_seen": 1734148096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048266666666666667,
      "loss": 2.4995,
      "theoretical_loss": 3.4719999462207083,
      "tokens_seen": 1734213632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826464646464647,
      "loss": 2.6906,
      "theoretical_loss": 3.4719886975594463,
      "tokens_seen": 1734279168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048262626262626264,
      "loss": 2.3716,
      "theoretical_loss": 3.471977449442263,
      "tokens_seen": 1734344704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826060606060606,
      "loss": 2.4103,
      "theoretical_loss": 3.47196620186911,
      "tokens_seen": 1734410240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048258585858585855,
      "loss": 2.7933,
      "theoretical_loss": 3.471954954839942,
      "tokens_seen": 1734475776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825656565656566,
      "loss": 2.6078,
      "theoretical_loss": 3.4719437083547113,
      "tokens_seen": 1734541312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825454545454546,
      "loss": 2.5246,
      "theoretical_loss": 3.4719324624133714,
      "tokens_seen": 1734606848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048252525252525253,
      "loss": 2.755,
      "theoretical_loss": 3.4719212170158755,
      "tokens_seen": 1734672384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825050505050505,
      "loss": 2.2628,
      "theoretical_loss": 3.471909972162176,
      "tokens_seen": 1734737920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048248484848484845,
      "loss": 2.6222,
      "theoretical_loss": 3.471898727852227,
      "tokens_seen": 1734803456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824646464646465,
      "loss": 2.6469,
      "theoretical_loss": 3.471887484085981,
      "tokens_seen": 1734868992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048244444444444447,
      "loss": 2.6871,
      "theoretical_loss": 3.4718762408633914,
      "tokens_seen": 1734934528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824242424242424,
      "loss": 2.5698,
      "theoretical_loss": 3.471864998184411,
      "tokens_seen": 1735000064
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824040404040404,
      "loss": 2.8557,
      "theoretical_loss": 3.471853756048994,
      "tokens_seen": 1735065600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823838383838384,
      "loss": 2.6444,
      "theoretical_loss": 3.4718425144570926,
      "tokens_seen": 1735131136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823636363636364,
      "loss": 2.6187,
      "theoretical_loss": 3.471831273408661,
      "tokens_seen": 1735196672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048234343434343436,
      "loss": 2.5576,
      "theoretical_loss": 3.471820032903651,
      "tokens_seen": 1735262208
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 957795,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0068306922912598,
      "objective/train/theoretical_loss": 3.471808792942017,
      "objective/train/tokens_used": 94186976,
      "theoretical_loss": 3.471808792942017,
      "tokens_seen": 1735327744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823232323232323,
      "loss": 2.839,
      "theoretical_loss": 3.471808792942017,
      "tokens_seen": 1735327744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004823030303030303,
      "loss": 2.612,
      "theoretical_loss": 3.4717975535237118,
      "tokens_seen": 1735393280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822828282828283,
      "loss": 2.5873,
      "theoretical_loss": 3.4717863146486887,
      "tokens_seen": 1735458816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822626262626263,
      "loss": 2.7101,
      "theoretical_loss": 3.4717750763169004,
      "tokens_seen": 1735524352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048224242424242425,
      "loss": 2.5618,
      "theoretical_loss": 3.471763838528301,
      "tokens_seen": 1735589888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822222222222222,
      "loss": 2.7326,
      "theoretical_loss": 3.471752601282843,
      "tokens_seen": 1735655424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822020202020202,
      "loss": 2.3858,
      "theoretical_loss": 3.47174136458048,
      "tokens_seen": 1735720960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821818181818182,
      "loss": 2.6821,
      "theoretical_loss": 3.4717301284211657,
      "tokens_seen": 1735786496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821616161616162,
      "loss": 2.6055,
      "theoretical_loss": 3.4717188928048524,
      "tokens_seen": 1735852032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048214141414141415,
      "loss": 2.8098,
      "theoretical_loss": 3.4717076577314936,
      "tokens_seen": 1735917568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821212121212121,
      "loss": 2.555,
      "theoretical_loss": 3.471696423201043,
      "tokens_seen": 1735983104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821010101010101,
      "loss": 2.7927,
      "theoretical_loss": 3.4716851892134537,
      "tokens_seen": 1736048640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048208080808080807,
      "loss": 2.7408,
      "theoretical_loss": 3.4716739557686793,
      "tokens_seen": 1736114176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820606060606061,
      "loss": 2.9396,
      "theoretical_loss": 3.471662722866672,
      "tokens_seen": 1736179712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048204040404040404,
      "loss": 2.5873,
      "theoretical_loss": 3.4716514905073863,
      "tokens_seen": 1736245248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048202020202020205,
      "loss": 2.5494,
      "theoretical_loss": 3.471640258690775,
      "tokens_seen": 1736310784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482,
      "loss": 2.6813,
      "theoretical_loss": 3.4716290274167907,
      "tokens_seen": 1736376320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048197979797979796,
      "loss": 2.6974,
      "theoretical_loss": 3.471617796685388,
      "tokens_seen": 1736441856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481959595959596,
      "loss": 2.6864,
      "theoretical_loss": 3.4716065664965194,
      "tokens_seen": 1736507392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048193939393939393,
      "loss": 2.7289,
      "theoretical_loss": 3.4715953368501387,
      "tokens_seen": 1736572928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048191919191919194,
      "loss": 2.7123,
      "theoretical_loss": 3.4715841077461986,
      "tokens_seen": 1736638464
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818989898989899,
      "loss": 2.6387,
      "theoretical_loss": 3.471572879184653,
      "tokens_seen": 1736704000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818787878787879,
      "loss": 2.763,
      "theoretical_loss": 3.4715616511654552,
      "tokens_seen": 1736769536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048185858585858587,
      "loss": 2.7991,
      "theoretical_loss": 3.4715504236885577,
      "tokens_seen": 1736835072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818383838383839,
      "loss": 2.7678,
      "theoretical_loss": 3.4715391967539153,
      "tokens_seen": 1736900608
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 958846,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5784518718719482,
      "objective/train/theoretical_loss": 3.47152797036148,
      "objective/train/tokens_used": 95825376,
      "theoretical_loss": 3.47152797036148,
      "tokens_seen": 1736966144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048181818181818184,
      "loss": 2.7097,
      "theoretical_loss": 3.47152797036148,
      "tokens_seen": 1736966144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817979797979798,
      "loss": 2.537,
      "theoretical_loss": 3.471516744511206,
      "tokens_seen": 1737031680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817777777777778,
      "loss": 2.6579,
      "theoretical_loss": 3.471505519203046,
      "tokens_seen": 1737097216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048175757575757576,
      "loss": 2.7483,
      "theoretical_loss": 3.471494294436954,
      "tokens_seen": 1737162752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048173737373737377,
      "loss": 2.5049,
      "theoretical_loss": 3.471483070212883,
      "tokens_seen": 1737228288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048171717171717173,
      "loss": 2.7261,
      "theoretical_loss": 3.471471846530787,
      "tokens_seen": 1737293824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816969696969697,
      "loss": 2.6989,
      "theoretical_loss": 3.471460623390618,
      "tokens_seen": 1737359360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816767676767677,
      "loss": 2.6896,
      "theoretical_loss": 3.4714494007923307,
      "tokens_seen": 1737424896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816565656565657,
      "loss": 2.8564,
      "theoretical_loss": 3.471438178735878,
      "tokens_seen": 1737490432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048163636363636366,
      "loss": 2.6221,
      "theoretical_loss": 3.4714269572212135,
      "tokens_seen": 1737555968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816161616161616,
      "loss": 2.733,
      "theoretical_loss": 3.4714157362482907,
      "tokens_seen": 1737621504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815959595959596,
      "loss": 2.8006,
      "theoretical_loss": 3.4714045158170626,
      "tokens_seen": 1737687040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815757575757576,
      "loss": 2.64,
      "theoretical_loss": 3.4713932959274825,
      "tokens_seen": 1737752576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815555555555556,
      "loss": 2.7127,
      "theoretical_loss": 3.4713820765795047,
      "tokens_seen": 1737818112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048153535353535356,
      "loss": 2.929,
      "theoretical_loss": 3.4713708577730817,
      "tokens_seen": 1737883648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815151515151515,
      "loss": 2.7397,
      "theoretical_loss": 3.471359639508168,
      "tokens_seen": 1737949184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048149494949494947,
      "loss": 2.7401,
      "theoretical_loss": 3.4713484217847155,
      "tokens_seen": 1738014720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048147474747474754,
      "loss": 2.6971,
      "theoretical_loss": 3.471337204602679,
      "tokens_seen": 1738080256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814545454545455,
      "loss": 2.7005,
      "theoretical_loss": 3.471325987962011,
      "tokens_seen": 1738145792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048143434343434345,
      "loss": 2.5562,
      "theoretical_loss": 3.471314771862666,
      "tokens_seen": 1738211328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814141414141414,
      "loss": 2.667,
      "theoretical_loss": 3.4713035563045964,
      "tokens_seen": 1738276864
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048139393939393936,
      "loss": 2.7304,
      "theoretical_loss": 3.4712923412877563,
      "tokens_seen": 1738342400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048137373737373743,
      "loss": 2.521,
      "theoretical_loss": 3.4712811268120993,
      "tokens_seen": 1738407936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813535353535354,
      "loss": 2.5026,
      "theoretical_loss": 3.4712699128775784,
      "tokens_seen": 1738473472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048133333333333334,
      "loss": 2.799,
      "theoretical_loss": 3.4712586994841477,
      "tokens_seen": 1738539008
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 959288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.807804584503174,
      "objective/train/theoretical_loss": 3.4712474866317597,
      "objective/train/tokens_used": 97463776,
      "theoretical_loss": 3.4712474866317597,
      "tokens_seen": 1738604544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813131313131313,
      "loss": 2.7561,
      "theoretical_loss": 3.4712474866317597,
      "tokens_seen": 1738604544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048129292929292926,
      "loss": 2.7609,
      "theoretical_loss": 3.4712362743203693,
      "tokens_seen": 1738670080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812727272727273,
      "loss": 2.797,
      "theoretical_loss": 3.4712250625499284,
      "tokens_seen": 1738735616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812525252525253,
      "loss": 3.041,
      "theoretical_loss": 3.471213851320392,
      "tokens_seen": 1738801152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048123232323232323,
      "loss": 2.8472,
      "theoretical_loss": 3.471202640631713,
      "tokens_seen": 1738866688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812121212121212,
      "loss": 2.5309,
      "theoretical_loss": 3.4711914304838443,
      "tokens_seen": 1738932224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811919191919192,
      "loss": 2.6006,
      "theoretical_loss": 3.4711802208767404,
      "tokens_seen": 1738997760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811717171717172,
      "loss": 2.6453,
      "theoretical_loss": 3.471169011810354,
      "tokens_seen": 1739063296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048115151515151517,
      "loss": 2.7669,
      "theoretical_loss": 3.4711578032846404,
      "tokens_seen": 1739128832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048113131313131313,
      "loss": 2.3778,
      "theoretical_loss": 3.4711465952995506,
      "tokens_seen": 1739194368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811111111111111,
      "loss": 2.7263,
      "theoretical_loss": 3.47113538785504,
      "tokens_seen": 1739259904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810909090909091,
      "loss": 2.5269,
      "theoretical_loss": 3.471124180951061,
      "tokens_seen": 1739325440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810707070707071,
      "loss": 2.8646,
      "theoretical_loss": 3.471112974587568,
      "tokens_seen": 1739390976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048105050505050506,
      "loss": 2.6309,
      "theoretical_loss": 3.4711017687645147,
      "tokens_seen": 1739456512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481030303030303,
      "loss": 2.4525,
      "theoretical_loss": 3.471090563481854,
      "tokens_seen": 1739522048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048101010101010103,
      "loss": 2.6909,
      "theoretical_loss": 3.4710793587395394,
      "tokens_seen": 1739587584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480989898989899,
      "loss": 2.8344,
      "theoretical_loss": 3.4710681545375253,
      "tokens_seen": 1739653120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480969696969697,
      "loss": 2.8379,
      "theoretical_loss": 3.4710569508757647,
      "tokens_seen": 1739718656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048094949494949496,
      "loss": 2.708,
      "theoretical_loss": 3.471045747754211,
      "tokens_seen": 1739784192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809292929292929,
      "loss": 2.6413,
      "theoretical_loss": 3.4710345451728184,
      "tokens_seen": 1739849728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809090909090909,
      "loss": 2.6477,
      "theoretical_loss": 3.4710233431315407,
      "tokens_seen": 1739915264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808888888888889,
      "loss": 2.4457,
      "theoretical_loss": 3.4710121416303306,
      "tokens_seen": 1739980800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808686868686869,
      "loss": 2.6397,
      "theoretical_loss": 3.471000940669142,
      "tokens_seen": 1740046336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048084848484848485,
      "loss": 2.6021,
      "theoretical_loss": 3.470989740247929,
      "tokens_seen": 1740111872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048082828282828286,
      "loss": 2.6974,
      "theoretical_loss": 3.4709785403666453,
      "tokens_seen": 1740177408
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 960545,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9208710193634033,
      "objective/train/theoretical_loss": 3.4709673410252435,
      "objective/train/tokens_used": 99102176,
      "theoretical_loss": 3.4709673410252435,
      "tokens_seen": 1740242944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808080808080808,
      "loss": 2.7373,
      "theoretical_loss": 3.4709673410252435,
      "tokens_seen": 1740242944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807878787878788,
      "loss": 2.8115,
      "theoretical_loss": 3.4709561422236783,
      "tokens_seen": 1740308480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807676767676768,
      "loss": 2.7494,
      "theoretical_loss": 3.470944943961903,
      "tokens_seen": 1740374016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048074747474747474,
      "loss": 2.6765,
      "theoretical_loss": 3.4709337462398713,
      "tokens_seen": 1740439552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048072727272727275,
      "loss": 2.9857,
      "theoretical_loss": 3.470922549057537,
      "tokens_seen": 1740505088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807070707070707,
      "loss": 2.443,
      "theoretical_loss": 3.470911352414853,
      "tokens_seen": 1740570624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048068686868686867,
      "loss": 2.6088,
      "theoretical_loss": 3.4709001563117736,
      "tokens_seen": 1740636160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806666666666667,
      "loss": 2.8324,
      "theoretical_loss": 3.470888960748253,
      "tokens_seen": 1740701696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806464646464647,
      "loss": 2.7392,
      "theoretical_loss": 3.470877765724244,
      "tokens_seen": 1740767232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048062626262626264,
      "loss": 2.5289,
      "theoretical_loss": 3.4708665712397004,
      "tokens_seen": 1740832768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806060606060606,
      "loss": 2.7885,
      "theoretical_loss": 3.4708553772945763,
      "tokens_seen": 1740898304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048058585858585856,
      "loss": 2.5816,
      "theoretical_loss": 3.4708441838888255,
      "tokens_seen": 1740963840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048056565656565657,
      "loss": 2.7401,
      "theoretical_loss": 3.4708329910224007,
      "tokens_seen": 1741029376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805454545454546,
      "loss": 2.6518,
      "theoretical_loss": 3.4708217986952574,
      "tokens_seen": 1741094912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048052525252525254,
      "loss": 2.6621,
      "theoretical_loss": 3.470810606907347,
      "tokens_seen": 1741160448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805050505050505,
      "loss": 2.468,
      "theoretical_loss": 3.4707994156586253,
      "tokens_seen": 1741225984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048048484848484845,
      "loss": 2.56,
      "theoretical_loss": 3.470788224949045,
      "tokens_seen": 1741291520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804646464646465,
      "loss": 2.6133,
      "theoretical_loss": 3.47077703477856,
      "tokens_seen": 1741357056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804444444444445,
      "loss": 2.4266,
      "theoretical_loss": 3.470765845147124,
      "tokens_seen": 1741422592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048042424242424243,
      "loss": 2.5242,
      "theoretical_loss": 3.470754656054691,
      "tokens_seen": 1741488128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804040404040404,
      "loss": 2.733,
      "theoretical_loss": 3.4707434675012148,
      "tokens_seen": 1741553664
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048038383838383834,
      "loss": 2.8107,
      "theoretical_loss": 3.470732279486649,
      "tokens_seen": 1741619200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803636363636364,
      "loss": 2.6628,
      "theoretical_loss": 3.4707210920109466,
      "tokens_seen": 1741684736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048034343434343437,
      "loss": 2.7073,
      "theoretical_loss": 3.470709905074062,
      "tokens_seen": 1741750272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803232323232323,
      "loss": 2.6392,
      "theoretical_loss": 3.47069871867595,
      "tokens_seen": 1741815808
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 961825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8929381370544434,
      "objective/train/theoretical_loss": 3.470687532816563,
      "objective/train/tokens_used": 100740576,
      "theoretical_loss": 3.470687532816563,
      "tokens_seen": 1741881344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803030303030303,
      "loss": 2.7658,
      "theoretical_loss": 3.470687532816563,
      "tokens_seen": 1741881344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802828282828283,
      "loss": 2.7095,
      "theoretical_loss": 3.4706763474958553,
      "tokens_seen": 1741946880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802626262626263,
      "loss": 2.6271,
      "theoretical_loss": 3.4706651627137806,
      "tokens_seen": 1742012416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048024242424242426,
      "loss": 2.6202,
      "theoretical_loss": 3.4706539784702928,
      "tokens_seen": 1742077952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802222222222222,
      "loss": 2.6904,
      "theoretical_loss": 3.470642794765345,
      "tokens_seen": 1742143488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048020202020202017,
      "loss": 2.2801,
      "theoretical_loss": 3.4706316115988924,
      "tokens_seen": 1742209024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801818181818182,
      "loss": 2.7515,
      "theoretical_loss": 3.4706204289708875,
      "tokens_seen": 1742274560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801616161616162,
      "loss": 2.8955,
      "theoretical_loss": 3.4706092468812852,
      "tokens_seen": 1742340096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048014141414141415,
      "loss": 2.5899,
      "theoretical_loss": 3.4705980653300386,
      "tokens_seen": 1742405632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801212121212121,
      "loss": 2.5155,
      "theoretical_loss": 3.4705868843171017,
      "tokens_seen": 1742471168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801010101010101,
      "loss": 2.469,
      "theoretical_loss": 3.4705757038424285,
      "tokens_seen": 1742536704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048008080808080813,
      "loss": 2.6819,
      "theoretical_loss": 3.4705645239059724,
      "tokens_seen": 1742602240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800606060606061,
      "loss": 2.9573,
      "theoretical_loss": 3.470553344507688,
      "tokens_seen": 1742667776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048004040404040404,
      "loss": 2.6073,
      "theoretical_loss": 3.4705421656475286,
      "tokens_seen": 1742733312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480020202020202,
      "loss": 2.5603,
      "theoretical_loss": 3.470530987325448,
      "tokens_seen": 1742798848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048,
      "loss": 2.5829,
      "theoretical_loss": 3.4705198095414005,
      "tokens_seen": 1742864384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479979797979798,
      "loss": 2.6065,
      "theoretical_loss": 3.4705086322953393,
      "tokens_seen": 1742929920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479959595959596,
      "loss": 2.7619,
      "theoretical_loss": 3.4704974555872194,
      "tokens_seen": 1742995456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047993939393939394,
      "loss": 2.5569,
      "theoretical_loss": 3.4704862794169937,
      "tokens_seen": 1743060992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047991919191919195,
      "loss": 2.7131,
      "theoretical_loss": 3.470475103784616,
      "tokens_seen": 1743126528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798989898989899,
      "loss": 2.5933,
      "theoretical_loss": 3.470463928690041,
      "tokens_seen": 1743192064
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798787878787879,
      "loss": 2.505,
      "theoretical_loss": 3.4704527541332224,
      "tokens_seen": 1743257600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047985858585858587,
      "loss": 2.608,
      "theoretical_loss": 3.4704415801141133,
      "tokens_seen": 1743323136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047983838383838383,
      "loss": 2.523,
      "theoretical_loss": 3.4704304066326683,
      "tokens_seen": 1743388672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047981818181818184,
      "loss": 2.5742,
      "theoretical_loss": 3.4704192336888413,
      "tokens_seen": 1743454208
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 962337,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5374813079833984,
      "objective/train/theoretical_loss": 3.470408061282587,
      "objective/train/tokens_used": 102378976,
      "theoretical_loss": 3.470408061282587,
      "tokens_seen": 1743519744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797979797979798,
      "loss": 2.3709,
      "theoretical_loss": 3.470408061282587,
      "tokens_seen": 1743519744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797777777777778,
      "loss": 2.8128,
      "theoretical_loss": 3.4703968894138573,
      "tokens_seen": 1743585280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047975757575757577,
      "loss": 2.4218,
      "theoretical_loss": 3.470385718082608,
      "tokens_seen": 1743650816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797373737373738,
      "loss": 2.7803,
      "theoretical_loss": 3.470374547288792,
      "tokens_seen": 1743716352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047971717171717173,
      "loss": 2.8138,
      "theoretical_loss": 3.470363377032364,
      "tokens_seen": 1743781888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796969696969697,
      "loss": 2.3758,
      "theoretical_loss": 3.4703522073132773,
      "tokens_seen": 1743847424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796767676767677,
      "loss": 2.7284,
      "theoretical_loss": 3.470341038131486,
      "tokens_seen": 1743912960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047965656565656566,
      "loss": 2.6017,
      "theoretical_loss": 3.470329869486944,
      "tokens_seen": 1743978496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047963636363636367,
      "loss": 3.0104,
      "theoretical_loss": 3.470318701379606,
      "tokens_seen": 1744044032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796161616161616,
      "loss": 2.6543,
      "theoretical_loss": 3.4703075338094256,
      "tokens_seen": 1744109568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795959595959596,
      "loss": 2.7775,
      "theoretical_loss": 3.4702963667763562,
      "tokens_seen": 1744175104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795757575757576,
      "loss": 2.5585,
      "theoretical_loss": 3.470285200280352,
      "tokens_seen": 1744240640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795555555555556,
      "loss": 2.6517,
      "theoretical_loss": 3.470274034321368,
      "tokens_seen": 1744306176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047953535353535356,
      "loss": 2.394,
      "theoretical_loss": 3.470262868899357,
      "tokens_seen": 1744371712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795151515151515,
      "loss": 2.7431,
      "theoretical_loss": 3.470251704014273,
      "tokens_seen": 1744437248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794949494949495,
      "loss": 2.6337,
      "theoretical_loss": 3.4702405396660705,
      "tokens_seen": 1744502784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794747474747475,
      "loss": 2.5813,
      "theoretical_loss": 3.4702293758547036,
      "tokens_seen": 1744568320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794545454545455,
      "loss": 2.7258,
      "theoretical_loss": 3.470218212580126,
      "tokens_seen": 1744633856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047943434343434345,
      "loss": 2.6717,
      "theoretical_loss": 3.4702070498422923,
      "tokens_seen": 1744699392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794141414141414,
      "loss": 2.3807,
      "theoretical_loss": 3.4701958876411556,
      "tokens_seen": 1744764928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047939393939393937,
      "loss": 2.7787,
      "theoretical_loss": 3.4701847259766705,
      "tokens_seen": 1744830464
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793737373737374,
      "loss": 2.4776,
      "theoretical_loss": 3.4701735648487912,
      "tokens_seen": 1744896000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793535353535354,
      "loss": 2.5712,
      "theoretical_loss": 3.4701624042574717,
      "tokens_seen": 1744961536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047933333333333335,
      "loss": 2.6089,
      "theoretical_loss": 3.4701512442026656,
      "tokens_seen": 1745027072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793131313131313,
      "loss": 2.6412,
      "theoretical_loss": 3.470140084684327,
      "tokens_seen": 1745092608
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 963376,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.749994993209839,
      "objective/train/theoretical_loss": 3.4701289257024106,
      "objective/train/tokens_used": 104017376,
      "theoretical_loss": 3.4701289257024106,
      "tokens_seen": 1745158144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047929292929292926,
      "loss": 2.6664,
      "theoretical_loss": 3.4701289257024106,
      "tokens_seen": 1745158144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792727272727273,
      "loss": 2.6718,
      "theoretical_loss": 3.47011776725687,
      "tokens_seen": 1745223680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792525252525253,
      "loss": 2.7504,
      "theoretical_loss": 3.47010660934766,
      "tokens_seen": 1745289216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047923232323232324,
      "loss": 2.894,
      "theoretical_loss": 3.470095451974733,
      "tokens_seen": 1745354752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792121212121212,
      "loss": 2.6854,
      "theoretical_loss": 3.4700842951380446,
      "tokens_seen": 1745420288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047919191919191915,
      "loss": 2.602,
      "theoretical_loss": 3.4700731388375483,
      "tokens_seen": 1745485824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791717171717172,
      "loss": 2.5376,
      "theoretical_loss": 3.4700619830731982,
      "tokens_seen": 1745551360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791515151515152,
      "loss": 2.6504,
      "theoretical_loss": 3.470050827844949,
      "tokens_seen": 1745616896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047913131313131313,
      "loss": 2.6948,
      "theoretical_loss": 3.4700396731527543,
      "tokens_seen": 1745682432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791111111111111,
      "loss": 2.7571,
      "theoretical_loss": 3.470028518996568,
      "tokens_seen": 1745747968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790909090909091,
      "loss": 2.4343,
      "theoretical_loss": 3.4700173653763446,
      "tokens_seen": 1745813504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790707070707071,
      "loss": 2.7223,
      "theoretical_loss": 3.470006212292038,
      "tokens_seen": 1745879040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047905050505050507,
      "loss": 2.7366,
      "theoretical_loss": 3.469995059743603,
      "tokens_seen": 1745944576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479030303030303,
      "loss": 2.7484,
      "theoretical_loss": 3.4699839077309926,
      "tokens_seen": 1746010112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479010101010101,
      "loss": 2.5738,
      "theoretical_loss": 3.469972756254162,
      "tokens_seen": 1746075648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478989898989899,
      "loss": 2.7647,
      "theoretical_loss": 3.469961605313064,
      "tokens_seen": 1746141184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478969696969697,
      "loss": 2.8169,
      "theoretical_loss": 3.4699504549076545,
      "tokens_seen": 1746206720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047894949494949496,
      "loss": 2.7003,
      "theoretical_loss": 3.4699393050378866,
      "tokens_seen": 1746272256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789292929292929,
      "loss": 2.7409,
      "theoretical_loss": 3.4699281557037147,
      "tokens_seen": 1746337792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047890909090909093,
      "loss": 2.7155,
      "theoretical_loss": 3.469917006905093,
      "tokens_seen": 1746403328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788888888888889,
      "loss": 2.8076,
      "theoretical_loss": 3.4699058586419755,
      "tokens_seen": 1746468864
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788686868686869,
      "loss": 2.7611,
      "theoretical_loss": 3.469894710914317,
      "tokens_seen": 1746534400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047884848484848485,
      "loss": 2.6351,
      "theoretical_loss": 3.46988356372207,
      "tokens_seen": 1746599936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788282828282828,
      "loss": 2.8063,
      "theoretical_loss": 3.469872417065191,
      "tokens_seen": 1746665472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788080808080808,
      "loss": 2.7563,
      "theoretical_loss": 3.4698612709436327,
      "tokens_seen": 1746731008
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 963934,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.45017409324646,
      "objective/train/theoretical_loss": 3.46985012535735,
      "objective/train/tokens_used": 105655776,
      "theoretical_loss": 3.46985012535735,
      "tokens_seen": 1746796544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787878787878788,
      "loss": 2.6769,
      "theoretical_loss": 3.46985012535735,
      "tokens_seen": 1746796544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787676767676768,
      "loss": 2.6498,
      "theoretical_loss": 3.469838980306297,
      "tokens_seen": 1746862080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047874747474747475,
      "loss": 2.7676,
      "theoretical_loss": 3.469827835790427,
      "tokens_seen": 1746927616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047872727272727276,
      "loss": 2.7206,
      "theoretical_loss": 3.4698166918096955,
      "tokens_seen": 1746993152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787070707070707,
      "loss": 2.6543,
      "theoretical_loss": 3.469805548364056,
      "tokens_seen": 1747058688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047868686868686867,
      "loss": 2.7194,
      "theoretical_loss": 3.469794405453463,
      "tokens_seen": 1747124224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786666666666667,
      "loss": 2.7218,
      "theoretical_loss": 3.4697832630778707,
      "tokens_seen": 1747189760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047864646464646464,
      "loss": 2.6617,
      "theoretical_loss": 3.4697721212372334,
      "tokens_seen": 1747255296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047862626262626265,
      "loss": 2.4966,
      "theoretical_loss": 3.4697609799315052,
      "tokens_seen": 1747320832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786060606060606,
      "loss": 2.5697,
      "theoretical_loss": 3.4697498391606403,
      "tokens_seen": 1747386368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047858585858585856,
      "loss": 2.792,
      "theoretical_loss": 3.4697386989245933,
      "tokens_seen": 1747451904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785656565656566,
      "loss": 2.5246,
      "theoretical_loss": 3.469727559223318,
      "tokens_seen": 1747517440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785454545454546,
      "loss": 2.7315,
      "theoretical_loss": 3.469716420056769,
      "tokens_seen": 1747582976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047852525252525254,
      "loss": 2.7625,
      "theoretical_loss": 3.469705281424901,
      "tokens_seen": 1747648512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785050505050505,
      "loss": 2.4715,
      "theoretical_loss": 3.4696941433276676,
      "tokens_seen": 1747714048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047848484848484846,
      "loss": 2.7081,
      "theoretical_loss": 3.469683005765023,
      "tokens_seen": 1747779584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047846464646464647,
      "loss": 2.5652,
      "theoretical_loss": 3.4696718687369223,
      "tokens_seen": 1747845120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784444444444445,
      "loss": 2.6408,
      "theoretical_loss": 3.469660732243319,
      "tokens_seen": 1747910656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047842424242424244,
      "loss": 2.6097,
      "theoretical_loss": 3.4696495962841674,
      "tokens_seen": 1747976192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784040404040404,
      "loss": 2.6966,
      "theoretical_loss": 3.469638460859423,
      "tokens_seen": 1748041728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783838383838384,
      "loss": 2.6111,
      "theoretical_loss": 3.4696273259690384,
      "tokens_seen": 1748107264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783636363636364,
      "loss": 2.7655,
      "theoretical_loss": 3.4696161916129693,
      "tokens_seen": 1748172800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047834343434343437,
      "loss": 2.5875,
      "theoretical_loss": 3.4696050577911692,
      "tokens_seen": 1748238336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047832323232323233,
      "loss": 2.698,
      "theoretical_loss": 3.469593924503593,
      "tokens_seen": 1748303872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783030303030303,
      "loss": 2.7526,
      "theoretical_loss": 3.4695827917501947,
      "tokens_seen": 1748369408
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 965147,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.617971897125244,
      "objective/train/theoretical_loss": 3.4695716595309287,
      "objective/train/tokens_used": 107294176,
      "theoretical_loss": 3.4695716595309287,
      "tokens_seen": 1748434944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782828282828283,
      "loss": 2.5613,
      "theoretical_loss": 3.4695716595309287,
      "tokens_seen": 1748434944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782626262626263,
      "loss": 2.9621,
      "theoretical_loss": 3.4695605278457493,
      "tokens_seen": 1748500480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047824242424242426,
      "loss": 2.642,
      "theoretical_loss": 3.469549396694611,
      "tokens_seen": 1748566016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782222222222222,
      "loss": 2.7141,
      "theoretical_loss": 3.4695382660774685,
      "tokens_seen": 1748631552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782020202020202,
      "loss": 2.5499,
      "theoretical_loss": 3.469527135994275,
      "tokens_seen": 1748697088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047818181818181824,
      "loss": 2.4784,
      "theoretical_loss": 3.4695160064449864,
      "tokens_seen": 1748762624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781616161616162,
      "loss": 2.7111,
      "theoretical_loss": 3.469504877429556,
      "tokens_seen": 1748828160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047814141414141416,
      "loss": 2.5869,
      "theoretical_loss": 3.4694937489479383,
      "tokens_seen": 1748893696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781212121212121,
      "loss": 2.4923,
      "theoretical_loss": 3.4694826210000884,
      "tokens_seen": 1748959232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047810101010101007,
      "loss": 2.6399,
      "theoretical_loss": 3.46947149358596,
      "tokens_seen": 1749024768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047808080808080814,
      "loss": 2.807,
      "theoretical_loss": 3.4694603667055075,
      "tokens_seen": 1749090304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780606060606061,
      "loss": 2.3863,
      "theoretical_loss": 3.469449240358686,
      "tokens_seen": 1749155840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047804040404040405,
      "loss": 2.8256,
      "theoretical_loss": 3.4694381145454494,
      "tokens_seen": 1749221376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478020202020202,
      "loss": 2.7606,
      "theoretical_loss": 3.469426989265752,
      "tokens_seen": 1749286912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047799999999999996,
      "loss": 2.6758,
      "theoretical_loss": 3.469415864519548,
      "tokens_seen": 1749352448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047797979797979803,
      "loss": 2.6428,
      "theoretical_loss": 3.469404740306793,
      "tokens_seen": 1749417984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000477959595959596,
      "loss": 2.7873,
      "theoretical_loss": 3.469393616627441,
      "tokens_seen": 1749483520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047793939393939394,
      "loss": 2.7165,
      "theoretical_loss": 3.469382493481445,
      "tokens_seen": 1749549056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779191919191919,
      "loss": 2.7791,
      "theoretical_loss": 3.4693713708687612,
      "tokens_seen": 1749614592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778989898989899,
      "loss": 2.8045,
      "theoretical_loss": 3.469360248789343,
      "tokens_seen": 1749680128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778787878787879,
      "loss": 2.5269,
      "theoretical_loss": 3.4693491272431456,
      "tokens_seen": 1749745664
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778585858585859,
      "loss": 2.742,
      "theoretical_loss": 3.469338006230123,
      "tokens_seen": 1749811200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047783838383838384,
      "loss": 2.7999,
      "theoretical_loss": 3.46932688575023,
      "tokens_seen": 1749876736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778181818181818,
      "loss": 2.782,
      "theoretical_loss": 3.4693157658034206,
      "tokens_seen": 1749942272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777979797979798,
      "loss": 2.8005,
      "theoretical_loss": 3.46930464638965,
      "tokens_seen": 1750007808
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 966110,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6748225688934326,
      "objective/train/theoretical_loss": 3.469293527508872,
      "objective/train/tokens_used": 108932576,
      "theoretical_loss": 3.469293527508872,
      "tokens_seen": 1750073344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777777777777778,
      "loss": 2.6659,
      "theoretical_loss": 3.469293527508872,
      "tokens_seen": 1750073344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047775757575757577,
      "loss": 2.7751,
      "theoretical_loss": 3.4692824091610412,
      "tokens_seen": 1750138880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047773737373737373,
      "loss": 2.6498,
      "theoretical_loss": 3.4692712913461126,
      "tokens_seen": 1750204416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047771717171717174,
      "loss": 2.6655,
      "theoretical_loss": 3.4692601740640403,
      "tokens_seen": 1750269952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776969696969697,
      "loss": 2.6462,
      "theoretical_loss": 3.4692490573147787,
      "tokens_seen": 1750335488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776767676767677,
      "loss": 2.8046,
      "theoretical_loss": 3.4692379410982825,
      "tokens_seen": 1750401024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047765656565656566,
      "loss": 2.5931,
      "theoretical_loss": 3.4692268254145064,
      "tokens_seen": 1750466560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776363636363636,
      "loss": 2.7954,
      "theoretical_loss": 3.4692157102634047,
      "tokens_seen": 1750532096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047761616161616163,
      "loss": 2.6183,
      "theoretical_loss": 3.469204595644932,
      "tokens_seen": 1750597632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775959595959596,
      "loss": 2.779,
      "theoretical_loss": 3.4691934815590426,
      "tokens_seen": 1750663168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775757575757576,
      "loss": 2.8523,
      "theoretical_loss": 3.4691823680056917,
      "tokens_seen": 1750728704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047755555555555556,
      "loss": 2.5914,
      "theoretical_loss": 3.4691712549848326,
      "tokens_seen": 1750794240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047753535353535357,
      "loss": 2.6619,
      "theoretical_loss": 3.4691601424964214,
      "tokens_seen": 1750859776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775151515151515,
      "loss": 2.845,
      "theoretical_loss": 3.469149030540412,
      "tokens_seen": 1750925312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774949494949495,
      "loss": 2.6006,
      "theoretical_loss": 3.469137919116758,
      "tokens_seen": 1750990848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774747474747475,
      "loss": 2.6609,
      "theoretical_loss": 3.4691268082254156,
      "tokens_seen": 1751056384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047745454545454545,
      "loss": 2.7653,
      "theoretical_loss": 3.4691156978663384,
      "tokens_seen": 1751121920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047743434343434346,
      "loss": 2.7009,
      "theoretical_loss": 3.4691045880394817,
      "tokens_seen": 1751187456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774141414141414,
      "loss": 2.8441,
      "theoretical_loss": 3.469093478744799,
      "tokens_seen": 1751252992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773939393939394,
      "loss": 2.6236,
      "theoretical_loss": 3.4690823699822455,
      "tokens_seen": 1751318528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773737373737374,
      "loss": 2.6359,
      "theoretical_loss": 3.4690712617517763,
      "tokens_seen": 1751384064
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773535353535354,
      "loss": 2.6701,
      "theoretical_loss": 3.469060154053345,
      "tokens_seen": 1751449600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047733333333333335,
      "loss": 2.5259,
      "theoretical_loss": 3.469049046886907,
      "tokens_seen": 1751515136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773131313131313,
      "loss": 2.7385,
      "theoretical_loss": 3.4690379402524165,
      "tokens_seen": 1751580672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047729292929292927,
      "loss": 2.4743,
      "theoretical_loss": 3.4690268341498287,
      "tokens_seen": 1751646208
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 966555,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6258327960968018,
      "objective/train/theoretical_loss": 3.4690157285790972,
      "objective/train/tokens_used": 110570976,
      "theoretical_loss": 3.4690157285790972,
      "tokens_seen": 1751711744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772727272727273,
      "loss": 2.6509,
      "theoretical_loss": 3.4690157285790972,
      "tokens_seen": 1751711744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772525252525253,
      "loss": 2.809,
      "theoretical_loss": 3.4690046235401777,
      "tokens_seen": 1751777280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047723232323232325,
      "loss": 2.6795,
      "theoretical_loss": 3.468993519033024,
      "tokens_seen": 1751842816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772121212121212,
      "loss": 2.7073,
      "theoretical_loss": 3.4689824150575914,
      "tokens_seen": 1751908352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047719191919191916,
      "loss": 2.6694,
      "theoretical_loss": 3.468971311613834,
      "tokens_seen": 1751973888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771717171717172,
      "loss": 2.6556,
      "theoretical_loss": 3.4689602087017066,
      "tokens_seen": 1752039424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771515151515152,
      "loss": 2.7105,
      "theoretical_loss": 3.468949106321164,
      "tokens_seen": 1752104960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047713131313131314,
      "loss": 2.5831,
      "theoretical_loss": 3.468938004472161,
      "tokens_seen": 1752170496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771111111111111,
      "loss": 2.644,
      "theoretical_loss": 3.468926903154652,
      "tokens_seen": 1752236032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047709090909090905,
      "loss": 2.5805,
      "theoretical_loss": 3.468915802368592,
      "tokens_seen": 1752301568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770707070707071,
      "loss": 2.8613,
      "theoretical_loss": 3.4689047021139356,
      "tokens_seen": 1752367104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770505050505051,
      "loss": 2.6725,
      "theoretical_loss": 3.4688936023906374,
      "tokens_seen": 1752432640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047703030303030303,
      "loss": 2.838,
      "theoretical_loss": 3.4688825031986514,
      "tokens_seen": 1752498176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477010101010101,
      "loss": 2.4936,
      "theoretical_loss": 3.468871404537934,
      "tokens_seen": 1752563712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476989898989899,
      "loss": 2.7542,
      "theoretical_loss": 3.4688603064084376,
      "tokens_seen": 1752629248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476969696969697,
      "loss": 2.6908,
      "theoretical_loss": 3.468849208810119,
      "tokens_seen": 1752694784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047694949494949497,
      "loss": 2.6054,
      "theoretical_loss": 3.4688381117429317,
      "tokens_seen": 1752760320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769292929292929,
      "loss": 2.5833,
      "theoretical_loss": 3.468827015206831,
      "tokens_seen": 1752825856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769090909090909,
      "loss": 3.0322,
      "theoretical_loss": 3.4688159192017713,
      "tokens_seen": 1752891392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768888888888889,
      "loss": 2.5106,
      "theoretical_loss": 3.468804823727708,
      "tokens_seen": 1752956928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768686868686869,
      "loss": 2.5625,
      "theoretical_loss": 3.4687937287845947,
      "tokens_seen": 1753022464
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047684848484848486,
      "loss": 2.8008,
      "theoretical_loss": 3.468782634372387,
      "tokens_seen": 1753088000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768282828282828,
      "loss": 2.5919,
      "theoretical_loss": 3.4687715404910398,
      "tokens_seen": 1753153536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047680808080808083,
      "loss": 2.6055,
      "theoretical_loss": 3.4687604471405065,
      "tokens_seen": 1753219072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767878787878788,
      "loss": 2.7941,
      "theoretical_loss": 3.4687493543207433,
      "tokens_seen": 1753284608
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 967544,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.789055347442627,
      "objective/train/theoretical_loss": 3.468738262031705,
      "objective/train/tokens_used": 112209376,
      "theoretical_loss": 3.468738262031705,
      "tokens_seen": 1753350144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767676767676768,
      "loss": 2.6272,
      "theoretical_loss": 3.468738262031705,
      "tokens_seen": 1753350144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047674747474747475,
      "loss": 2.7313,
      "theoretical_loss": 3.468727170273345,
      "tokens_seen": 1753415680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767272727272727,
      "loss": 2.6265,
      "theoretical_loss": 3.4687160790456195,
      "tokens_seen": 1753481216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767070707070707,
      "loss": 2.5517,
      "theoretical_loss": 3.468704988348483,
      "tokens_seen": 1753546752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047668686868686873,
      "loss": 2.6153,
      "theoretical_loss": 3.4686938981818893,
      "tokens_seen": 1753612288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766666666666667,
      "loss": 2.698,
      "theoretical_loss": 3.468682808545794,
      "tokens_seen": 1753677824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047664646464646465,
      "loss": 2.5592,
      "theoretical_loss": 3.4686717194401524,
      "tokens_seen": 1753743360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047662626262626266,
      "loss": 2.6959,
      "theoretical_loss": 3.4686606308649184,
      "tokens_seen": 1753808896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766060606060606,
      "loss": 2.7546,
      "theoretical_loss": 3.468649542820047,
      "tokens_seen": 1753874432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765858585858586,
      "loss": 2.7334,
      "theoretical_loss": 3.468638455305493,
      "tokens_seen": 1753939968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765656565656566,
      "loss": 2.7488,
      "theoretical_loss": 3.4686273683212114,
      "tokens_seen": 1754005504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047654545454545454,
      "loss": 2.5704,
      "theoretical_loss": 3.468616281867157,
      "tokens_seen": 1754071040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047652525252525255,
      "loss": 2.7951,
      "theoretical_loss": 3.468605195943285,
      "tokens_seen": 1754136576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765050505050505,
      "loss": 2.7617,
      "theoretical_loss": 3.46859411054955,
      "tokens_seen": 1754202112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764848484848485,
      "loss": 2.7249,
      "theoretical_loss": 3.468583025685906,
      "tokens_seen": 1754267648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764646464646465,
      "loss": 2.8578,
      "theoretical_loss": 3.468571941352309,
      "tokens_seen": 1754333184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764444444444445,
      "loss": 2.7968,
      "theoretical_loss": 3.4685608575487135,
      "tokens_seen": 1754398720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047642424242424244,
      "loss": 2.5969,
      "theoretical_loss": 3.4685497742750737,
      "tokens_seen": 1754464256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764040404040404,
      "loss": 2.6532,
      "theoretical_loss": 3.4685386915313456,
      "tokens_seen": 1754529792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763838383838384,
      "loss": 2.5712,
      "theoretical_loss": 3.4685276093174835,
      "tokens_seen": 1754595328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047636363636363637,
      "loss": 2.7184,
      "theoretical_loss": 3.468516527633442,
      "tokens_seen": 1754660864
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763434343434344,
      "loss": 2.666,
      "theoretical_loss": 3.4685054464791767,
      "tokens_seen": 1754726400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047632323232323233,
      "loss": 2.704,
      "theoretical_loss": 3.4684943658546414,
      "tokens_seen": 1754791936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763030303030303,
      "loss": 2.5184,
      "theoretical_loss": 3.4684832857597923,
      "tokens_seen": 1754857472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762828282828283,
      "loss": 2.5821,
      "theoretical_loss": 3.4684722061945834,
      "tokens_seen": 1754923008
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 968097,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9240682125091553,
      "objective/train/theoretical_loss": 3.46846112715897,
      "objective/train/tokens_used": 113847776,
      "theoretical_loss": 3.46846112715897,
      "tokens_seen": 1754988544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762626262626263,
      "loss": 2.6846,
      "theoretical_loss": 3.46846112715897,
      "tokens_seen": 1754988544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047624242424242427,
      "loss": 2.686,
      "theoretical_loss": 3.4684500486529064,
      "tokens_seen": 1755054080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762222222222222,
      "loss": 2.6209,
      "theoretical_loss": 3.468438970676348,
      "tokens_seen": 1755119616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762020202020202,
      "loss": 2.5909,
      "theoretical_loss": 3.4684278932292503,
      "tokens_seen": 1755185152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761818181818182,
      "loss": 2.7275,
      "theoretical_loss": 3.468416816311567,
      "tokens_seen": 1755250688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761616161616162,
      "loss": 2.7313,
      "theoretical_loss": 3.4684057399232544,
      "tokens_seen": 1755316224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047614141414141416,
      "loss": 2.5685,
      "theoretical_loss": 3.468394664064266,
      "tokens_seen": 1755381760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761212121212121,
      "loss": 2.6424,
      "theoretical_loss": 3.468383588734558,
      "tokens_seen": 1755447296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761010101010101,
      "loss": 2.3658,
      "theoretical_loss": 3.468372513934084,
      "tokens_seen": 1755512832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760808080808081,
      "loss": 2.8822,
      "theoretical_loss": 3.468361439662801,
      "tokens_seen": 1755578368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760606060606061,
      "loss": 2.768,
      "theoretical_loss": 3.468350365920662,
      "tokens_seen": 1755643904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047604040404040406,
      "loss": 2.4781,
      "theoretical_loss": 3.468339292707623,
      "tokens_seen": 1755709440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476020202020202,
      "loss": 2.509,
      "theoretical_loss": 3.468328220023638,
      "tokens_seen": 1755774976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047599999999999997,
      "loss": 2.556,
      "theoretical_loss": 3.468317147868663,
      "tokens_seen": 1755840512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047597979797979803,
      "loss": 2.6309,
      "theoretical_loss": 3.4683060762426523,
      "tokens_seen": 1755906048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475959595959596,
      "loss": 2.6173,
      "theoretical_loss": 3.4682950051455617,
      "tokens_seen": 1755971584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047593939393939395,
      "loss": 2.502,
      "theoretical_loss": 3.4682839345773457,
      "tokens_seen": 1756037120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759191919191919,
      "loss": 2.7037,
      "theoretical_loss": 3.468272864537959,
      "tokens_seen": 1756102656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047589898989898986,
      "loss": 2.9021,
      "theoretical_loss": 3.468261795027357,
      "tokens_seen": 1756168192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047587878787878793,
      "loss": 2.6436,
      "theoretical_loss": 3.4682507260454942,
      "tokens_seen": 1756233728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758585858585859,
      "loss": 2.5381,
      "theoretical_loss": 3.4682396575923264,
      "tokens_seen": 1756299264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047583838383838384,
      "loss": 2.6176,
      "theoretical_loss": 3.4682285896678087,
      "tokens_seen": 1756364800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758181818181818,
      "loss": 2.5754,
      "theoretical_loss": 3.468217522271895,
      "tokens_seen": 1756430336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757979797979798,
      "loss": 2.6936,
      "theoretical_loss": 3.468206455404541,
      "tokens_seen": 1756495872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757777777777778,
      "loss": 2.4587,
      "theoretical_loss": 3.468195389065702,
      "tokens_seen": 1756561408
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 969105,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5639123916625977,
      "objective/train/theoretical_loss": 3.4681843232553327,
      "objective/train/tokens_used": 115486176,
      "theoretical_loss": 3.4681843232553327,
      "tokens_seen": 1756626944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757575757575758,
      "loss": 2.5915,
      "theoretical_loss": 3.4681843232553327,
      "tokens_seen": 1756626944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047573737373737373,
      "loss": 2.5335,
      "theoretical_loss": 3.4681732579733886,
      "tokens_seen": 1756692480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757171717171717,
      "loss": 2.4869,
      "theoretical_loss": 3.4681621932198237,
      "tokens_seen": 1756758016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756969696969697,
      "loss": 2.8309,
      "theoretical_loss": 3.4681511289945943,
      "tokens_seen": 1756823552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756767676767677,
      "loss": 2.832,
      "theoretical_loss": 3.4681400652976544,
      "tokens_seen": 1756889088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047565656565656567,
      "loss": 2.6746,
      "theoretical_loss": 3.46812900212896,
      "tokens_seen": 1756954624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756363636363636,
      "loss": 2.6248,
      "theoretical_loss": 3.4681179394884656,
      "tokens_seen": 1757020160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047561616161616164,
      "loss": 2.5682,
      "theoretical_loss": 3.468106877376126,
      "tokens_seen": 1757085696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755959595959596,
      "loss": 2.6822,
      "theoretical_loss": 3.4680958157918975,
      "tokens_seen": 1757151232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755757575757576,
      "loss": 2.8153,
      "theoretical_loss": 3.468084754735734,
      "tokens_seen": 1757216768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047555555555555556,
      "loss": 2.7997,
      "theoretical_loss": 3.4680736942075914,
      "tokens_seen": 1757282304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755353535353535,
      "loss": 2.6609,
      "theoretical_loss": 3.468062634207424,
      "tokens_seen": 1757347840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047551515151515153,
      "loss": 2.8563,
      "theoretical_loss": 3.4680515747351874,
      "tokens_seen": 1757413376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754949494949495,
      "loss": 2.6059,
      "theoretical_loss": 3.4680405157908365,
      "tokens_seen": 1757478912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754747474747475,
      "loss": 2.8125,
      "theoretical_loss": 3.4680294573743264,
      "tokens_seen": 1757544448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047545454545454545,
      "loss": 2.7202,
      "theoretical_loss": 3.4680183994856133,
      "tokens_seen": 1757609984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047543434343434347,
      "loss": 2.6924,
      "theoretical_loss": 3.4680073421246504,
      "tokens_seen": 1757675520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754141414141414,
      "loss": 2.6198,
      "theoretical_loss": 3.4679962852913944,
      "tokens_seen": 1757741056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753939393939394,
      "loss": 2.6276,
      "theoretical_loss": 3.4679852289857998,
      "tokens_seen": 1757806592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753737373737374,
      "loss": 2.7586,
      "theoretical_loss": 3.4679741732078213,
      "tokens_seen": 1757872128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047535353535353535,
      "loss": 2.4646,
      "theoretical_loss": 3.4679631179574155,
      "tokens_seen": 1757937664
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047533333333333336,
      "loss": 2.4745,
      "theoretical_loss": 3.467952063234536,
      "tokens_seen": 1758003200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753131313131313,
      "loss": 2.7784,
      "theoretical_loss": 3.467941009039139,
      "tokens_seen": 1758068736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047529292929292927,
      "loss": 2.7118,
      "theoretical_loss": 3.4679299553711793,
      "tokens_seen": 1758134272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752727272727273,
      "loss": 2.8808,
      "theoretical_loss": 3.4679189022306116,
      "tokens_seen": 1758199808
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 970436,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.550333023071289,
      "objective/train/theoretical_loss": 3.467907849617392,
      "objective/train/tokens_used": 117124576,
      "theoretical_loss": 3.467907849617392,
      "tokens_seen": 1758265344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752525252525253,
      "loss": 2.6393,
      "theoretical_loss": 3.467907849617392,
      "tokens_seen": 1758265344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047523232323232325,
      "loss": 2.488,
      "theoretical_loss": 3.467896797531475,
      "tokens_seen": 1758330880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752121212121212,
      "loss": 2.7108,
      "theoretical_loss": 3.467885745972816,
      "tokens_seen": 1758396416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047519191919191916,
      "loss": 2.7059,
      "theoretical_loss": 3.4678746949413704,
      "tokens_seen": 1758461952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751717171717172,
      "loss": 2.5608,
      "theoretical_loss": 3.467863644437093,
      "tokens_seen": 1758527488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751515151515152,
      "loss": 2.7159,
      "theoretical_loss": 3.467852594459939,
      "tokens_seen": 1758593024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047513131313131314,
      "loss": 2.6068,
      "theoretical_loss": 3.4678415450098643,
      "tokens_seen": 1758658560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751111111111111,
      "loss": 2.5219,
      "theoretical_loss": 3.4678304960868234,
      "tokens_seen": 1758724096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047509090909090906,
      "loss": 2.5771,
      "theoretical_loss": 3.467819447690772,
      "tokens_seen": 1758789632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750707070707071,
      "loss": 2.6642,
      "theoretical_loss": 3.4678083998216653,
      "tokens_seen": 1758855168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750505050505051,
      "loss": 2.6359,
      "theoretical_loss": 3.467797352479458,
      "tokens_seen": 1758920704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047503030303030304,
      "loss": 2.6904,
      "theoretical_loss": 3.4677863056641054,
      "tokens_seen": 1758986240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475010101010101,
      "loss": 2.5726,
      "theoretical_loss": 3.4677752593755633,
      "tokens_seen": 1759051776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474989898989899,
      "loss": 2.689,
      "theoretical_loss": 3.467764213613787,
      "tokens_seen": 1759117312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474969696969697,
      "loss": 2.5023,
      "theoretical_loss": 3.4677531683787306,
      "tokens_seen": 1759182848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047494949494949497,
      "loss": 2.5113,
      "theoretical_loss": 3.4677421236703507,
      "tokens_seen": 1759248384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047492929292929293,
      "loss": 2.7619,
      "theoretical_loss": 3.4677310794886025,
      "tokens_seen": 1759313920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749090909090909,
      "loss": 2.5726,
      "theoretical_loss": 3.46772003583344,
      "tokens_seen": 1759379456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047488888888888895,
      "loss": 2.424,
      "theoretical_loss": 3.46770899270482,
      "tokens_seen": 1759444992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748686868686869,
      "loss": 2.705,
      "theoretical_loss": 3.467697950102697,
      "tokens_seen": 1759510528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047484848484848487,
      "loss": 2.8012,
      "theoretical_loss": 3.467686908027026,
      "tokens_seen": 1759576064
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748282828282828,
      "loss": 2.6163,
      "theoretical_loss": 3.467675866477763,
      "tokens_seen": 1759641600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748080808080808,
      "loss": 2.7308,
      "theoretical_loss": 3.4676648254548628,
      "tokens_seen": 1759707136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047478787878787884,
      "loss": 2.6315,
      "theoretical_loss": 3.467653784958281,
      "tokens_seen": 1759772672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747676767676768,
      "loss": 2.8835,
      "theoretical_loss": 3.467642744987973,
      "tokens_seen": 1759838208
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 971034,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.658268690109253,
      "objective/train/theoretical_loss": 3.467631705543894,
      "objective/train/tokens_used": 118762976,
      "theoretical_loss": 3.467631705543894,
      "tokens_seen": 1759903744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047474747474747476,
      "loss": 2.441,
      "theoretical_loss": 3.467631705543894,
      "tokens_seen": 1759903744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747272727272727,
      "loss": 2.4756,
      "theoretical_loss": 3.467620666625999,
      "tokens_seen": 1759969280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047470707070707067,
      "loss": 2.5877,
      "theoretical_loss": 3.4676096282342432,
      "tokens_seen": 1760034816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047468686868686874,
      "loss": 2.4157,
      "theoretical_loss": 3.4675985903685826,
      "tokens_seen": 1760100352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746666666666667,
      "loss": 2.6185,
      "theoretical_loss": 3.4675875530289724,
      "tokens_seen": 1760165888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047464646464646465,
      "loss": 2.626,
      "theoretical_loss": 3.4675765162153676,
      "tokens_seen": 1760231424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746262626262626,
      "loss": 2.6303,
      "theoretical_loss": 3.467565479927724,
      "tokens_seen": 1760296960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746060606060606,
      "loss": 2.4665,
      "theoretical_loss": 3.4675544441659962,
      "tokens_seen": 1760362496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047458585858585863,
      "loss": 2.4859,
      "theoretical_loss": 3.4675434089301405,
      "tokens_seen": 1760428032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745656565656566,
      "loss": 2.7329,
      "theoretical_loss": 3.467532374220112,
      "tokens_seen": 1760493568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047454545454545454,
      "loss": 2.7522,
      "theoretical_loss": 3.4675213400358653,
      "tokens_seen": 1760559104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745252525252525,
      "loss": 2.6709,
      "theoretical_loss": 3.4675103063773567,
      "tokens_seen": 1760624640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745050505050505,
      "loss": 2.3443,
      "theoretical_loss": 3.4674992732445413,
      "tokens_seen": 1760690176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744848484848485,
      "loss": 2.5856,
      "theoretical_loss": 3.4674882406373744,
      "tokens_seen": 1760755712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744646464646465,
      "loss": 2.8132,
      "theoretical_loss": 3.4674772085558114,
      "tokens_seen": 1760821248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047444444444444444,
      "loss": 2.691,
      "theoretical_loss": 3.4674661769998076,
      "tokens_seen": 1760886784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047442424242424245,
      "loss": 2.8995,
      "theoretical_loss": 3.4674551459693186,
      "tokens_seen": 1760952320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744040404040404,
      "loss": 2.5649,
      "theoretical_loss": 3.4674441154643,
      "tokens_seen": 1761017856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743838383838384,
      "loss": 2.7209,
      "theoretical_loss": 3.467433085484706,
      "tokens_seen": 1761083392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047436363636363637,
      "loss": 2.8425,
      "theoretical_loss": 3.467422056030494,
      "tokens_seen": 1761148928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047434343434343433,
      "loss": 2.7739,
      "theoretical_loss": 3.467411027101618,
      "tokens_seen": 1761214464
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047432323232323234,
      "loss": 2.6837,
      "theoretical_loss": 3.467399998698034,
      "tokens_seen": 1761280000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743030303030303,
      "loss": 2.7335,
      "theoretical_loss": 3.467388970819697,
      "tokens_seen": 1761345536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742828282828283,
      "loss": 2.6222,
      "theoretical_loss": 3.4673779434665635,
      "tokens_seen": 1761411072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047426262626262626,
      "loss": 2.6726,
      "theoretical_loss": 3.4673669166385874,
      "tokens_seen": 1761476608
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 972196,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6714515686035156,
      "objective/train/theoretical_loss": 3.4673558903357247,
      "objective/train/tokens_used": 120401376,
      "theoretical_loss": 3.4673558903357247,
      "tokens_seen": 1761542144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742424242424243,
      "loss": 2.6306,
      "theoretical_loss": 3.4673558903357247,
      "tokens_seen": 1761542144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047422222222222223,
      "loss": 2.8951,
      "theoretical_loss": 3.4673448645579317,
      "tokens_seen": 1761607680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742020202020202,
      "loss": 2.7546,
      "theoretical_loss": 3.467333839305163,
      "tokens_seen": 1761673216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741818181818182,
      "loss": 2.5618,
      "theoretical_loss": 3.4673228145773747,
      "tokens_seen": 1761738752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047416161616161616,
      "loss": 2.5558,
      "theoretical_loss": 3.467311790374521,
      "tokens_seen": 1761804288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047414141414141417,
      "loss": 2.3849,
      "theoretical_loss": 3.467300766696559,
      "tokens_seen": 1761869824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741212121212121,
      "loss": 2.6992,
      "theoretical_loss": 3.467289743543443,
      "tokens_seen": 1761935360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741010101010101,
      "loss": 2.6239,
      "theoretical_loss": 3.4672787209151292,
      "tokens_seen": 1762000896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740808080808081,
      "loss": 2.61,
      "theoretical_loss": 3.4672676988115727,
      "tokens_seen": 1762066432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740606060606061,
      "loss": 2.7707,
      "theoretical_loss": 3.4672566772327293,
      "tokens_seen": 1762131968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047404040404040406,
      "loss": 2.5288,
      "theoretical_loss": 3.4672456561785543,
      "tokens_seen": 1762197504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474020202020202,
      "loss": 2.5156,
      "theoretical_loss": 3.467234635649003,
      "tokens_seen": 1762263040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474,
      "loss": 2.3634,
      "theoretical_loss": 3.4672236156440315,
      "tokens_seen": 1762328576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473979797979798,
      "loss": 2.3909,
      "theoretical_loss": 3.467212596163595,
      "tokens_seen": 1762394112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473959595959596,
      "loss": 2.6485,
      "theoretical_loss": 3.467201577207649,
      "tokens_seen": 1762459648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047393939393939395,
      "loss": 2.9205,
      "theoretical_loss": 3.467190558776149,
      "tokens_seen": 1762525184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739191919191919,
      "loss": 2.4281,
      "theoretical_loss": 3.4671795408690507,
      "tokens_seen": 1762590720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047389898989898987,
      "loss": 2.6767,
      "theoretical_loss": 3.4671685234863094,
      "tokens_seen": 1762656256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047387878787878793,
      "loss": 2.6607,
      "theoretical_loss": 3.467157506627881,
      "tokens_seen": 1762721792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738585858585859,
      "loss": 2.4989,
      "theoretical_loss": 3.4671464902937204,
      "tokens_seen": 1762787328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047383838383838385,
      "loss": 2.7566,
      "theoretical_loss": 3.4671354744837837,
      "tokens_seen": 1762852864
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738181818181818,
      "loss": 2.4784,
      "theoretical_loss": 3.4671244591980264,
      "tokens_seen": 1762918400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047379797979797976,
      "loss": 2.6094,
      "theoretical_loss": 3.4671134444364045,
      "tokens_seen": 1762983936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737777777777778,
      "loss": 2.3573,
      "theoretical_loss": 3.467102430198873,
      "tokens_seen": 1763049472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737575757575758,
      "loss": 2.7262,
      "theoretical_loss": 3.4670914164853874,
      "tokens_seen": 1763115008
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 972718,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.458674430847168,
      "objective/train/theoretical_loss": 3.4670804032959035,
      "objective/train/tokens_used": 122039776,
      "theoretical_loss": 3.4670804032959035,
      "tokens_seen": 1763180544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047373737373737374,
      "loss": 2.5665,
      "theoretical_loss": 3.4670804032959035,
      "tokens_seen": 1763180544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737171717171717,
      "loss": 2.4811,
      "theoretical_loss": 3.4670693906303773,
      "tokens_seen": 1763246080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736969696969697,
      "loss": 2.4546,
      "theoretical_loss": 3.4670583784887636,
      "tokens_seen": 1763311616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736767676767677,
      "loss": 2.5793,
      "theoretical_loss": 3.4670473668710184,
      "tokens_seen": 1763377152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736565656565657,
      "loss": 2.5163,
      "theoretical_loss": 3.467036355777098,
      "tokens_seen": 1763442688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047363636363636363,
      "loss": 2.5888,
      "theoretical_loss": 3.467025345206956,
      "tokens_seen": 1763508224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736161616161616,
      "loss": 2.7923,
      "theoretical_loss": 3.4670143351605502,
      "tokens_seen": 1763573760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735959595959596,
      "loss": 2.7194,
      "theoretical_loss": 3.4670033256378354,
      "tokens_seen": 1763639296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735757575757576,
      "loss": 2.4734,
      "theoretical_loss": 3.466992316638767,
      "tokens_seen": 1763704832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047355555555555557,
      "loss": 2.7535,
      "theoretical_loss": 3.4669813081633007,
      "tokens_seen": 1763770368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735353535353535,
      "loss": 2.7308,
      "theoretical_loss": 3.466970300211393,
      "tokens_seen": 1763835904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047351515151515154,
      "loss": 2.579,
      "theoretical_loss": 3.466959292782998,
      "tokens_seen": 1763901440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734949494949495,
      "loss": 2.5358,
      "theoretical_loss": 3.4669482858780727,
      "tokens_seen": 1763966976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734747474747475,
      "loss": 2.7295,
      "theoretical_loss": 3.4669372794965723,
      "tokens_seen": 1764032512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047345454545454546,
      "loss": 2.474,
      "theoretical_loss": 3.4669262736384523,
      "tokens_seen": 1764098048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734343434343434,
      "loss": 2.7381,
      "theoretical_loss": 3.4669152683036684,
      "tokens_seen": 1764163584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047341414141414143,
      "loss": 2.7042,
      "theoretical_loss": 3.466904263492176,
      "tokens_seen": 1764229120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733939393939394,
      "loss": 2.8351,
      "theoretical_loss": 3.466893259203932,
      "tokens_seen": 1764294656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733737373737374,
      "loss": 2.7247,
      "theoretical_loss": 3.4668822554388905,
      "tokens_seen": 1764360192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047335353535353535,
      "loss": 2.5802,
      "theoretical_loss": 3.4668712521970084,
      "tokens_seen": 1764425728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047333333333333336,
      "loss": 2.6002,
      "theoretical_loss": 3.4668602494782403,
      "tokens_seen": 1764491264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733131313131313,
      "loss": 2.6269,
      "theoretical_loss": 3.466849247282543,
      "tokens_seen": 1764556800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732929292929293,
      "loss": 2.6215,
      "theoretical_loss": 3.466838245609872,
      "tokens_seen": 1764622336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732727272727273,
      "loss": 2.6113,
      "theoretical_loss": 3.466827244460182,
      "tokens_seen": 1764687872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047325252525252525,
      "loss": 2.534,
      "theoretical_loss": 3.46681624383343,
      "tokens_seen": 1764753408
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 973844,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.547422170639038,
      "objective/train/theoretical_loss": 3.466805243729571,
      "objective/train/tokens_used": 123678176,
      "theoretical_loss": 3.466805243729571,
      "tokens_seen": 1764818944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047323232323232326,
      "loss": 2.5461,
      "theoretical_loss": 3.466805243729571,
      "tokens_seen": 1764818944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732121212121212,
      "loss": 2.4588,
      "theoretical_loss": 3.466794244148561,
      "tokens_seen": 1764884480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731919191919192,
      "loss": 2.4346,
      "theoretical_loss": 3.4667832450903555,
      "tokens_seen": 1764950016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731717171717172,
      "loss": 2.6804,
      "theoretical_loss": 3.4667722465549105,
      "tokens_seen": 1765015552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731515151515152,
      "loss": 2.5802,
      "theoretical_loss": 3.4667612485421815,
      "tokens_seen": 1765081088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047313131313131315,
      "loss": 2.6148,
      "theoretical_loss": 3.4667502510521246,
      "tokens_seen": 1765146624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731111111111111,
      "loss": 2.625,
      "theoretical_loss": 3.466739254084695,
      "tokens_seen": 1765212160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730909090909091,
      "loss": 2.6565,
      "theoretical_loss": 3.466728257639849,
      "tokens_seen": 1765277696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730707070707071,
      "loss": 2.6265,
      "theoretical_loss": 3.4667172617175424,
      "tokens_seen": 1765343232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730505050505051,
      "loss": 2.5729,
      "theoretical_loss": 3.4667062663177304,
      "tokens_seen": 1765408768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047303030303030304,
      "loss": 2.706,
      "theoretical_loss": 3.4666952714403694,
      "tokens_seen": 1765474304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473010101010101,
      "loss": 2.5028,
      "theoretical_loss": 3.4666842770854145,
      "tokens_seen": 1765539840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472989898989899,
      "loss": 2.7347,
      "theoretical_loss": 3.466673283252822,
      "tokens_seen": 1765605376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472969696969697,
      "loss": 2.2926,
      "theoretical_loss": 3.4666622899425477,
      "tokens_seen": 1765670912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472949494949495,
      "loss": 2.7329,
      "theoretical_loss": 3.4666512971545473,
      "tokens_seen": 1765736448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047292929292929293,
      "loss": 2.7155,
      "theoretical_loss": 3.4666403048887764,
      "tokens_seen": 1765801984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729090909090909,
      "loss": 2.676,
      "theoretical_loss": 3.4666293131451917,
      "tokens_seen": 1765867520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728888888888889,
      "loss": 2.5448,
      "theoretical_loss": 3.4666183219237476,
      "tokens_seen": 1765933056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728686868686869,
      "loss": 2.6503,
      "theoretical_loss": 3.4666073312244006,
      "tokens_seen": 1765998592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047284848484848487,
      "loss": 2.7058,
      "theoretical_loss": 3.466596341047107,
      "tokens_seen": 1766064128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047282828282828283,
      "loss": 2.517,
      "theoretical_loss": 3.4665853513918217,
      "tokens_seen": 1766129664
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728080808080808,
      "loss": 2.626,
      "theoretical_loss": 3.466574362258501,
      "tokens_seen": 1766195200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727878787878788,
      "loss": 2.6928,
      "theoretical_loss": 3.4665633736471015,
      "tokens_seen": 1766260736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727676767676768,
      "loss": 2.6162,
      "theoretical_loss": 3.466552385557578,
      "tokens_seen": 1766326272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047274747474747476,
      "loss": 2.5867,
      "theoretical_loss": 3.4665413979898863,
      "tokens_seen": 1766391808
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 974264,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8634729385375977,
      "objective/train/theoretical_loss": 3.466530410943983,
      "objective/train/tokens_used": 125316576,
      "theoretical_loss": 3.466530410943983,
      "tokens_seen": 1766457344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727272727272727,
      "loss": 2.755,
      "theoretical_loss": 3.466530410943983,
      "tokens_seen": 1766457344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727070707070707,
      "loss": 2.6146,
      "theoretical_loss": 3.466519424419823,
      "tokens_seen": 1766522880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047268686868686874,
      "loss": 2.8209,
      "theoretical_loss": 3.4665084384173634,
      "tokens_seen": 1766588416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726666666666667,
      "loss": 2.7291,
      "theoretical_loss": 3.4664974529365593,
      "tokens_seen": 1766653952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047264646464646466,
      "loss": 2.7675,
      "theoretical_loss": 3.466486467977367,
      "tokens_seen": 1766719488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726262626262626,
      "loss": 2.4818,
      "theoretical_loss": 3.4664754835397416,
      "tokens_seen": 1766785024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047260606060606057,
      "loss": 2.5268,
      "theoretical_loss": 3.4664644996236396,
      "tokens_seen": 1766850560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047258585858585864,
      "loss": 2.5537,
      "theoretical_loss": 3.4664535162290164,
      "tokens_seen": 1766916096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725656565656566,
      "loss": 2.643,
      "theoretical_loss": 3.466442533355829,
      "tokens_seen": 1766981632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047254545454545455,
      "loss": 2.6887,
      "theoretical_loss": 3.466431551004032,
      "tokens_seen": 1767047168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725252525252525,
      "loss": 2.6244,
      "theoretical_loss": 3.4664205691735823,
      "tokens_seen": 1767112704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725050505050505,
      "loss": 2.5747,
      "theoretical_loss": 3.466409587864436,
      "tokens_seen": 1767178240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047248484848484853,
      "loss": 2.3783,
      "theoretical_loss": 3.466398607076547,
      "tokens_seen": 1767243776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724646464646465,
      "loss": 2.6642,
      "theoretical_loss": 3.4663876268098734,
      "tokens_seen": 1767309312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047244444444444444,
      "loss": 2.4038,
      "theoretical_loss": 3.466376647064371,
      "tokens_seen": 1767374848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724242424242424,
      "loss": 2.5637,
      "theoretical_loss": 3.4663656678399946,
      "tokens_seen": 1767440384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724040404040404,
      "loss": 2.6683,
      "theoretical_loss": 3.4663546891367005,
      "tokens_seen": 1767505920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723838383838384,
      "loss": 2.6663,
      "theoretical_loss": 3.4663437109544453,
      "tokens_seen": 1767571456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723636363636364,
      "loss": 2.7499,
      "theoretical_loss": 3.466332733293184,
      "tokens_seen": 1767636992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047234343434343433,
      "loss": 2.6059,
      "theoretical_loss": 3.4663217561528734,
      "tokens_seen": 1767702528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047232323232323235,
      "loss": 2.8539,
      "theoretical_loss": 3.466310779533469,
      "tokens_seen": 1767768064
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723030303030303,
      "loss": 2.6061,
      "theoretical_loss": 3.466299803434927,
      "tokens_seen": 1767833600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722828282828283,
      "loss": 2.7606,
      "theoretical_loss": 3.4662888278572033,
      "tokens_seen": 1767899136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047226262626262627,
      "loss": 2.602,
      "theoretical_loss": 3.4662778528002534,
      "tokens_seen": 1767964672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722424242424242,
      "loss": 2.6908,
      "theoretical_loss": 3.466266878264034,
      "tokens_seen": 1768030208
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 975251,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7586162090301514,
      "objective/train/theoretical_loss": 3.4662559042485013,
      "objective/train/tokens_used": 126954976,
      "theoretical_loss": 3.4662559042485013,
      "tokens_seen": 1768095744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047222222222222224,
      "loss": 2.6865,
      "theoretical_loss": 3.4662559042485013,
      "tokens_seen": 1768095744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004722020202020202,
      "loss": 2.5986,
      "theoretical_loss": 3.4662449307536103,
      "tokens_seen": 1768161280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721818181818182,
      "loss": 2.5451,
      "theoretical_loss": 3.466233957779318,
      "tokens_seen": 1768226816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047216161616161616,
      "loss": 2.5286,
      "theoretical_loss": 3.4662229853255795,
      "tokens_seen": 1768292352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721414141414142,
      "loss": 2.6367,
      "theoretical_loss": 3.466212013392351,
      "tokens_seen": 1768357888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047212121212121213,
      "loss": 2.6325,
      "theoretical_loss": 3.4662010419795894,
      "tokens_seen": 1768423424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721010101010101,
      "loss": 2.5137,
      "theoretical_loss": 3.46619007108725,
      "tokens_seen": 1768488960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720808080808081,
      "loss": 2.4045,
      "theoretical_loss": 3.4661791007152885,
      "tokens_seen": 1768554496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047206060606060606,
      "loss": 2.6393,
      "theoretical_loss": 3.466168130863662,
      "tokens_seen": 1768620032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047204040404040407,
      "loss": 2.4247,
      "theoretical_loss": 3.4661571615323252,
      "tokens_seen": 1768685568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000472020202020202,
      "loss": 2.4163,
      "theoretical_loss": 3.466146192721235,
      "tokens_seen": 1768751104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000472,
      "loss": 2.6915,
      "theoretical_loss": 3.4661352244303476,
      "tokens_seen": 1768816640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000471979797979798,
      "loss": 2.4705,
      "theoretical_loss": 3.4661242566596187,
      "tokens_seen": 1768882176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000471959595959596,
      "loss": 2.6915,
      "theoretical_loss": 3.4661132894090043,
      "tokens_seen": 1768947712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047193939393939396,
      "loss": 2.5054,
      "theoretical_loss": 3.46610232267846,
      "tokens_seen": 1769013248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719191919191919,
      "loss": 2.6938,
      "theoretical_loss": 3.4660913564679436,
      "tokens_seen": 1769078784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718989898989899,
      "loss": 2.5973,
      "theoretical_loss": 3.4660803907774094,
      "tokens_seen": 1769144320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718787878787879,
      "loss": 2.6252,
      "theoretical_loss": 3.4660694256068143,
      "tokens_seen": 1769209856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718585858585859,
      "loss": 2.7995,
      "theoretical_loss": 3.4660584609561145,
      "tokens_seen": 1769275392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047183838383838385,
      "loss": 2.5113,
      "theoretical_loss": 3.466047496825265,
      "tokens_seen": 1769340928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718181818181818,
      "loss": 2.58,
      "theoretical_loss": 3.4660365332142233,
      "tokens_seen": 1769406464
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047179797979797977,
      "loss": 2.6229,
      "theoretical_loss": 3.4660255701229445,
      "tokens_seen": 1769472000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047177777777777783,
      "loss": 2.625,
      "theoretical_loss": 3.4660146075513856,
      "tokens_seen": 1769537536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717575757575758,
      "loss": 2.4516,
      "theoretical_loss": 3.466003645499502,
      "tokens_seen": 1769603072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047173737373737374,
      "loss": 2.6413,
      "theoretical_loss": 3.46599268396725,
      "tokens_seen": 1769668608
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 976290,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.645479679107666,
      "objective/train/theoretical_loss": 3.465981722954586,
      "objective/train/tokens_used": 128593376,
      "theoretical_loss": 3.465981722954586,
      "tokens_seen": 1769734144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717171717171717,
      "loss": 2.7062,
      "theoretical_loss": 3.465981722954586,
      "tokens_seen": 1769734144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047169696969696966,
      "loss": 2.7589,
      "theoretical_loss": 3.465970762461466,
      "tokens_seen": 1769799680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716767676767677,
      "loss": 2.8116,
      "theoretical_loss": 3.4659598024878457,
      "tokens_seen": 1769865216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716565656565657,
      "loss": 2.8078,
      "theoretical_loss": 3.4659488430336816,
      "tokens_seen": 1769930752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047163636363636364,
      "loss": 2.2874,
      "theoretical_loss": 3.46593788409893,
      "tokens_seen": 1769996288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716161616161616,
      "loss": 2.2027,
      "theoretical_loss": 3.4659269256835468,
      "tokens_seen": 1770061824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715959595959596,
      "loss": 2.7392,
      "theoretical_loss": 3.4659159677874882,
      "tokens_seen": 1770127360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715757575757576,
      "loss": 2.8195,
      "theoretical_loss": 3.4659050104107103,
      "tokens_seen": 1770192896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715555555555556,
      "loss": 2.6863,
      "theoretical_loss": 3.4658940535531695,
      "tokens_seen": 1770258432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047153535353535353,
      "loss": 2.4416,
      "theoretical_loss": 3.465883097214822,
      "tokens_seen": 1770323968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715151515151515,
      "loss": 2.5736,
      "theoretical_loss": 3.4658721413956237,
      "tokens_seen": 1770389504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047149494949494955,
      "loss": 2.5487,
      "theoretical_loss": 3.465861186095531,
      "tokens_seen": 1770455040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714747474747475,
      "loss": 2.9009,
      "theoretical_loss": 3.4658502313144997,
      "tokens_seen": 1770520576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047145454545454547,
      "loss": 2.8561,
      "theoretical_loss": 3.4658392770524866,
      "tokens_seen": 1770586112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714343434343434,
      "loss": 2.646,
      "theoretical_loss": 3.4658283233094473,
      "tokens_seen": 1770651648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714141414141414,
      "loss": 2.6522,
      "theoretical_loss": 3.4658173700853387,
      "tokens_seen": 1770717184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047139393939393944,
      "loss": 2.7383,
      "theoretical_loss": 3.4658064173801164,
      "tokens_seen": 1770782720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713737373737374,
      "loss": 2.4971,
      "theoretical_loss": 3.465795465193737,
      "tokens_seen": 1770848256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047135353535353536,
      "loss": 2.8238,
      "theoretical_loss": 3.465784513526156,
      "tokens_seen": 1770913792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713333333333333,
      "loss": 2.5457,
      "theoretical_loss": 3.465773562377331,
      "tokens_seen": 1770979328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713131313131313,
      "loss": 2.7584,
      "theoretical_loss": 3.465762611747217,
      "tokens_seen": 1771044864
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047129292929292934,
      "loss": 2.5373,
      "theoretical_loss": 3.4657516616357706,
      "tokens_seen": 1771110400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712727272727273,
      "loss": 2.7518,
      "theoretical_loss": 3.4657407120429484,
      "tokens_seen": 1771175936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047125252525252525,
      "loss": 2.6246,
      "theoretical_loss": 3.4657297629687056,
      "tokens_seen": 1771241472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712323232323232,
      "loss": 2.6083,
      "theoretical_loss": 3.465718814413,
      "tokens_seen": 1771307008
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 976909,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4419713020324707,
      "objective/train/theoretical_loss": 3.4657078663757868,
      "objective/train/tokens_used": 130231776,
      "theoretical_loss": 3.4657078663757868,
      "tokens_seen": 1771372544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712121212121212,
      "loss": 2.683,
      "theoretical_loss": 3.4657078663757868,
      "tokens_seen": 1771372544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047119191919191923,
      "loss": 2.7196,
      "theoretical_loss": 3.4656969188570224,
      "tokens_seen": 1771438080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711717171717172,
      "loss": 2.5106,
      "theoretical_loss": 3.4656859718566633,
      "tokens_seen": 1771503616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047115151515151514,
      "loss": 2.4445,
      "theoretical_loss": 3.465675025374666,
      "tokens_seen": 1771569152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047113131313131316,
      "loss": 2.4538,
      "theoretical_loss": 3.4656640794109856,
      "tokens_seen": 1771634688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711111111111111,
      "loss": 2.6856,
      "theoretical_loss": 3.4656531339655796,
      "tokens_seen": 1771700224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710909090909091,
      "loss": 2.771,
      "theoretical_loss": 3.4656421890384044,
      "tokens_seen": 1771765760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710707070707071,
      "loss": 2.5379,
      "theoretical_loss": 3.4656312446294155,
      "tokens_seen": 1771831296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047105050505050504,
      "loss": 2.587,
      "theoretical_loss": 3.465620300738569,
      "tokens_seen": 1771896832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047103030303030305,
      "loss": 2.5692,
      "theoretical_loss": 3.4656093573658224,
      "tokens_seen": 1771962368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000471010101010101,
      "loss": 2.7043,
      "theoretical_loss": 3.465598414511131,
      "tokens_seen": 1772027904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470989898989899,
      "loss": 2.6211,
      "theoretical_loss": 3.4655874721744517,
      "tokens_seen": 1772093440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047096969696969697,
      "loss": 2.6399,
      "theoretical_loss": 3.46557653035574,
      "tokens_seen": 1772158976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470949494949495,
      "loss": 2.5832,
      "theoretical_loss": 3.465565589054954,
      "tokens_seen": 1772224512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047092929292929294,
      "loss": 2.6915,
      "theoretical_loss": 3.4655546482720476,
      "tokens_seen": 1772290048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709090909090909,
      "loss": 2.5416,
      "theoretical_loss": 3.4655437080069786,
      "tokens_seen": 1772355584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708888888888889,
      "loss": 2.6132,
      "theoretical_loss": 3.4655327682597035,
      "tokens_seen": 1772421120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047086868686868687,
      "loss": 2.4983,
      "theoretical_loss": 3.4655218290301786,
      "tokens_seen": 1772486656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708484848484849,
      "loss": 2.6079,
      "theoretical_loss": 3.465510890318359,
      "tokens_seen": 1772552192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047082828282828283,
      "loss": 2.7404,
      "theoretical_loss": 3.465499952124203,
      "tokens_seen": 1772617728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708080808080808,
      "loss": 2.7041,
      "theoretical_loss": 3.465489014447665,
      "tokens_seen": 1772683264
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707878787878788,
      "loss": 2.6753,
      "theoretical_loss": 3.4654780772887026,
      "tokens_seen": 1772748800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707676767676768,
      "loss": 2.8614,
      "theoretical_loss": 3.465467140647272,
      "tokens_seen": 1772814336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047074747474747477,
      "loss": 2.5925,
      "theoretical_loss": 3.4654562045233295,
      "tokens_seen": 1772879872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707272727272727,
      "loss": 2.7646,
      "theoretical_loss": 3.465445268916832,
      "tokens_seen": 1772945408
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 977811,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4676389694213867,
      "objective/train/theoretical_loss": 3.4654343338277345,
      "objective/train/tokens_used": 131870176,
      "theoretical_loss": 3.4654343338277345,
      "tokens_seen": 1773010944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707070707070707,
      "loss": 2.5732,
      "theoretical_loss": 3.4654343338277345,
      "tokens_seen": 1773010944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706868686868687,
      "loss": 2.8103,
      "theoretical_loss": 3.4654233992559944,
      "tokens_seen": 1773076480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706666666666667,
      "loss": 2.4505,
      "theoretical_loss": 3.465412465201568,
      "tokens_seen": 1773142016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047064646464646466,
      "loss": 2.3781,
      "theoretical_loss": 3.465401531664412,
      "tokens_seen": 1773207552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706262626262626,
      "loss": 2.8003,
      "theoretical_loss": 3.4653905986444817,
      "tokens_seen": 1773273088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706060606060606,
      "loss": 2.6426,
      "theoretical_loss": 3.465379666141735,
      "tokens_seen": 1773338624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047058585858585864,
      "loss": 2.5942,
      "theoretical_loss": 3.4653687341561277,
      "tokens_seen": 1773404160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705656565656566,
      "loss": 2.628,
      "theoretical_loss": 3.4653578026876155,
      "tokens_seen": 1773469696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047054545454545455,
      "loss": 2.6492,
      "theoretical_loss": 3.4653468717361555,
      "tokens_seen": 1773535232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705252525252525,
      "loss": 2.6411,
      "theoretical_loss": 3.4653359413017046,
      "tokens_seen": 1773600768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047050505050505047,
      "loss": 2.7541,
      "theoretical_loss": 3.465325011384219,
      "tokens_seen": 1773666304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047048484848484853,
      "loss": 2.4253,
      "theoretical_loss": 3.465314081983654,
      "tokens_seen": 1773731840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704646464646465,
      "loss": 2.7165,
      "theoretical_loss": 3.4653031530999674,
      "tokens_seen": 1773797376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047044444444444445,
      "loss": 2.7943,
      "theoretical_loss": 3.4652922247331155,
      "tokens_seen": 1773862912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704242424242424,
      "loss": 2.5949,
      "theoretical_loss": 3.465281296883054,
      "tokens_seen": 1773928448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704040404040404,
      "loss": 2.5194,
      "theoretical_loss": 3.4652703695497404,
      "tokens_seen": 1773993984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703838383838384,
      "loss": 2.8284,
      "theoretical_loss": 3.46525944273313,
      "tokens_seen": 1774059520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703636363636364,
      "loss": 2.5426,
      "theoretical_loss": 3.4652485164331805,
      "tokens_seen": 1774125056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047034343434343434,
      "loss": 2.6549,
      "theoretical_loss": 3.4652375906498474,
      "tokens_seen": 1774190592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703232323232323,
      "loss": 2.5567,
      "theoretical_loss": 3.4652266653830877,
      "tokens_seen": 1774256128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703030303030303,
      "loss": 2.6891,
      "theoretical_loss": 3.465215740632858,
      "tokens_seen": 1774321664
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702828282828283,
      "loss": 2.6164,
      "theoretical_loss": 3.465204816399114,
      "tokens_seen": 1774387200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702626262626263,
      "loss": 2.6243,
      "theoretical_loss": 3.465193892681813,
      "tokens_seen": 1774452736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047024242424242423,
      "loss": 2.6201,
      "theoretical_loss": 3.4651829694809115,
      "tokens_seen": 1774518272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047022222222222224,
      "loss": 2.6683,
      "theoretical_loss": 3.465172046796366,
      "tokens_seen": 1774583808
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 978517,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.493600368499756,
      "objective/train/theoretical_loss": 3.4651611246281324,
      "objective/train/tokens_used": 133508576,
      "theoretical_loss": 3.4651611246281324,
      "tokens_seen": 1774649344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702020202020202,
      "loss": 2.5756,
      "theoretical_loss": 3.4651611246281324,
      "tokens_seen": 1774649344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701818181818182,
      "loss": 2.7666,
      "theoretical_loss": 3.465150202976168,
      "tokens_seen": 1774714880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047016161616161617,
      "loss": 2.6612,
      "theoretical_loss": 3.465139281840429,
      "tokens_seen": 1774780416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701414141414141,
      "loss": 2.5653,
      "theoretical_loss": 3.465128361220872,
      "tokens_seen": 1774845952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047012121212121214,
      "loss": 2.6465,
      "theoretical_loss": 3.465117441117453,
      "tokens_seen": 1774911488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701010101010101,
      "loss": 2.5144,
      "theoretical_loss": 3.4651065215301298,
      "tokens_seen": 1774977024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700808080808081,
      "loss": 2.7176,
      "theoretical_loss": 3.4650956024588577,
      "tokens_seen": 1775042560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047006060606060606,
      "loss": 2.7295,
      "theoretical_loss": 3.4650846839035943,
      "tokens_seen": 1775108096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047004040404040407,
      "loss": 2.5483,
      "theoretical_loss": 3.465073765864295,
      "tokens_seen": 1775173632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047002020202020203,
      "loss": 2.5907,
      "theoretical_loss": 3.465062848340917,
      "tokens_seen": 1775239168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047,
      "loss": 2.6084,
      "theoretical_loss": 3.4650519313334174,
      "tokens_seen": 1775304704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469979797979798,
      "loss": 2.4807,
      "theoretical_loss": 3.465041014841751,
      "tokens_seen": 1775370240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046995959595959595,
      "loss": 2.6201,
      "theoretical_loss": 3.465030098865877,
      "tokens_seen": 1775435776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046993939393939396,
      "loss": 2.4878,
      "theoretical_loss": 3.4650191834057504,
      "tokens_seen": 1775501312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699191919191919,
      "loss": 2.6224,
      "theoretical_loss": 3.465008268461328,
      "tokens_seen": 1775566848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698989898989899,
      "loss": 2.653,
      "theoretical_loss": 3.4649973540325663,
      "tokens_seen": 1775632384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698787878787879,
      "loss": 2.5147,
      "theoretical_loss": 3.4649864401194215,
      "tokens_seen": 1775697920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698585858585859,
      "loss": 2.4981,
      "theoretical_loss": 3.4649755267218514,
      "tokens_seen": 1775763456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046983838383838386,
      "loss": 2.6706,
      "theoretical_loss": 3.464964613839812,
      "tokens_seen": 1775828992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698181818181818,
      "loss": 2.6469,
      "theoretical_loss": 3.4649537014732594,
      "tokens_seen": 1775894528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697979797979798,
      "loss": 2.6877,
      "theoretical_loss": 3.464942789622151,
      "tokens_seen": 1775960064
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697777777777778,
      "loss": 2.6401,
      "theoretical_loss": 3.464931878286443,
      "tokens_seen": 1776025600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697575757575758,
      "loss": 2.7416,
      "theoretical_loss": 3.4649209674660923,
      "tokens_seen": 1776091136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046973737373737375,
      "loss": 2.716,
      "theoretical_loss": 3.464910057161056,
      "tokens_seen": 1776156672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697171717171717,
      "loss": 2.775,
      "theoretical_loss": 3.4648991473712893,
      "tokens_seen": 1776222208
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 979554,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.655245542526245,
      "objective/train/theoretical_loss": 3.4648882380967505,
      "objective/train/tokens_used": 135146976,
      "theoretical_loss": 3.4648882380967505,
      "tokens_seen": 1776287744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696969696969697,
      "loss": 2.735,
      "theoretical_loss": 3.4648882380967505,
      "tokens_seen": 1776287744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046967676767676773,
      "loss": 2.7324,
      "theoretical_loss": 3.464877329337395,
      "tokens_seen": 1776353280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696565656565657,
      "loss": 2.6467,
      "theoretical_loss": 3.4648664210931797,
      "tokens_seen": 1776418816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046963636363636364,
      "loss": 2.6666,
      "theoretical_loss": 3.4648555133640624,
      "tokens_seen": 1776484352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696161616161616,
      "loss": 2.694,
      "theoretical_loss": 3.464844606149998,
      "tokens_seen": 1776549888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695959595959596,
      "loss": 2.5421,
      "theoretical_loss": 3.4648336994509448,
      "tokens_seen": 1776615424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695757575757576,
      "loss": 2.4765,
      "theoretical_loss": 3.4648227932668583,
      "tokens_seen": 1776680960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695555555555556,
      "loss": 2.7016,
      "theoretical_loss": 3.464811887597696,
      "tokens_seen": 1776746496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046953535353535354,
      "loss": 2.3606,
      "theoretical_loss": 3.464800982443414,
      "tokens_seen": 1776812032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695151515151515,
      "loss": 2.7383,
      "theoretical_loss": 3.4647900778039693,
      "tokens_seen": 1776877568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694949494949495,
      "loss": 2.5567,
      "theoretical_loss": 3.464779173679319,
      "tokens_seen": 1776943104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694747474747475,
      "loss": 2.6411,
      "theoretical_loss": 3.4647682700694187,
      "tokens_seen": 1777008640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046945454545454547,
      "loss": 2.4005,
      "theoretical_loss": 3.464757366974226,
      "tokens_seen": 1777074176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046943434343434343,
      "loss": 2.7334,
      "theoretical_loss": 3.4647464643936976,
      "tokens_seen": 1777139712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694141414141414,
      "loss": 2.5828,
      "theoretical_loss": 3.4647355623277902,
      "tokens_seen": 1777205248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046939393939393945,
      "loss": 2.3715,
      "theoretical_loss": 3.4647246607764606,
      "tokens_seen": 1777270784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693737373737374,
      "loss": 2.6547,
      "theoretical_loss": 3.4647137597396647,
      "tokens_seen": 1777336320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046935353535353536,
      "loss": 2.4004,
      "theoretical_loss": 3.4647028592173603,
      "tokens_seen": 1777401856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693333333333333,
      "loss": 2.7303,
      "theoretical_loss": 3.464691959209503,
      "tokens_seen": 1777467392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693131313131313,
      "loss": 2.5245,
      "theoretical_loss": 3.4646810597160513,
      "tokens_seen": 1777532928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046929292929292934,
      "loss": 2.6391,
      "theoretical_loss": 3.46467016073696,
      "tokens_seen": 1777598464
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692727272727273,
      "loss": 2.6976,
      "theoretical_loss": 3.4646592622721872,
      "tokens_seen": 1777664000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046925252525252526,
      "loss": 2.5603,
      "theoretical_loss": 3.4646483643216897,
      "tokens_seen": 1777729536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692323232323232,
      "loss": 2.4618,
      "theoretical_loss": 3.4646374668854234,
      "tokens_seen": 1777795072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692121212121212,
      "loss": 2.8756,
      "theoretical_loss": 3.4646265699633454,
      "tokens_seen": 1777860608
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 979848,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.122527599334717,
      "objective/train/theoretical_loss": 3.464615673555413,
      "objective/train/tokens_used": 136785376,
      "theoretical_loss": 3.464615673555413,
      "tokens_seen": 1777926144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046919191919191924,
      "loss": 2.7605,
      "theoretical_loss": 3.464615673555413,
      "tokens_seen": 1777926144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691717171717172,
      "loss": 2.6706,
      "theoretical_loss": 3.4646047776615823,
      "tokens_seen": 1777991680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046915151515151515,
      "loss": 2.5914,
      "theoretical_loss": 3.4645938822818105,
      "tokens_seen": 1778057216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691313131313131,
      "loss": 2.8862,
      "theoretical_loss": 3.4645829874160543,
      "tokens_seen": 1778122752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691111111111111,
      "loss": 2.6125,
      "theoretical_loss": 3.4645720930642705,
      "tokens_seen": 1778188288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046909090909090913,
      "loss": 2.6488,
      "theoretical_loss": 3.4645611992264156,
      "tokens_seen": 1778253824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690707070707071,
      "loss": 2.5569,
      "theoretical_loss": 3.4645503059024474,
      "tokens_seen": 1778319360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046905050505050504,
      "loss": 2.4418,
      "theoretical_loss": 3.4645394130923215,
      "tokens_seen": 1778384896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046903030303030305,
      "loss": 2.6375,
      "theoretical_loss": 3.4645285207959953,
      "tokens_seen": 1778450432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469010101010101,
      "loss": 2.5087,
      "theoretical_loss": 3.464517629013426,
      "tokens_seen": 1778515968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468989898989899,
      "loss": 2.4809,
      "theoretical_loss": 3.4645067377445695,
      "tokens_seen": 1778581504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468969696969697,
      "loss": 2.696,
      "theoretical_loss": 3.464495846989384,
      "tokens_seen": 1778647040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046894949494949494,
      "loss": 2.7445,
      "theoretical_loss": 3.4644849567478246,
      "tokens_seen": 1778712576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046892929292929295,
      "loss": 2.7525,
      "theoretical_loss": 3.4644740670198497,
      "tokens_seen": 1778778112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689090909090909,
      "loss": 2.6765,
      "theoretical_loss": 3.464463177805415,
      "tokens_seen": 1778843648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688888888888889,
      "loss": 2.6821,
      "theoretical_loss": 3.4644522891044787,
      "tokens_seen": 1778909184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046886868686868687,
      "loss": 2.8012,
      "theoretical_loss": 3.4644414009169964,
      "tokens_seen": 1778974720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688484848484849,
      "loss": 2.573,
      "theoretical_loss": 3.4644305132429256,
      "tokens_seen": 1779040256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046882828282828284,
      "loss": 2.8178,
      "theoretical_loss": 3.464419626082223,
      "tokens_seen": 1779105792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688080808080808,
      "loss": 2.7468,
      "theoretical_loss": 3.4644087394348455,
      "tokens_seen": 1779171328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687878787878788,
      "loss": 2.5975,
      "theoretical_loss": 3.46439785330075,
      "tokens_seen": 1779236864
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046876767676767676,
      "loss": 2.7253,
      "theoretical_loss": 3.4643869676798937,
      "tokens_seen": 1779302400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687474747474748,
      "loss": 2.6003,
      "theoretical_loss": 3.464376082572233,
      "tokens_seen": 1779367936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046872727272727273,
      "loss": 2.6038,
      "theoretical_loss": 3.464365197977725,
      "tokens_seen": 1779433472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687070707070707,
      "loss": 2.7135,
      "theoretical_loss": 3.464354313896327,
      "tokens_seen": 1779499008
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 981168,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4408981800079346,
      "objective/train/theoretical_loss": 3.464343430327995,
      "objective/train/tokens_used": 138423776,
      "theoretical_loss": 3.464343430327995,
      "tokens_seen": 1779564544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686868686868687,
      "loss": 2.6659,
      "theoretical_loss": 3.464343430327995,
      "tokens_seen": 1779564544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686666666666667,
      "loss": 2.545,
      "theoretical_loss": 3.4643325472726865,
      "tokens_seen": 1779630080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046864646464646467,
      "loss": 2.5786,
      "theoretical_loss": 3.4643216647303583,
      "tokens_seen": 1779695616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686262626262626,
      "loss": 2.6326,
      "theoretical_loss": 3.4643107827009674,
      "tokens_seen": 1779761152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686060606060606,
      "loss": 2.476,
      "theoretical_loss": 3.4642999011844715,
      "tokens_seen": 1779826688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685858585858586,
      "loss": 2.5703,
      "theoretical_loss": 3.464289020180826,
      "tokens_seen": 1779892224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685656565656566,
      "loss": 2.5252,
      "theoretical_loss": 3.464278139689989,
      "tokens_seen": 1779957760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046854545454545456,
      "loss": 2.653,
      "theoretical_loss": 3.4642672597119173,
      "tokens_seen": 1780023296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685252525252525,
      "loss": 2.5044,
      "theoretical_loss": 3.4642563802465673,
      "tokens_seen": 1780088832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685050505050505,
      "loss": 2.5998,
      "theoretical_loss": 3.464245501293896,
      "tokens_seen": 1780154368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046848484848484854,
      "loss": 2.5963,
      "theoretical_loss": 3.464234622853861,
      "tokens_seen": 1780219904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684646464646465,
      "loss": 2.6996,
      "theoretical_loss": 3.4642237449264193,
      "tokens_seen": 1780285440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046844444444444445,
      "loss": 2.702,
      "theoretical_loss": 3.4642128675115273,
      "tokens_seen": 1780350976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684242424242424,
      "loss": 2.7233,
      "theoretical_loss": 3.4642019906091424,
      "tokens_seen": 1780416512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046840404040404037,
      "loss": 2.7893,
      "theoretical_loss": 3.464191114219221,
      "tokens_seen": 1780482048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046838383838383843,
      "loss": 2.4928,
      "theoretical_loss": 3.464180238341721,
      "tokens_seen": 1780547584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683636363636364,
      "loss": 2.7807,
      "theoretical_loss": 3.4641693629765986,
      "tokens_seen": 1780613120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046834343434343435,
      "loss": 2.4966,
      "theoretical_loss": 3.464158488123811,
      "tokens_seen": 1780678656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683232323232323,
      "loss": 2.6617,
      "theoretical_loss": 3.4641476137833154,
      "tokens_seen": 1780744192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046830303030303026,
      "loss": 2.8356,
      "theoretical_loss": 3.464136739955069,
      "tokens_seen": 1780809728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682828282828283,
      "loss": 2.577,
      "theoretical_loss": 3.464125866639028,
      "tokens_seen": 1780875264
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682626262626263,
      "loss": 2.6046,
      "theoretical_loss": 3.4641149938351505,
      "tokens_seen": 1780940800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046824242424242424,
      "loss": 2.7569,
      "theoretical_loss": 3.464104121543393,
      "tokens_seen": 1781006336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682222222222222,
      "loss": 2.6343,
      "theoretical_loss": 3.4640932497637125,
      "tokens_seen": 1781071872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682020202020202,
      "loss": 2.9327,
      "theoretical_loss": 3.464082378496066,
      "tokens_seen": 1781137408
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 982366,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5046448707580566,
      "objective/train/theoretical_loss": 3.4640715077404103,
      "objective/train/tokens_used": 140062176,
      "theoretical_loss": 3.4640715077404103,
      "tokens_seen": 1781202944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681818181818182,
      "loss": 2.4701,
      "theoretical_loss": 3.4640715077404103,
      "tokens_seen": 1781202944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681616161616162,
      "loss": 2.5085,
      "theoretical_loss": 3.4640606374967033,
      "tokens_seen": 1781268480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046814141414141413,
      "loss": 2.7434,
      "theoretical_loss": 3.464049767764901,
      "tokens_seen": 1781334016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681212121212121,
      "loss": 2.681,
      "theoretical_loss": 3.4640388985449615,
      "tokens_seen": 1781399552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046810101010101015,
      "loss": 2.8738,
      "theoretical_loss": 3.4640280298368413,
      "tokens_seen": 1781465088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680808080808081,
      "loss": 2.5953,
      "theoretical_loss": 3.4640171616404976,
      "tokens_seen": 1781530624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046806060606060607,
      "loss": 2.8082,
      "theoretical_loss": 3.4640062939558867,
      "tokens_seen": 1781596160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468040404040404,
      "loss": 2.6266,
      "theoretical_loss": 3.4639954267829673,
      "tokens_seen": 1781661696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046802020202020203,
      "loss": 2.7488,
      "theoretical_loss": 3.4639845601216948,
      "tokens_seen": 1781727232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046800000000000005,
      "loss": 2.6267,
      "theoretical_loss": 3.4639736939720276,
      "tokens_seen": 1781792768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467979797979798,
      "loss": 2.8528,
      "theoretical_loss": 3.4639628283339223,
      "tokens_seen": 1781858304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046795959595959596,
      "loss": 2.4593,
      "theoretical_loss": 3.463951963207336,
      "tokens_seen": 1781923840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679393939393939,
      "loss": 2.7433,
      "theoretical_loss": 3.463941098592225,
      "tokens_seen": 1781989376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046791919191919193,
      "loss": 2.7575,
      "theoretical_loss": 3.463930234488548,
      "tokens_seen": 1782054912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046789898989898994,
      "loss": 2.8066,
      "theoretical_loss": 3.4639193708962615,
      "tokens_seen": 1782120448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678787878787879,
      "loss": 2.6911,
      "theoretical_loss": 3.4639085078153222,
      "tokens_seen": 1782185984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046785858585858585,
      "loss": 2.5327,
      "theoretical_loss": 3.4638976452456873,
      "tokens_seen": 1782251520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046783838383838386,
      "loss": 2.713,
      "theoretical_loss": 3.4638867831873137,
      "tokens_seen": 1782317056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678181818181818,
      "loss": 2.7197,
      "theoretical_loss": 3.4638759216401596,
      "tokens_seen": 1782382592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046779797979797983,
      "loss": 2.7073,
      "theoretical_loss": 3.463865060604181,
      "tokens_seen": 1782448128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677777777777778,
      "loss": 2.5768,
      "theoretical_loss": 3.4638542000793358,
      "tokens_seen": 1782513664
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046775757575757574,
      "loss": 2.7761,
      "theoretical_loss": 3.463843340065581,
      "tokens_seen": 1782579200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046773737373737376,
      "loss": 2.67,
      "theoretical_loss": 3.463832480562873,
      "tokens_seen": 1782644736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677171717171717,
      "loss": 2.6575,
      "theoretical_loss": 3.463821621571171,
      "tokens_seen": 1782710272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676969696969697,
      "loss": 2.5326,
      "theoretical_loss": 3.4638107630904296,
      "tokens_seen": 1782775808
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 982914,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4014761447906494,
      "objective/train/theoretical_loss": 3.4637999051206076,
      "objective/train/tokens_used": 141700576,
      "theoretical_loss": 3.4637999051206076,
      "tokens_seen": 1782841344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676767676767677,
      "loss": 2.5757,
      "theoretical_loss": 3.4637999051206076,
      "tokens_seen": 1782841344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676565656565657,
      "loss": 2.7185,
      "theoretical_loss": 3.4637890476616615,
      "tokens_seen": 1782906880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046763636363636365,
      "loss": 2.4657,
      "theoretical_loss": 3.4637781907135485,
      "tokens_seen": 1782972416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676161616161616,
      "loss": 2.753,
      "theoretical_loss": 3.4637673342762265,
      "tokens_seen": 1783037952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675959595959596,
      "loss": 2.6718,
      "theoretical_loss": 3.463756478349652,
      "tokens_seen": 1783103488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675757575757576,
      "loss": 2.8216,
      "theoretical_loss": 3.463745622933782,
      "tokens_seen": 1783169024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675555555555556,
      "loss": 2.5559,
      "theoretical_loss": 3.4637347680285746,
      "tokens_seen": 1783234560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046753535353535354,
      "loss": 2.7446,
      "theoretical_loss": 3.463723913633987,
      "tokens_seen": 1783300096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675151515151515,
      "loss": 2.7602,
      "theoretical_loss": 3.463713059749975,
      "tokens_seen": 1783365632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674949494949495,
      "loss": 2.5877,
      "theoretical_loss": 3.4637022063764977,
      "tokens_seen": 1783431168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674747474747475,
      "loss": 2.7978,
      "theoretical_loss": 3.4636913535135108,
      "tokens_seen": 1783496704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674545454545455,
      "loss": 2.547,
      "theoretical_loss": 3.4636805011609724,
      "tokens_seen": 1783562240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046743434343434343,
      "loss": 2.8632,
      "theoretical_loss": 3.463669649318839,
      "tokens_seen": 1783627776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674141414141414,
      "loss": 2.57,
      "theoretical_loss": 3.4636587979870685,
      "tokens_seen": 1783693312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673939393939394,
      "loss": 2.5764,
      "theoretical_loss": 3.463647947165618,
      "tokens_seen": 1783758848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673737373737374,
      "loss": 2.9032,
      "theoretical_loss": 3.4636370968544448,
      "tokens_seen": 1783824384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046735353535353537,
      "loss": 2.7317,
      "theoretical_loss": 3.463626247053506,
      "tokens_seen": 1783889920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673333333333333,
      "loss": 2.6338,
      "theoretical_loss": 3.4636153977627586,
      "tokens_seen": 1783955456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673131313131313,
      "loss": 2.7822,
      "theoretical_loss": 3.4636045489821607,
      "tokens_seen": 1784020992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046729292929292935,
      "loss": 2.497,
      "theoretical_loss": 3.4635937007116686,
      "tokens_seen": 1784086528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672727272727273,
      "loss": 2.509,
      "theoretical_loss": 3.4635828529512405,
      "tokens_seen": 1784152064
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046725252525252526,
      "loss": 2.6947,
      "theoretical_loss": 3.4635720057008332,
      "tokens_seen": 1784217600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672323232323232,
      "loss": 2.5064,
      "theoretical_loss": 3.463561158960404,
      "tokens_seen": 1784283136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672121212121212,
      "loss": 2.3541,
      "theoretical_loss": 3.4635503127299097,
      "tokens_seen": 1784348672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046719191919191924,
      "loss": 2.5577,
      "theoretical_loss": 3.4635394670093085,
      "tokens_seen": 1784414208
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 984019,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.804720401763916,
      "objective/train/theoretical_loss": 3.4635286217985572,
      "objective/train/tokens_used": 143338976,
      "theoretical_loss": 3.4635286217985572,
      "tokens_seen": 1784479744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004671717171717172,
      "loss": 2.8734,
      "theoretical_loss": 3.4635286217985572,
      "tokens_seen": 1784479744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046715151515151516,
      "loss": 2.841,
      "theoretical_loss": 3.4635177770976133,
      "tokens_seen": 1784545280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671313131313131,
      "loss": 2.4586,
      "theoretical_loss": 3.463506932906434,
      "tokens_seen": 1784610816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671111111111111,
      "loss": 2.3573,
      "theoretical_loss": 3.463496089224977,
      "tokens_seen": 1784676352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046709090909090913,
      "loss": 2.7561,
      "theoretical_loss": 3.4634852460531986,
      "tokens_seen": 1784741888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670707070707071,
      "loss": 2.8304,
      "theoretical_loss": 3.463474403391057,
      "tokens_seen": 1784807424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046705050505050505,
      "loss": 2.566,
      "theoretical_loss": 3.46346356123851,
      "tokens_seen": 1784872960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000467030303030303,
      "loss": 2.7447,
      "theoretical_loss": 3.4634527195955136,
      "tokens_seen": 1784938496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000467010101010101,
      "loss": 2.6607,
      "theoretical_loss": 3.463441878462026,
      "tokens_seen": 1785004032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466989898989899,
      "loss": 2.4519,
      "theoretical_loss": 3.463431037838004,
      "tokens_seen": 1785069568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466969696969697,
      "loss": 2.7172,
      "theoretical_loss": 3.463420197723406,
      "tokens_seen": 1785135104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046694949494949494,
      "loss": 2.4123,
      "theoretical_loss": 3.463409358118188,
      "tokens_seen": 1785200640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046692929292929295,
      "loss": 2.7201,
      "theoretical_loss": 3.4633985190223084,
      "tokens_seen": 1785266176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669090909090909,
      "loss": 2.7307,
      "theoretical_loss": 3.4633876804357246,
      "tokens_seen": 1785331712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668888888888889,
      "loss": 2.5424,
      "theoretical_loss": 3.463376842358393,
      "tokens_seen": 1785397248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668686868686869,
      "loss": 2.8041,
      "theoretical_loss": 3.463366004790272,
      "tokens_seen": 1785462784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046684848484848483,
      "loss": 2.7058,
      "theoretical_loss": 3.4633551677313186,
      "tokens_seen": 1785528320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046682828282828284,
      "loss": 2.6283,
      "theoretical_loss": 3.46334433118149,
      "tokens_seen": 1785593856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668080808080808,
      "loss": 2.4156,
      "theoretical_loss": 3.463333495140744,
      "tokens_seen": 1785659392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667878787878788,
      "loss": 2.4195,
      "theoretical_loss": 3.4633226596090374,
      "tokens_seen": 1785724928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046676767676767677,
      "loss": 2.4871,
      "theoretical_loss": 3.463311824586328,
      "tokens_seen": 1785790464
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667474747474748,
      "loss": 2.4098,
      "theoretical_loss": 3.4633009900725735,
      "tokens_seen": 1785856000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046672727272727274,
      "loss": 2.8379,
      "theoretical_loss": 3.463290156067731,
      "tokens_seen": 1785921536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667070707070707,
      "loss": 2.5884,
      "theoretical_loss": 3.4632793225717573,
      "tokens_seen": 1785987072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666868686868687,
      "loss": 2.7167,
      "theoretical_loss": 3.463268489584611,
      "tokens_seen": 1786052608
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 984652,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4918596744537354,
      "objective/train/theoretical_loss": 3.463257657106249,
      "objective/train/tokens_used": 144977376,
      "theoretical_loss": 3.463257657106249,
      "tokens_seen": 1786118144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046666666666666666,
      "loss": 2.6791,
      "theoretical_loss": 3.463257657106249,
      "tokens_seen": 1786118144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666464646464647,
      "loss": 2.5787,
      "theoretical_loss": 3.463246825136628,
      "tokens_seen": 1786183680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046662626262626263,
      "loss": 2.4831,
      "theoretical_loss": 3.463235993675707,
      "tokens_seen": 1786249216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666060606060606,
      "loss": 2.6653,
      "theoretical_loss": 3.4632251627234423,
      "tokens_seen": 1786314752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665858585858586,
      "loss": 2.604,
      "theoretical_loss": 3.4632143322797915,
      "tokens_seen": 1786380288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665656565656566,
      "loss": 2.458,
      "theoretical_loss": 3.4632035023447125,
      "tokens_seen": 1786445824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046654545454545457,
      "loss": 2.6402,
      "theoretical_loss": 3.4631926729181624,
      "tokens_seen": 1786511360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665252525252525,
      "loss": 2.5807,
      "theoretical_loss": 3.4631818440000988,
      "tokens_seen": 1786576896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665050505050505,
      "loss": 2.7028,
      "theoretical_loss": 3.4631710155904787,
      "tokens_seen": 1786642432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664848484848485,
      "loss": 2.6881,
      "theoretical_loss": 3.4631601876892604,
      "tokens_seen": 1786707968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664646464646465,
      "loss": 2.6738,
      "theoretical_loss": 3.4631493602964007,
      "tokens_seen": 1786773504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046644444444444446,
      "loss": 2.5745,
      "theoretical_loss": 3.4631385334118576,
      "tokens_seen": 1786839040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664242424242424,
      "loss": 2.6133,
      "theoretical_loss": 3.4631277070355884,
      "tokens_seen": 1786904576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046640404040404037,
      "loss": 2.6516,
      "theoretical_loss": 3.463116881167551,
      "tokens_seen": 1786970112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663838383838384,
      "loss": 2.6627,
      "theoretical_loss": 3.4631060558077014,
      "tokens_seen": 1787035648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663636363636364,
      "loss": 2.6789,
      "theoretical_loss": 3.463095230955999,
      "tokens_seen": 1787101184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046634343434343435,
      "loss": 2.7396,
      "theoretical_loss": 3.4630844066124,
      "tokens_seen": 1787166720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663232323232323,
      "loss": 2.5693,
      "theoretical_loss": 3.463073582776863,
      "tokens_seen": 1787232256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663030303030303,
      "loss": 2.8965,
      "theoretical_loss": 3.463062759449344,
      "tokens_seen": 1787297792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046628282828282833,
      "loss": 2.7113,
      "theoretical_loss": 3.4630519366298023,
      "tokens_seen": 1787363328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662626262626263,
      "loss": 2.7535,
      "theoretical_loss": 3.4630411143181945,
      "tokens_seen": 1787428864
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046624242424242424,
      "loss": 2.6914,
      "theoretical_loss": 3.4630302925144782,
      "tokens_seen": 1787494400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662222222222222,
      "loss": 2.4684,
      "theoretical_loss": 3.463019471218611,
      "tokens_seen": 1787559936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662020202020202,
      "loss": 2.4305,
      "theoretical_loss": 3.4630086504305497,
      "tokens_seen": 1787625472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661818181818182,
      "loss": 2.5705,
      "theoretical_loss": 3.462997830150253,
      "tokens_seen": 1787691008
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 985931,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.233638286590576,
      "objective/train/theoretical_loss": 3.4629870103776783,
      "objective/train/tokens_used": 146615776,
      "theoretical_loss": 3.4629870103776783,
      "tokens_seen": 1787756544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661616161616162,
      "loss": 2.5084,
      "theoretical_loss": 3.4629870103776783,
      "tokens_seen": 1787756544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046614141414141414,
      "loss": 2.4923,
      "theoretical_loss": 3.462976191112783,
      "tokens_seen": 1787822080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661212121212121,
      "loss": 2.4434,
      "theoretical_loss": 3.4629653723555247,
      "tokens_seen": 1787887616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046610101010101016,
      "loss": 2.669,
      "theoretical_loss": 3.4629545541058606,
      "tokens_seen": 1787953152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660808080808081,
      "loss": 2.6602,
      "theoretical_loss": 3.4629437363637483,
      "tokens_seen": 1788018688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046606060606060607,
      "loss": 2.6326,
      "theoretical_loss": 3.462932919129146,
      "tokens_seen": 1788084224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046604040404040403,
      "loss": 2.5248,
      "theoretical_loss": 3.4629221024020107,
      "tokens_seen": 1788149760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466020202020202,
      "loss": 2.4801,
      "theoretical_loss": 3.4629112861823,
      "tokens_seen": 1788215296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046600000000000005,
      "loss": 2.6761,
      "theoretical_loss": 3.4629004704699726,
      "tokens_seen": 1788280832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465979797979798,
      "loss": 2.6667,
      "theoretical_loss": 3.462889655264984,
      "tokens_seen": 1788346368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046595959595959597,
      "loss": 2.8018,
      "theoretical_loss": 3.462878840567294,
      "tokens_seen": 1788411904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659393939393939,
      "loss": 2.7818,
      "theoretical_loss": 3.4628680263768588,
      "tokens_seen": 1788477440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046591919191919193,
      "loss": 2.6761,
      "theoretical_loss": 3.4628572126936366,
      "tokens_seen": 1788542976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046589898989898994,
      "loss": 2.5054,
      "theoretical_loss": 3.4628463995175847,
      "tokens_seen": 1788608512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658787878787879,
      "loss": 2.6396,
      "theoretical_loss": 3.4628355868486613,
      "tokens_seen": 1788674048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046585858585858586,
      "loss": 2.595,
      "theoretical_loss": 3.4628247746868235,
      "tokens_seen": 1788739584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658383838383838,
      "loss": 2.6262,
      "theoretical_loss": 3.462813963032029,
      "tokens_seen": 1788805120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658181818181818,
      "loss": 2.6693,
      "theoretical_loss": 3.4628031518842355,
      "tokens_seen": 1788870656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046579797979797984,
      "loss": 2.5535,
      "theoretical_loss": 3.462792341243401,
      "tokens_seen": 1788936192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657777777777778,
      "loss": 2.4625,
      "theoretical_loss": 3.4627815311094823,
      "tokens_seen": 1789001728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046575757575757575,
      "loss": 2.4874,
      "theoretical_loss": 3.462770721482438,
      "tokens_seen": 1789067264
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046573737373737376,
      "loss": 2.3284,
      "theoretical_loss": 3.4627599123622255,
      "tokens_seen": 1789132800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657171717171717,
      "loss": 2.6288,
      "theoretical_loss": 3.462749103748802,
      "tokens_seen": 1789198336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046569696969696973,
      "loss": 2.4654,
      "theoretical_loss": 3.4627382956421258,
      "tokens_seen": 1789263872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656767676767677,
      "loss": 2.5071,
      "theoretical_loss": 3.4627274880421544,
      "tokens_seen": 1789329408
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 986490,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.360241413116455,
      "objective/train/theoretical_loss": 3.462716680948845,
      "objective/train/tokens_used": 148254176,
      "theoretical_loss": 3.462716680948845,
      "tokens_seen": 1789394944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046565656565656564,
      "loss": 2.8045,
      "theoretical_loss": 3.462716680948845,
      "tokens_seen": 1789394944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046563636363636365,
      "loss": 2.7073,
      "theoretical_loss": 3.462705874362156,
      "tokens_seen": 1789460480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656161616161616,
      "loss": 2.4757,
      "theoretical_loss": 3.462695068282045,
      "tokens_seen": 1789526016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655959595959596,
      "loss": 2.5969,
      "theoretical_loss": 3.4626842627084695,
      "tokens_seen": 1789591552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655757575757576,
      "loss": 2.7488,
      "theoretical_loss": 3.4626734576413867,
      "tokens_seen": 1789657088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655555555555556,
      "loss": 2.6564,
      "theoretical_loss": 3.4626626530807547,
      "tokens_seen": 1789722624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046553535353535355,
      "loss": 2.5831,
      "theoretical_loss": 3.462651849026532,
      "tokens_seen": 1789788160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655151515151515,
      "loss": 2.6404,
      "theoretical_loss": 3.4626410454786756,
      "tokens_seen": 1789853696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654949494949495,
      "loss": 2.4886,
      "theoretical_loss": 3.462630242437143,
      "tokens_seen": 1789919232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046547474747474747,
      "loss": 2.7851,
      "theoretical_loss": 3.462619439901892,
      "tokens_seen": 1789984768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654545454545455,
      "loss": 2.4352,
      "theoretical_loss": 3.462608637872881,
      "tokens_seen": 1790050304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046543434343434344,
      "loss": 2.5899,
      "theoretical_loss": 3.462597836350067,
      "tokens_seen": 1790115840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654141414141414,
      "loss": 2.4442,
      "theoretical_loss": 3.4625870353334083,
      "tokens_seen": 1790181376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653939393939394,
      "loss": 2.4023,
      "theoretical_loss": 3.462576234822863,
      "tokens_seen": 1790246912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653737373737374,
      "loss": 2.5182,
      "theoretical_loss": 3.462565434818387,
      "tokens_seen": 1790312448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653535353535354,
      "loss": 2.6444,
      "theoretical_loss": 3.46255463531994,
      "tokens_seen": 1790377984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046533333333333333,
      "loss": 2.5226,
      "theoretical_loss": 3.4625438363274794,
      "tokens_seen": 1790443520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653131313131313,
      "loss": 2.7895,
      "theoretical_loss": 3.462533037840962,
      "tokens_seen": 1790509056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652929292929293,
      "loss": 2.5776,
      "theoretical_loss": 3.4625222398603466,
      "tokens_seen": 1790574592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652727272727273,
      "loss": 2.5917,
      "theoretical_loss": 3.4625114423855905,
      "tokens_seen": 1790640128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046525252525252527,
      "loss": 2.7743,
      "theoretical_loss": 3.4625006454166516,
      "tokens_seen": 1790705664
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652323232323232,
      "loss": 2.7714,
      "theoretical_loss": 3.4624898489534877,
      "tokens_seen": 1790771200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652121212121212,
      "loss": 2.5768,
      "theoretical_loss": 3.4624790529960574,
      "tokens_seen": 1790836736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046519191919191925,
      "loss": 2.751,
      "theoretical_loss": 3.462468257544317,
      "tokens_seen": 1790902272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651717171717172,
      "loss": 2.5792,
      "theoretical_loss": 3.4624574625982247,
      "tokens_seen": 1790967808
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 987679,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5805740356445312,
      "objective/train/theoretical_loss": 3.462446668157739,
      "objective/train/tokens_used": 149892576,
      "theoretical_loss": 3.462446668157739,
      "tokens_seen": 1791033344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046515151515151516,
      "loss": 2.4389,
      "theoretical_loss": 3.462446668157739,
      "tokens_seen": 1791033344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651313131313131,
      "loss": 2.5047,
      "theoretical_loss": 3.4624358742228174,
      "tokens_seen": 1791098880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651111111111111,
      "loss": 2.6406,
      "theoretical_loss": 3.462425080793418,
      "tokens_seen": 1791164416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046509090909090914,
      "loss": 2.5138,
      "theoretical_loss": 3.462414287869498,
      "tokens_seen": 1791229952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650707070707071,
      "loss": 2.7199,
      "theoretical_loss": 3.4624034954510154,
      "tokens_seen": 1791295488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046505050505050505,
      "loss": 2.5726,
      "theoretical_loss": 3.4623927035379283,
      "tokens_seen": 1791361024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465030303030303,
      "loss": 2.5712,
      "theoretical_loss": 3.462381912130194,
      "tokens_seen": 1791426560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046501010101010097,
      "loss": 2.4468,
      "theoretical_loss": 3.462371121227772,
      "tokens_seen": 1791492096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046498989898989903,
      "loss": 2.628,
      "theoretical_loss": 3.462360330830618,
      "tokens_seen": 1791557632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464969696969697,
      "loss": 2.8937,
      "theoretical_loss": 3.462349540938691,
      "tokens_seen": 1791623168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046494949494949495,
      "loss": 2.5733,
      "theoretical_loss": 3.4623387515519486,
      "tokens_seen": 1791688704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649292929292929,
      "loss": 2.7154,
      "theoretical_loss": 3.462327962670349,
      "tokens_seen": 1791754240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649090909090909,
      "loss": 2.3878,
      "theoretical_loss": 3.462317174293849,
      "tokens_seen": 1791819776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648888888888889,
      "loss": 2.773,
      "theoretical_loss": 3.462306386422408,
      "tokens_seen": 1791885312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648686868686869,
      "loss": 2.6715,
      "theoretical_loss": 3.462295599055983,
      "tokens_seen": 1791950848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046484848484848484,
      "loss": 2.6902,
      "theoretical_loss": 3.4622848121945324,
      "tokens_seen": 1792016384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648282828282828,
      "loss": 2.551,
      "theoretical_loss": 3.4622740258380134,
      "tokens_seen": 1792081920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648080808080808,
      "loss": 2.4135,
      "theoretical_loss": 3.4622632399863846,
      "tokens_seen": 1792147456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647878787878788,
      "loss": 2.5175,
      "theoretical_loss": 3.4622524546396027,
      "tokens_seen": 1792212992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647676767676768,
      "loss": 2.7417,
      "theoretical_loss": 3.4622416697976277,
      "tokens_seen": 1792278528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046474747474747473,
      "loss": 2.4755,
      "theoretical_loss": 3.4622308854604156,
      "tokens_seen": 1792344064
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046472727272727274,
      "loss": 2.5535,
      "theoretical_loss": 3.462220101627925,
      "tokens_seen": 1792409600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647070707070707,
      "loss": 2.6122,
      "theoretical_loss": 3.462209318300114,
      "tokens_seen": 1792475136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646868686868687,
      "loss": 2.7892,
      "theoretical_loss": 3.46219853547694,
      "tokens_seen": 1792540672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046466666666666667,
      "loss": 2.6861,
      "theoretical_loss": 3.4621877531583616,
      "tokens_seen": 1792606208
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 988330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5130808353424072,
      "objective/train/theoretical_loss": 3.462176971344337,
      "objective/train/tokens_used": 151530976,
      "theoretical_loss": 3.462176971344337,
      "tokens_seen": 1792671744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646464646464646,
      "loss": 2.5728,
      "theoretical_loss": 3.462176971344337,
      "tokens_seen": 1792671744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046462626262626264,
      "loss": 2.4717,
      "theoretical_loss": 3.4621661900348224,
      "tokens_seen": 1792737280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046460606060606065,
      "loss": 2.6032,
      "theoretical_loss": 3.4621554092297777,
      "tokens_seen": 1792802816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645858585858586,
      "loss": 2.6182,
      "theoretical_loss": 3.46214462892916,
      "tokens_seen": 1792868352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046456565656565656,
      "loss": 2.7285,
      "theoretical_loss": 3.462133849132927,
      "tokens_seen": 1792933888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046454545454545457,
      "loss": 2.4959,
      "theoretical_loss": 3.462123069841038,
      "tokens_seen": 1792999424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046452525252525253,
      "loss": 2.5876,
      "theoretical_loss": 3.462112291053449,
      "tokens_seen": 1793064960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046450505050505054,
      "loss": 2.52,
      "theoretical_loss": 3.4621015127701193,
      "tokens_seen": 1793130496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644848484848485,
      "loss": 2.5387,
      "theoretical_loss": 3.4620907349910066,
      "tokens_seen": 1793196032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046446464646464645,
      "loss": 2.4564,
      "theoretical_loss": 3.462079957716069,
      "tokens_seen": 1793261568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046444444444444446,
      "loss": 2.7585,
      "theoretical_loss": 3.4620691809452637,
      "tokens_seen": 1793327104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644242424242424,
      "loss": 2.5675,
      "theoretical_loss": 3.46205840467855,
      "tokens_seen": 1793392640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046440404040404043,
      "loss": 2.6589,
      "theoretical_loss": 3.462047628915885,
      "tokens_seen": 1793458176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643838383838384,
      "loss": 2.5341,
      "theoretical_loss": 3.462036853657227,
      "tokens_seen": 1793523712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643636363636364,
      "loss": 2.81,
      "theoretical_loss": 3.4620260789025337,
      "tokens_seen": 1793589248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046434343434343436,
      "loss": 2.6829,
      "theoretical_loss": 3.4620153046517634,
      "tokens_seen": 1793654784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643232323232323,
      "loss": 2.7504,
      "theoretical_loss": 3.462004530904874,
      "tokens_seen": 1793720320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643030303030303,
      "loss": 2.6757,
      "theoretical_loss": 3.461993757661824,
      "tokens_seen": 1793785856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642828282828283,
      "loss": 2.5854,
      "theoretical_loss": 3.461982984922571,
      "tokens_seen": 1793851392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642626262626263,
      "loss": 2.7475,
      "theoretical_loss": 3.4619722126870727,
      "tokens_seen": 1793916928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046424242424242425,
      "loss": 2.6829,
      "theoretical_loss": 3.461961440955288,
      "tokens_seen": 1793982464
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642222222222222,
      "loss": 2.5497,
      "theoretical_loss": 3.461950669727174,
      "tokens_seen": 1794048000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642020202020202,
      "loss": 2.5674,
      "theoretical_loss": 3.4619398990026893,
      "tokens_seen": 1794113536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046418181818181823,
      "loss": 2.6553,
      "theoretical_loss": 3.461929128781792,
      "tokens_seen": 1794179072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641616161616162,
      "loss": 2.7199,
      "theoretical_loss": 3.4619183590644402,
      "tokens_seen": 1794244608
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 989671,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.687277317047119,
      "objective/train/theoretical_loss": 3.4619075898505915,
      "objective/train/tokens_used": 153169376,
      "theoretical_loss": 3.4619075898505915,
      "tokens_seen": 1794310144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046414141414141414,
      "loss": 2.7491,
      "theoretical_loss": 3.4619075898505915,
      "tokens_seen": 1794310144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641212121212121,
      "loss": 2.6619,
      "theoretical_loss": 3.461896821140204,
      "tokens_seen": 1794375680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641010101010101,
      "loss": 2.6628,
      "theoretical_loss": 3.4618860529332363,
      "tokens_seen": 1794441216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640808080808081,
      "loss": 2.4893,
      "theoretical_loss": 3.461875285229646,
      "tokens_seen": 1794506752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640606060606061,
      "loss": 2.6323,
      "theoretical_loss": 3.461864518029392,
      "tokens_seen": 1794572288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046404040404040403,
      "loss": 2.7328,
      "theoretical_loss": 3.4618537513324315,
      "tokens_seen": 1794637824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464020202020202,
      "loss": 2.6513,
      "theoretical_loss": 3.4618429851387233,
      "tokens_seen": 1794703360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046400000000000006,
      "loss": 2.5925,
      "theoretical_loss": 3.4618322194482243,
      "tokens_seen": 1794768896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463979797979798,
      "loss": 2.4511,
      "theoretical_loss": 3.4618214542608943,
      "tokens_seen": 1794834432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046395959595959597,
      "loss": 2.7825,
      "theoretical_loss": 3.46181068957669,
      "tokens_seen": 1794899968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046393939393939393,
      "loss": 2.6791,
      "theoretical_loss": 3.46179992539557,
      "tokens_seen": 1794965504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639191919191919,
      "loss": 2.6882,
      "theoretical_loss": 3.4617891617174923,
      "tokens_seen": 1795031040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046389898989898995,
      "loss": 2.7291,
      "theoretical_loss": 3.4617783985424158,
      "tokens_seen": 1795096576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638787878787879,
      "loss": 2.7468,
      "theoretical_loss": 3.4617676358702973,
      "tokens_seen": 1795162112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046385858585858586,
      "loss": 2.6849,
      "theoretical_loss": 3.461756873701096,
      "tokens_seen": 1795227648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638383838383838,
      "loss": 2.5844,
      "theoretical_loss": 3.4617461120347697,
      "tokens_seen": 1795293184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046381818181818183,
      "loss": 2.5013,
      "theoretical_loss": 3.461735350871277,
      "tokens_seen": 1795358720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046379797979797984,
      "loss": 2.564,
      "theoretical_loss": 3.461724590210575,
      "tokens_seen": 1795424256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637777777777778,
      "loss": 2.6563,
      "theoretical_loss": 3.4617138300526222,
      "tokens_seen": 1795489792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046375757575757576,
      "loss": 2.4432,
      "theoretical_loss": 3.4617030703973777,
      "tokens_seen": 1795555328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637373737373737,
      "loss": 2.7619,
      "theoretical_loss": 3.461692311244798,
      "tokens_seen": 1795620864
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637171717171717,
      "loss": 2.8329,
      "theoretical_loss": 3.4616815525948432,
      "tokens_seen": 1795686400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046369696969696974,
      "loss": 2.434,
      "theoretical_loss": 3.4616707944474703,
      "tokens_seen": 1795751936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636767676767677,
      "loss": 2.5695,
      "theoretical_loss": 3.4616600368026376,
      "tokens_seen": 1795817472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046365656565656565,
      "loss": 2.3648,
      "theoretical_loss": 3.4616492796603033,
      "tokens_seen": 1795883008
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 990185,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.496757984161377,
      "objective/train/theoretical_loss": 3.461638523020426,
      "objective/train/tokens_used": 154807776,
      "theoretical_loss": 3.461638523020426,
      "tokens_seen": 1795948544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046363636363636366,
      "loss": 2.5464,
      "theoretical_loss": 3.461638523020426,
      "tokens_seen": 1795948544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636161616161616,
      "loss": 2.6964,
      "theoretical_loss": 3.4616277668829634,
      "tokens_seen": 1796014080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046359595959595963,
      "loss": 2.5699,
      "theoretical_loss": 3.461617011247874,
      "tokens_seen": 1796079616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635757575757576,
      "loss": 2.6995,
      "theoretical_loss": 3.461606256115116,
      "tokens_seen": 1796145152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046355555555555554,
      "loss": 2.5355,
      "theoretical_loss": 3.461595501484647,
      "tokens_seen": 1796210688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046353535353535355,
      "loss": 2.4066,
      "theoretical_loss": 3.4615847473564263,
      "tokens_seen": 1796276224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635151515151515,
      "loss": 2.573,
      "theoretical_loss": 3.4615739937304113,
      "tokens_seen": 1796341760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634949494949495,
      "loss": 2.5721,
      "theoretical_loss": 3.4615632406065604,
      "tokens_seen": 1796407296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634747474747475,
      "loss": 2.782,
      "theoretical_loss": 3.461552487984832,
      "tokens_seen": 1796472832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634545454545455,
      "loss": 2.3226,
      "theoretical_loss": 3.4615417358651843,
      "tokens_seen": 1796538368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046343434343434345,
      "loss": 2.6606,
      "theoretical_loss": 3.461530984247575,
      "tokens_seen": 1796603904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634141414141414,
      "loss": 2.6738,
      "theoretical_loss": 3.4615202331319637,
      "tokens_seen": 1796669440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633939393939394,
      "loss": 2.6571,
      "theoretical_loss": 3.4615094825183066,
      "tokens_seen": 1796734976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046337373737373737,
      "loss": 2.5698,
      "theoretical_loss": 3.4614987324065645,
      "tokens_seen": 1796800512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633535353535354,
      "loss": 2.6218,
      "theoretical_loss": 3.4614879827966933,
      "tokens_seen": 1796866048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046333333333333334,
      "loss": 2.6216,
      "theoretical_loss": 3.461477233688653,
      "tokens_seen": 1796931584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633131313131313,
      "loss": 2.6791,
      "theoretical_loss": 3.4614664850824,
      "tokens_seen": 1796997120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632929292929293,
      "loss": 2.6627,
      "theoretical_loss": 3.4614557369778947,
      "tokens_seen": 1797062656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632727272727273,
      "loss": 2.7609,
      "theoretical_loss": 3.461444989375094,
      "tokens_seen": 1797128192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632525252525253,
      "loss": 2.689,
      "theoretical_loss": 3.461434242273957,
      "tokens_seen": 1797193728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046323232323232323,
      "loss": 2.352,
      "theoretical_loss": 3.461423495674441,
      "tokens_seen": 1797259264
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632121212121212,
      "loss": 2.7198,
      "theoretical_loss": 3.461412749576505,
      "tokens_seen": 1797324800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631919191919192,
      "loss": 2.5821,
      "theoretical_loss": 3.4614020039801074,
      "tokens_seen": 1797390336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631717171717172,
      "loss": 2.5262,
      "theoretical_loss": 3.4613912588852065,
      "tokens_seen": 1797455872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046315151515151517,
      "loss": 2.5769,
      "theoretical_loss": 3.46138051429176,
      "tokens_seen": 1797521408
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 991489,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6060218811035156,
      "objective/train/theoretical_loss": 3.4613697701997266,
      "objective/train/tokens_used": 156446176,
      "theoretical_loss": 3.4613697701997266,
      "tokens_seen": 1797586944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631313131313131,
      "loss": 2.5123,
      "theoretical_loss": 3.4613697701997266,
      "tokens_seen": 1797586944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631111111111111,
      "loss": 2.7132,
      "theoretical_loss": 3.461359026609065,
      "tokens_seen": 1797652480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630909090909091,
      "loss": 2.5873,
      "theoretical_loss": 3.4613482835197322,
      "tokens_seen": 1797718016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630707070707071,
      "loss": 2.737,
      "theoretical_loss": 3.461337540931688,
      "tokens_seen": 1797783552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046305050505050506,
      "loss": 2.5273,
      "theoretical_loss": 3.4613267988448913,
      "tokens_seen": 1797849088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463030303030303,
      "loss": 2.561,
      "theoretical_loss": 3.4613160572592983,
      "tokens_seen": 1797914624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046301010101010097,
      "loss": 2.7165,
      "theoretical_loss": 3.4613053161748684,
      "tokens_seen": 1797980160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046298989898989904,
      "loss": 2.7665,
      "theoretical_loss": 3.46129457559156,
      "tokens_seen": 1798045696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462969696969697,
      "loss": 2.846,
      "theoretical_loss": 3.461283835509332,
      "tokens_seen": 1798111232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046294949494949495,
      "loss": 2.7031,
      "theoretical_loss": 3.4612730959281413,
      "tokens_seen": 1798176768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629292929292929,
      "loss": 2.7652,
      "theoretical_loss": 3.4612623568479477,
      "tokens_seen": 1798242304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629090909090909,
      "loss": 2.5421,
      "theoretical_loss": 3.461251618268709,
      "tokens_seen": 1798307840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046288888888888893,
      "loss": 2.7416,
      "theoretical_loss": 3.4612408801903833,
      "tokens_seen": 1798373376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628686868686869,
      "loss": 2.6124,
      "theoretical_loss": 3.4612301426129295,
      "tokens_seen": 1798438912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046284848484848484,
      "loss": 2.8337,
      "theoretical_loss": 3.4612194055363057,
      "tokens_seen": 1798504448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628282828282828,
      "loss": 2.5391,
      "theoretical_loss": 3.4612086689604706,
      "tokens_seen": 1798569984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046280808080808087,
      "loss": 2.6909,
      "theoretical_loss": 3.461197932885382,
      "tokens_seen": 1798635520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627878787878788,
      "loss": 2.4735,
      "theoretical_loss": 3.461187197310999,
      "tokens_seen": 1798701056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627676767676768,
      "loss": 2.4016,
      "theoretical_loss": 3.4611764622372796,
      "tokens_seen": 1798766592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046274747474747474,
      "loss": 2.7274,
      "theoretical_loss": 3.4611657276641816,
      "tokens_seen": 1798832128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627272727272727,
      "loss": 2.4681,
      "theoretical_loss": 3.4611549935916646,
      "tokens_seen": 1798897664
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046270707070707076,
      "loss": 2.6583,
      "theoretical_loss": 3.4611442600196867,
      "tokens_seen": 1798963200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626868686868687,
      "loss": 2.5414,
      "theoretical_loss": 3.461133526948206,
      "tokens_seen": 1799028736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626666666666667,
      "loss": 2.7138,
      "theoretical_loss": 3.4611227943771814,
      "tokens_seen": 1799094272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046264646464646463,
      "loss": 2.6,
      "theoretical_loss": 3.4611120623065705,
      "tokens_seen": 1799159808
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 992118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7015321254730225,
      "objective/train/theoretical_loss": 3.4611013307363323,
      "objective/train/tokens_used": 158084576,
      "theoretical_loss": 3.4611013307363323,
      "tokens_seen": 1799225344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046262626262626264,
      "loss": 2.5245,
      "theoretical_loss": 3.4611013307363323,
      "tokens_seen": 1799225344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046260606060606065,
      "loss": 2.5428,
      "theoretical_loss": 3.4610905996664254,
      "tokens_seen": 1799290880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625858585858586,
      "loss": 2.9118,
      "theoretical_loss": 3.4610798690968076,
      "tokens_seen": 1799356416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046256565656565657,
      "loss": 2.7124,
      "theoretical_loss": 3.461069139027438,
      "tokens_seen": 1799421952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625454545454545,
      "loss": 2.6946,
      "theoretical_loss": 3.461058409458275,
      "tokens_seen": 1799487488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046252525252525253,
      "loss": 2.7679,
      "theoretical_loss": 3.461047680389277,
      "tokens_seen": 1799553024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046250505050505054,
      "loss": 2.8121,
      "theoretical_loss": 3.461036951820402,
      "tokens_seen": 1799618560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624848484848485,
      "loss": 2.5779,
      "theoretical_loss": 3.4610262237516096,
      "tokens_seen": 1799684096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046246464646464646,
      "loss": 2.6062,
      "theoretical_loss": 3.4610154961828568,
      "tokens_seen": 1799749632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046244444444444447,
      "loss": 2.6283,
      "theoretical_loss": 3.461004769114103,
      "tokens_seen": 1799815168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624242424242424,
      "loss": 2.6631,
      "theoretical_loss": 3.4609940425453067,
      "tokens_seen": 1799880704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046240404040404044,
      "loss": 2.6806,
      "theoretical_loss": 3.460983316476426,
      "tokens_seen": 1799946240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623838383838384,
      "loss": 2.7087,
      "theoretical_loss": 3.46097259090742,
      "tokens_seen": 1800011776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046236363636363635,
      "loss": 2.4683,
      "theoretical_loss": 3.4609618658382466,
      "tokens_seen": 1800077312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046234343434343436,
      "loss": 2.8184,
      "theoretical_loss": 3.4609511412688647,
      "tokens_seen": 1800142848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623232323232323,
      "loss": 2.5615,
      "theoretical_loss": 3.4609404171992324,
      "tokens_seen": 1800208384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046230303030303033,
      "loss": 2.5703,
      "theoretical_loss": 3.460929693629309,
      "tokens_seen": 1800273920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622828282828283,
      "loss": 2.5853,
      "theoretical_loss": 3.4609189705590513,
      "tokens_seen": 1800339456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622626262626263,
      "loss": 2.8325,
      "theoretical_loss": 3.46090824798842,
      "tokens_seen": 1800404992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046224242424242425,
      "loss": 2.4077,
      "theoretical_loss": 3.460897525917373,
      "tokens_seen": 1800470528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622222222222222,
      "loss": 2.45,
      "theoretical_loss": 3.4608868043458676,
      "tokens_seen": 1800536064
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622020202020202,
      "loss": 2.7897,
      "theoretical_loss": 3.4608760832738636,
      "tokens_seen": 1800601600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621818181818182,
      "loss": 2.7197,
      "theoretical_loss": 3.4608653627013193,
      "tokens_seen": 1800667136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621616161616162,
      "loss": 2.5448,
      "theoretical_loss": 3.4608546426281928,
      "tokens_seen": 1800732672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046214141414141415,
      "loss": 2.6301,
      "theoretical_loss": 3.460843923054443,
      "tokens_seen": 1800798208
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 993246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.732495069503784,
      "objective/train/theoretical_loss": 3.4608332039800294,
      "objective/train/tokens_used": 159722976,
      "theoretical_loss": 3.4608332039800294,
      "tokens_seen": 1800863744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004621212121212121,
      "loss": 2.7341,
      "theoretical_loss": 3.4608332039800294,
      "tokens_seen": 1800863744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004621010101010101,
      "loss": 2.6709,
      "theoretical_loss": 3.460822485404909,
      "tokens_seen": 1800929280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620808080808081,
      "loss": 2.6878,
      "theoretical_loss": 3.460811767329041,
      "tokens_seen": 1800994816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620606060606061,
      "loss": 2.8197,
      "theoretical_loss": 3.460801049752384,
      "tokens_seen": 1801060352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046204040404040404,
      "loss": 2.4737,
      "theoretical_loss": 3.4607903326748968,
      "tokens_seen": 1801125888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000462020202020202,
      "loss": 2.6015,
      "theoretical_loss": 3.4607796160965374,
      "tokens_seen": 1801191424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000462,
      "loss": 2.7105,
      "theoretical_loss": 3.460768900017265,
      "tokens_seen": 1801256960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461979797979798,
      "loss": 2.7153,
      "theoretical_loss": 3.4607581844370383,
      "tokens_seen": 1801322496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461959595959596,
      "loss": 2.7215,
      "theoretical_loss": 3.460747469355815,
      "tokens_seen": 1801388032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046193939393939393,
      "loss": 2.7523,
      "theoretical_loss": 3.4607367547735546,
      "tokens_seen": 1801453568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619191919191919,
      "loss": 2.6356,
      "theoretical_loss": 3.460726040690215,
      "tokens_seen": 1801519104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046189898989898996,
      "loss": 2.8216,
      "theoretical_loss": 3.4607153271057554,
      "tokens_seen": 1801584640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618787878787879,
      "loss": 2.6657,
      "theoretical_loss": 3.460704614020135,
      "tokens_seen": 1801650176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046185858585858587,
      "loss": 2.5136,
      "theoretical_loss": 3.460693901433311,
      "tokens_seen": 1801715712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618383838383838,
      "loss": 2.7718,
      "theoretical_loss": 3.4606831893452425,
      "tokens_seen": 1801781248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618181818181818,
      "loss": 2.645,
      "theoretical_loss": 3.460672477755889,
      "tokens_seen": 1801846784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046179797979797985,
      "loss": 2.637,
      "theoretical_loss": 3.4606617666652078,
      "tokens_seen": 1801912320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617777777777778,
      "loss": 2.6257,
      "theoretical_loss": 3.4606510560731585,
      "tokens_seen": 1801977856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046175757575757576,
      "loss": 2.5434,
      "theoretical_loss": 3.4606403459796997,
      "tokens_seen": 1802043392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617373737373737,
      "loss": 2.5597,
      "theoretical_loss": 3.4606296363847897,
      "tokens_seen": 1802108928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617171717171717,
      "loss": 2.4782,
      "theoretical_loss": 3.4606189272883876,
      "tokens_seen": 1802174464
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046169696969696974,
      "loss": 2.5543,
      "theoretical_loss": 3.460608218690451,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616767676767677,
      "loss": 2.6303,
      "theoretical_loss": 3.46059751059094,
      "tokens_seen": 1802305536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046165656565656565,
      "loss": 2.3844,
      "theoretical_loss": 3.4605868029898126,
      "tokens_seen": 1802371072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616363636363636,
      "loss": 2.6324,
      "theoretical_loss": 3.4605760958870277,
      "tokens_seen": 1802436608
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 994266,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.712080717086792,
      "objective/train/theoretical_loss": 3.460565389282543,
      "objective/train/tokens_used": 161361376,
      "theoretical_loss": 3.460565389282543,
      "tokens_seen": 1802502144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616161616161616,
      "loss": 2.6,
      "theoretical_loss": 3.460565389282543,
      "tokens_seen": 1802502144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046159595959595963,
      "loss": 2.8756,
      "theoretical_loss": 3.4605546831763188,
      "tokens_seen": 1802567680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615757575757576,
      "loss": 2.7354,
      "theoretical_loss": 3.460543977568313,
      "tokens_seen": 1802633216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046155555555555555,
      "loss": 2.6013,
      "theoretical_loss": 3.460533272458484,
      "tokens_seen": 1802698752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615353535353535,
      "loss": 2.8307,
      "theoretical_loss": 3.4605225678467906,
      "tokens_seen": 1802764288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615151515151515,
      "loss": 2.7824,
      "theoretical_loss": 3.460511863733192,
      "tokens_seen": 1802829824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614949494949495,
      "loss": 2.4709,
      "theoretical_loss": 3.460501160117647,
      "tokens_seen": 1802895360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614747474747475,
      "loss": 2.6541,
      "theoretical_loss": 3.4604904570001134,
      "tokens_seen": 1802960896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046145454545454544,
      "loss": 2.684,
      "theoretical_loss": 3.4604797543805503,
      "tokens_seen": 1803026432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046143434343434345,
      "loss": 2.6394,
      "theoretical_loss": 3.460469052258917,
      "tokens_seen": 1803091968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614141414141414,
      "loss": 2.5222,
      "theoretical_loss": 3.460458350635172,
      "tokens_seen": 1803157504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613939393939394,
      "loss": 2.4386,
      "theoretical_loss": 3.460447649509274,
      "tokens_seen": 1803223040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613737373737374,
      "loss": 2.5735,
      "theoretical_loss": 3.4604369488811813,
      "tokens_seen": 1803288576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046135353535353533,
      "loss": 2.5963,
      "theoretical_loss": 3.4604262487508533,
      "tokens_seen": 1803354112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046133333333333334,
      "loss": 2.6435,
      "theoretical_loss": 3.460415549118248,
      "tokens_seen": 1803419648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613131313131313,
      "loss": 2.5554,
      "theoretical_loss": 3.460404849983325,
      "tokens_seen": 1803485184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612929292929293,
      "loss": 2.5415,
      "theoretical_loss": 3.4603941513460423,
      "tokens_seen": 1803550720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046127272727272727,
      "loss": 2.6456,
      "theoretical_loss": 3.4603834532063598,
      "tokens_seen": 1803616256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612525252525253,
      "loss": 2.6177,
      "theoretical_loss": 3.460372755564235,
      "tokens_seen": 1803681792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046123232323232324,
      "loss": 2.3797,
      "theoretical_loss": 3.4603620584196273,
      "tokens_seen": 1803747328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612121212121212,
      "loss": 2.5797,
      "theoretical_loss": 3.4603513617724957,
      "tokens_seen": 1803812864
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611919191919192,
      "loss": 2.6529,
      "theoretical_loss": 3.460340665622798,
      "tokens_seen": 1803878400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046117171717171716,
      "loss": 2.5033,
      "theoretical_loss": 3.460329969970495,
      "tokens_seen": 1803943936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046115151515151517,
      "loss": 2.7475,
      "theoretical_loss": 3.460319274815543,
      "tokens_seen": 1804009472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046113131313131313,
      "loss": 2.4978,
      "theoretical_loss": 3.460308580157902,
      "tokens_seen": 1804075008
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 994874,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3973050117492676,
      "objective/train/theoretical_loss": 3.460297885997531,
      "objective/train/tokens_used": 162999776,
      "theoretical_loss": 3.460297885997531,
      "tokens_seen": 1804140544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046111111111111114,
      "loss": 2.5142,
      "theoretical_loss": 3.460297885997531,
      "tokens_seen": 1804140544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610909090909091,
      "loss": 2.5792,
      "theoretical_loss": 3.4602871923343894,
      "tokens_seen": 1804206080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610707070707071,
      "loss": 2.6572,
      "theoretical_loss": 3.4602764991684345,
      "tokens_seen": 1804271616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046105050505050506,
      "loss": 2.7006,
      "theoretical_loss": 3.460265806499626,
      "tokens_seen": 1804337152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461030303030303,
      "loss": 2.5644,
      "theoretical_loss": 3.4602551143279228,
      "tokens_seen": 1804402688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046101010101010103,
      "loss": 2.5553,
      "theoretical_loss": 3.4602444226532834,
      "tokens_seen": 1804468224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460989898989899,
      "loss": 2.4872,
      "theoretical_loss": 3.4602337314756664,
      "tokens_seen": 1804533760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460969696969697,
      "loss": 2.5919,
      "theoretical_loss": 3.460223040795032,
      "tokens_seen": 1804599296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046094949494949496,
      "loss": 2.5868,
      "theoretical_loss": 3.4602123506113376,
      "tokens_seen": 1804664832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609292929292929,
      "loss": 2.557,
      "theoretical_loss": 3.4602016609245423,
      "tokens_seen": 1804730368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609090909090909,
      "loss": 2.4903,
      "theoretical_loss": 3.460190971734605,
      "tokens_seen": 1804795904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046088888888888894,
      "loss": 2.6423,
      "theoretical_loss": 3.4601802830414856,
      "tokens_seen": 1804861440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608686868686869,
      "loss": 2.607,
      "theoretical_loss": 3.4601695948451416,
      "tokens_seen": 1804926976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046084848484848485,
      "loss": 2.4623,
      "theoretical_loss": 3.4601589071455323,
      "tokens_seen": 1804992512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608282828282828,
      "loss": 2.5671,
      "theoretical_loss": 3.460148219942617,
      "tokens_seen": 1805058048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608080808080808,
      "loss": 2.7849,
      "theoretical_loss": 3.4601375332363538,
      "tokens_seen": 1805123584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046078787878787883,
      "loss": 2.4045,
      "theoretical_loss": 3.4601268470267024,
      "tokens_seen": 1805189120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607676767676768,
      "loss": 2.6255,
      "theoretical_loss": 3.4601161613136213,
      "tokens_seen": 1805254656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046074747474747474,
      "loss": 2.3319,
      "theoretical_loss": 3.4601054760970698,
      "tokens_seen": 1805320192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607272727272727,
      "loss": 2.6209,
      "theoretical_loss": 3.460094791377006,
      "tokens_seen": 1805385728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046070707070707077,
      "loss": 2.5799,
      "theoretical_loss": 3.4600841071533894,
      "tokens_seen": 1805451264
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606868686868687,
      "loss": 2.6091,
      "theoretical_loss": 3.460073423426179,
      "tokens_seen": 1805516800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606666666666667,
      "loss": 2.8417,
      "theoretical_loss": 3.460062740195333,
      "tokens_seen": 1805582336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046064646464646464,
      "loss": 2.5454,
      "theoretical_loss": 3.4600520574608113,
      "tokens_seen": 1805647872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606262626262626,
      "loss": 2.7244,
      "theoretical_loss": 3.4600413752225725,
      "tokens_seen": 1805713408
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 995886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.640719413757324,
      "objective/train/theoretical_loss": 3.460030693480575,
      "objective/train/tokens_used": 164638176,
      "theoretical_loss": 3.460030693480575,
      "tokens_seen": 1805778944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046060606060606066,
      "loss": 2.7838,
      "theoretical_loss": 3.460030693480575,
      "tokens_seen": 1805778944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605858585858586,
      "loss": 2.5454,
      "theoretical_loss": 3.460020012234778,
      "tokens_seen": 1805844480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046056565656565657,
      "loss": 2.737,
      "theoretical_loss": 3.460009331485141,
      "tokens_seen": 1805910016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046054545454545453,
      "loss": 2.6839,
      "theoretical_loss": 3.459998651231622,
      "tokens_seen": 1805975552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046052525252525254,
      "loss": 2.5493,
      "theoretical_loss": 3.459987971474181,
      "tokens_seen": 1806041088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046050505050505055,
      "loss": 2.4706,
      "theoretical_loss": 3.4599772922127765,
      "tokens_seen": 1806106624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604848484848485,
      "loss": 2.7013,
      "theoretical_loss": 3.4599666134473672,
      "tokens_seen": 1806172160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046046464646464646,
      "loss": 2.7035,
      "theoretical_loss": 3.459955935177912,
      "tokens_seen": 1806237696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604444444444444,
      "loss": 2.6302,
      "theoretical_loss": 3.45994525740437,
      "tokens_seen": 1806303232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046042424242424243,
      "loss": 2.5569,
      "theoretical_loss": 3.459934580126701,
      "tokens_seen": 1806368768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046040404040404044,
      "loss": 2.6291,
      "theoretical_loss": 3.459923903344863,
      "tokens_seen": 1806434304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603838383838384,
      "loss": 2.6722,
      "theoretical_loss": 3.4599132270588155,
      "tokens_seen": 1806499840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046036363636363636,
      "loss": 2.8284,
      "theoretical_loss": 3.4599025512685166,
      "tokens_seen": 1806565376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046034343434343437,
      "loss": 2.836,
      "theoretical_loss": 3.459891875973927,
      "tokens_seen": 1806630912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603232323232323,
      "loss": 2.5876,
      "theoretical_loss": 3.459881201175004,
      "tokens_seen": 1806696448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046030303030303034,
      "loss": 2.5903,
      "theoretical_loss": 3.4598705268717076,
      "tokens_seen": 1806761984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602828282828283,
      "loss": 2.5872,
      "theoretical_loss": 3.459859853063996,
      "tokens_seen": 1806827520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046026262626262625,
      "loss": 2.7085,
      "theoretical_loss": 3.4598491797518287,
      "tokens_seen": 1806893056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046024242424242426,
      "loss": 2.6058,
      "theoretical_loss": 3.4598385069351654,
      "tokens_seen": 1806958592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602222222222222,
      "loss": 2.431,
      "theoretical_loss": 3.459827834613964,
      "tokens_seen": 1807024128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046020202020202023,
      "loss": 2.6627,
      "theoretical_loss": 3.459817162788184,
      "tokens_seen": 1807089664
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601818181818182,
      "loss": 2.6105,
      "theoretical_loss": 3.4598064914577846,
      "tokens_seen": 1807155200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601616161616162,
      "loss": 2.6375,
      "theoretical_loss": 3.4597958206227246,
      "tokens_seen": 1807220736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046014141414141415,
      "loss": 2.6245,
      "theoretical_loss": 3.459785150282963,
      "tokens_seen": 1807286272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601212121212121,
      "loss": 2.7088,
      "theoretical_loss": 3.459774480438459,
      "tokens_seen": 1807351808
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 996582,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.828779935836792,
      "objective/train/theoretical_loss": 3.4597638110891715,
      "objective/train/tokens_used": 166276576,
      "theoretical_loss": 3.4597638110891715,
      "tokens_seen": 1807417344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601010101010101,
      "loss": 2.6118,
      "theoretical_loss": 3.4597638110891715,
      "tokens_seen": 1807417344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600808080808081,
      "loss": 2.55,
      "theoretical_loss": 3.4597531422350603,
      "tokens_seen": 1807482880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600606060606061,
      "loss": 2.9119,
      "theoretical_loss": 3.459742473876083,
      "tokens_seen": 1807548416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046004040404040405,
      "loss": 2.4969,
      "theoretical_loss": 3.4597318060122,
      "tokens_seen": 1807613952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460020202020202,
      "loss": 2.3569,
      "theoretical_loss": 3.4597211386433697,
      "tokens_seen": 1807679488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046,
      "loss": 2.8525,
      "theoretical_loss": 3.4597104717695517,
      "tokens_seen": 1807745024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459979797979798,
      "loss": 2.6927,
      "theoretical_loss": 3.459699805390705,
      "tokens_seen": 1807810560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459959595959596,
      "loss": 2.6665,
      "theoretical_loss": 3.4596891395067875,
      "tokens_seen": 1807876096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045993939393939394,
      "loss": 2.6101,
      "theoretical_loss": 3.45967847411776,
      "tokens_seen": 1807941632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599191919191919,
      "loss": 2.6743,
      "theoretical_loss": 3.4596678092235806,
      "tokens_seen": 1808007168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598989898989899,
      "loss": 2.7123,
      "theoretical_loss": 3.4596571448242086,
      "tokens_seen": 1808072704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598787878787879,
      "loss": 2.589,
      "theoretical_loss": 3.459646480919603,
      "tokens_seen": 1808138240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598585858585859,
      "loss": 2.4773,
      "theoretical_loss": 3.4596358175097235,
      "tokens_seen": 1808203776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045983838383838383,
      "loss": 2.4748,
      "theoretical_loss": 3.4596251545945282,
      "tokens_seen": 1808269312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598181818181818,
      "loss": 2.5673,
      "theoretical_loss": 3.4596144921739773,
      "tokens_seen": 1808334848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597979797979798,
      "loss": 2.4766,
      "theoretical_loss": 3.45960383024803,
      "tokens_seen": 1808400384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597777777777778,
      "loss": 2.7642,
      "theoretical_loss": 3.459593168816644,
      "tokens_seen": 1808465920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045975757575757577,
      "loss": 2.7496,
      "theoretical_loss": 3.4595825078797793,
      "tokens_seen": 1808531456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597373737373737,
      "loss": 2.8011,
      "theoretical_loss": 3.4595718474373953,
      "tokens_seen": 1808596992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597171717171717,
      "loss": 2.8709,
      "theoretical_loss": 3.459561187489451,
      "tokens_seen": 1808662528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045969696969696975,
      "loss": 2.6683,
      "theoretical_loss": 3.459550528035905,
      "tokens_seen": 1808728064
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596767676767677,
      "loss": 2.6207,
      "theoretical_loss": 3.4595398690767176,
      "tokens_seen": 1808793600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045965656565656566,
      "loss": 2.4637,
      "theoretical_loss": 3.459529210611847,
      "tokens_seen": 1808859136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596363636363636,
      "loss": 2.7937,
      "theoretical_loss": 3.4595185526412524,
      "tokens_seen": 1808924672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596161616161616,
      "loss": 2.6215,
      "theoretical_loss": 3.4595078951648937,
      "tokens_seen": 1808990208
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 997711,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4244730472564697,
      "objective/train/theoretical_loss": 3.4594972381827294,
      "objective/train/tokens_used": 167914976,
      "theoretical_loss": 3.4594972381827294,
      "tokens_seen": 1809055744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045959595959595964,
      "loss": 2.64,
      "theoretical_loss": 3.4594972381827294,
      "tokens_seen": 1809055744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595757575757576,
      "loss": 2.6338,
      "theoretical_loss": 3.4594865816947182,
      "tokens_seen": 1809121280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045955555555555555,
      "loss": 2.5629,
      "theoretical_loss": 3.4594759257008207,
      "tokens_seen": 1809186816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595353535353535,
      "loss": 2.5733,
      "theoretical_loss": 3.4594652702009956,
      "tokens_seen": 1809252352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595151515151515,
      "loss": 2.6077,
      "theoretical_loss": 3.4594546151952015,
      "tokens_seen": 1809317888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045949494949494953,
      "loss": 2.4091,
      "theoretical_loss": 3.4594439606833975,
      "tokens_seen": 1809383424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594747474747475,
      "loss": 2.6934,
      "theoretical_loss": 3.459433306665544,
      "tokens_seen": 1809448960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045945454545454545,
      "loss": 2.6156,
      "theoretical_loss": 3.4594226531415986,
      "tokens_seen": 1809514496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594343434343434,
      "loss": 2.8258,
      "theoretical_loss": 3.4594120001115223,
      "tokens_seen": 1809580032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045941414141414147,
      "loss": 2.7045,
      "theoretical_loss": 3.459401347575273,
      "tokens_seen": 1809645568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593939393939394,
      "loss": 2.6075,
      "theoretical_loss": 3.4593906955328104,
      "tokens_seen": 1809711104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593737373737374,
      "loss": 2.6631,
      "theoretical_loss": 3.4593800439840936,
      "tokens_seen": 1809776640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045935353535353534,
      "loss": 2.8016,
      "theoretical_loss": 3.459369392929082,
      "tokens_seen": 1809842176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045933333333333335,
      "loss": 2.7707,
      "theoretical_loss": 3.4593587423677348,
      "tokens_seen": 1809907712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045931313131313136,
      "loss": 2.5944,
      "theoretical_loss": 3.459348092300011,
      "tokens_seen": 1809973248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592929292929293,
      "loss": 2.536,
      "theoretical_loss": 3.45933744272587,
      "tokens_seen": 1810038784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592727272727273,
      "loss": 2.6021,
      "theoretical_loss": 3.459326793645271,
      "tokens_seen": 1810104320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045925252525252523,
      "loss": 2.6998,
      "theoretical_loss": 3.4593161450581738,
      "tokens_seen": 1810169856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045923232323232324,
      "loss": 2.8461,
      "theoretical_loss": 3.4593054969645367,
      "tokens_seen": 1810235392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045921212121212125,
      "loss": 2.7473,
      "theoretical_loss": 3.45929484936432,
      "tokens_seen": 1810300928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591919191919192,
      "loss": 2.6926,
      "theoretical_loss": 3.459284202257482,
      "tokens_seen": 1810366464
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045917171717171717,
      "loss": 2.6265,
      "theoretical_loss": 3.459273555643983,
      "tokens_seen": 1810432000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591515151515152,
      "loss": 2.5741,
      "theoretical_loss": 3.4592629095237815,
      "tokens_seen": 1810497536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045913131313131313,
      "loss": 2.6005,
      "theoretical_loss": 3.459252263896837,
      "tokens_seen": 1810563072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045911111111111115,
      "loss": 2.6592,
      "theoretical_loss": 3.4592416187631088,
      "tokens_seen": 1810628608
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 998243,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6315853595733643,
      "objective/train/theoretical_loss": 3.4592309741225558,
      "objective/train/tokens_used": 169553376,
      "theoretical_loss": 3.4592309741225558,
      "tokens_seen": 1810694144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004590909090909091,
      "loss": 2.5426,
      "theoretical_loss": 3.4592309741225558,
      "tokens_seen": 1810694144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045907070707070706,
      "loss": 2.5075,
      "theoretical_loss": 3.459220329975138,
      "tokens_seen": 1810759680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045905050505050507,
      "loss": 2.7955,
      "theoretical_loss": 3.4592096863208144,
      "tokens_seen": 1810825216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045903030303030303,
      "loss": 2.5196,
      "theoretical_loss": 3.4591990431595443,
      "tokens_seen": 1810890752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045901010101010104,
      "loss": 2.6317,
      "theoretical_loss": 3.4591884004912874,
      "tokens_seen": 1810956288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458989898989899,
      "loss": 2.3922,
      "theoretical_loss": 3.4591777583160024,
      "tokens_seen": 1811021824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458969696969697,
      "loss": 2.9019,
      "theoretical_loss": 3.4591671166336484,
      "tokens_seen": 1811087360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045894949494949496,
      "loss": 2.7095,
      "theoretical_loss": 3.459156475444186,
      "tokens_seen": 1811152896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589292929292929,
      "loss": 2.7239,
      "theoretical_loss": 3.4591458347475736,
      "tokens_seen": 1811218432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045890909090909093,
      "loss": 2.7473,
      "theoretical_loss": 3.4591351945437703,
      "tokens_seen": 1811283968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588888888888889,
      "loss": 2.6645,
      "theoretical_loss": 3.4591245548327363,
      "tokens_seen": 1811349504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588686868686869,
      "loss": 2.8029,
      "theoretical_loss": 3.45911391561443,
      "tokens_seen": 1811415040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045884848484848486,
      "loss": 2.5806,
      "theoretical_loss": 3.4591032768888117,
      "tokens_seen": 1811480576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588282828282828,
      "loss": 2.6284,
      "theoretical_loss": 3.4590926386558403,
      "tokens_seen": 1811546112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588080808080808,
      "loss": 2.767,
      "theoretical_loss": 3.459082000915475,
      "tokens_seen": 1811611648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045878787878787883,
      "loss": 2.792,
      "theoretical_loss": 3.4590713636676753,
      "tokens_seen": 1811677184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587676767676768,
      "loss": 2.7077,
      "theoretical_loss": 3.4590607269124005,
      "tokens_seen": 1811742720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045874747474747475,
      "loss": 2.6499,
      "theoretical_loss": 3.459050090649611,
      "tokens_seen": 1811808256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587272727272727,
      "loss": 2.8369,
      "theoretical_loss": 3.4590394548792642,
      "tokens_seen": 1811873792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587070707070707,
      "loss": 2.3474,
      "theoretical_loss": 3.459028819601321,
      "tokens_seen": 1811939328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045868686868686873,
      "loss": 2.7419,
      "theoretical_loss": 3.459018184815741,
      "tokens_seen": 1812004864
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586666666666667,
      "loss": 2.7241,
      "theoretical_loss": 3.459007550522482,
      "tokens_seen": 1812070400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045864646464646464,
      "loss": 2.5917,
      "theoretical_loss": 3.458996916721505,
      "tokens_seen": 1812135936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586262626262626,
      "loss": 2.5461,
      "theoretical_loss": 3.4589862834127683,
      "tokens_seen": 1812201472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045860606060606066,
      "loss": 2.429,
      "theoretical_loss": 3.458975650596232,
      "tokens_seen": 1812267008
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 999229,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6778602600097656,
      "objective/train/theoretical_loss": 3.458965018271855,
      "objective/train/tokens_used": 171191776,
      "theoretical_loss": 3.458965018271855,
      "tokens_seen": 1812332544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585858585858586,
      "loss": 2.6921,
      "theoretical_loss": 3.458965018271855,
      "tokens_seen": 1812332544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585656565656566,
      "loss": 2.6781,
      "theoretical_loss": 3.458954386439598,
      "tokens_seen": 1812398080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045854545454545453,
      "loss": 2.8004,
      "theoretical_loss": 3.458943755099419,
      "tokens_seen": 1812463616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585252525252525,
      "loss": 2.5061,
      "theoretical_loss": 3.4589331242512777,
      "tokens_seen": 1812529152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045850505050505056,
      "loss": 2.6991,
      "theoretical_loss": 3.458922493895134,
      "tokens_seen": 1812594688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584848484848485,
      "loss": 2.59,
      "theoretical_loss": 3.458911864030947,
      "tokens_seen": 1812660224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045846464646464647,
      "loss": 2.7665,
      "theoretical_loss": 3.4589012346586765,
      "tokens_seen": 1812725760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584444444444444,
      "loss": 2.5644,
      "theoretical_loss": 3.458890605778281,
      "tokens_seen": 1812791296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584242424242424,
      "loss": 2.6733,
      "theoretical_loss": 3.4588799773897216,
      "tokens_seen": 1812856832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045840404040404045,
      "loss": 2.495,
      "theoretical_loss": 3.4588693494929563,
      "tokens_seen": 1812922368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583838383838384,
      "loss": 2.4922,
      "theoretical_loss": 3.458858722087945,
      "tokens_seen": 1812987904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045836363636363636,
      "loss": 2.4791,
      "theoretical_loss": 3.4588480951746474,
      "tokens_seen": 1813053440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583434343434343,
      "loss": 2.6036,
      "theoretical_loss": 3.4588374687530226,
      "tokens_seen": 1813118976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045832323232323233,
      "loss": 2.4228,
      "theoretical_loss": 3.4588268428230307,
      "tokens_seen": 1813184512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045830303030303034,
      "loss": 2.747,
      "theoretical_loss": 3.4588162173846304,
      "tokens_seen": 1813250048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582828282828283,
      "loss": 2.4541,
      "theoretical_loss": 3.458805592437782,
      "tokens_seen": 1813315584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045826262626262626,
      "loss": 2.4754,
      "theoretical_loss": 3.458794967982444,
      "tokens_seen": 1813381120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582424242424242,
      "loss": 2.3365,
      "theoretical_loss": 3.458784344018577,
      "tokens_seen": 1813446656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582222222222222,
      "loss": 2.8325,
      "theoretical_loss": 3.45877372054614,
      "tokens_seen": 1813512192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045820202020202023,
      "loss": 2.7678,
      "theoretical_loss": 3.4587630975650923,
      "tokens_seen": 1813577728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581818181818182,
      "loss": 2.5188,
      "theoretical_loss": 3.4587524750753937,
      "tokens_seen": 1813643264
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045816161616161615,
      "loss": 2.6459,
      "theoretical_loss": 3.4587418530770035,
      "tokens_seen": 1813708800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045814141414141416,
      "loss": 2.477,
      "theoretical_loss": 3.4587312315698813,
      "tokens_seen": 1813774336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581212121212121,
      "loss": 2.5432,
      "theoretical_loss": 3.458720610553987,
      "tokens_seen": 1813839872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581010101010101,
      "loss": 2.7704,
      "theoretical_loss": 3.4587099900292797,
      "tokens_seen": 1813905408
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 999785,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5763378143310547,
      "objective/train/theoretical_loss": 3.4586993699957187,
      "objective/train/tokens_used": 172830176,
      "theoretical_loss": 3.4586993699957187,
      "tokens_seen": 1813970944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580808080808081,
      "loss": 2.5818,
      "theoretical_loss": 3.4586993699957187,
      "tokens_seen": 1813970944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045806060606060604,
      "loss": 2.3931,
      "theoretical_loss": 3.4586887504532644,
      "tokens_seen": 1814036480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045804040404040405,
      "loss": 2.455,
      "theoretical_loss": 3.458678131401876,
      "tokens_seen": 1814102016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458020202020202,
      "loss": 2.5927,
      "theoretical_loss": 3.4586675128415125,
      "tokens_seen": 1814167552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458,
      "loss": 2.9255,
      "theoretical_loss": 3.4586568947721337,
      "tokens_seen": 1814233088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457979797979798,
      "loss": 2.5883,
      "theoretical_loss": 3.4586462771936994,
      "tokens_seen": 1814298624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457959595959596,
      "loss": 2.618,
      "theoretical_loss": 3.4586356601061694,
      "tokens_seen": 1814364160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045793939393939394,
      "loss": 2.4486,
      "theoretical_loss": 3.4586250435095027,
      "tokens_seen": 1814429696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579191919191919,
      "loss": 2.7142,
      "theoretical_loss": 3.4586144274036594,
      "tokens_seen": 1814495232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578989898989899,
      "loss": 2.4711,
      "theoretical_loss": 3.4586038117885987,
      "tokens_seen": 1814560768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045787878787878787,
      "loss": 2.5087,
      "theoretical_loss": 3.4585931966642804,
      "tokens_seen": 1814626304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578585858585859,
      "loss": 2.622,
      "theoretical_loss": 3.458582582030664,
      "tokens_seen": 1814691840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045783838383838384,
      "loss": 2.472,
      "theoretical_loss": 3.4585719678877087,
      "tokens_seen": 1814757376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578181818181818,
      "loss": 2.8238,
      "theoretical_loss": 3.458561354235375,
      "tokens_seen": 1814822912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577979797979798,
      "loss": 2.4024,
      "theoretical_loss": 3.458550741073622,
      "tokens_seen": 1814888448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577777777777778,
      "loss": 2.6186,
      "theoretical_loss": 3.458540128402409,
      "tokens_seen": 1814953984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045775757575757577,
      "loss": 2.609,
      "theoretical_loss": 3.458529516221696,
      "tokens_seen": 1815019520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045773737373737373,
      "loss": 2.565,
      "theoretical_loss": 3.4585189045314424,
      "tokens_seen": 1815085056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045771717171717174,
      "loss": 2.7391,
      "theoretical_loss": 3.458508293331609,
      "tokens_seen": 1815150592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576969696969697,
      "loss": 2.5499,
      "theoretical_loss": 3.458497682622153,
      "tokens_seen": 1815216128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576767676767677,
      "loss": 2.7469,
      "theoretical_loss": 3.4584870724030363,
      "tokens_seen": 1815281664
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045765656565656567,
      "loss": 2.6686,
      "theoretical_loss": 3.4584764626742173,
      "tokens_seen": 1815347200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576363636363636,
      "loss": 2.6245,
      "theoretical_loss": 3.4584658534356567,
      "tokens_seen": 1815412736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045761616161616163,
      "loss": 2.6962,
      "theoretical_loss": 3.458455244687313,
      "tokens_seen": 1815478272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045759595959595964,
      "loss": 2.6647,
      "theoretical_loss": 3.458444636429146,
      "tokens_seen": 1815543808
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 1000504,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.835871458053589,
      "objective/train/theoretical_loss": 3.458434028661116,
      "objective/train/tokens_used": 174468576,
      "theoretical_loss": 3.458434028661116,
      "tokens_seen": 1815609344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575757575757576,
      "loss": 2.7497,
      "theoretical_loss": 3.458434028661116,
      "tokens_seen": 1815609344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045755555555555556,
      "loss": 2.6998,
      "theoretical_loss": 3.4584234213831824,
      "tokens_seen": 1815674880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575353535353535,
      "loss": 2.5959,
      "theoretical_loss": 3.458412814595305,
      "tokens_seen": 1815740416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575151515151515,
      "loss": 2.6133,
      "theoretical_loss": 3.458402208297443,
      "tokens_seen": 1815805952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045749494949494954,
      "loss": 2.6405,
      "theoretical_loss": 3.4583916024895567,
      "tokens_seen": 1815871488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574747474747475,
      "loss": 2.7544,
      "theoretical_loss": 3.458380997171605,
      "tokens_seen": 1815937024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045745454545454545,
      "loss": 2.6652,
      "theoretical_loss": 3.4583703923435483,
      "tokens_seen": 1816002560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574343434343434,
      "loss": 2.7187,
      "theoretical_loss": 3.458359788005346,
      "tokens_seen": 1816068096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574141414141415,
      "loss": 2.6787,
      "theoretical_loss": 3.458349184156958,
      "tokens_seen": 1816133632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045739393939393943,
      "loss": 2.6822,
      "theoretical_loss": 3.458338580798343,
      "tokens_seen": 1816199168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004573737373737374,
      "loss": 2.6664,
      "theoretical_loss": 3.4583279779294624,
      "tokens_seen": 1816264704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045735353535353534,
      "loss": 2.6565,
      "theoretical_loss": 3.458317375550275,
      "tokens_seen": 1816330240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004573333333333333,
      "loss": 2.6583,
      "theoretical_loss": 3.4583067736607402,
      "tokens_seen": 1816395776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045731313131313137,
      "loss": 2.9378,
      "theoretical_loss": 3.4582961722608188,
      "tokens_seen": 1816461312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572929292929293,
      "loss": 2.6101,
      "theoretical_loss": 3.458285571350469,
      "tokens_seen": 1816526848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572727272727273,
      "loss": 2.8678,
      "theoretical_loss": 3.4582749709296516,
      "tokens_seen": 1816592384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045725252525252524,
      "loss": 2.5882,
      "theoretical_loss": 3.458264370998326,
      "tokens_seen": 1816657920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045723232323232325,
      "loss": 2.5968,
      "theoretical_loss": 3.4582537715564525,
      "tokens_seen": 1816723456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045721212121212126,
      "loss": 2.7255,
      "theoretical_loss": 3.45824317260399,
      "tokens_seen": 1816788992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571919191919192,
      "loss": 2.7617,
      "theoretical_loss": 3.458232574140899,
      "tokens_seen": 1816854528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045717171717171717,
      "loss": 2.4575,
      "theoretical_loss": 3.4582219761671382,
      "tokens_seen": 1816920064
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045715151515151513,
      "loss": 2.66,
      "theoretical_loss": 3.4582113786826683,
      "tokens_seen": 1816985600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045713131313131314,
      "loss": 2.527,
      "theoretical_loss": 3.4582007816874487,
      "tokens_seen": 1817051136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045711111111111115,
      "loss": 2.4134,
      "theoretical_loss": 3.4581901851814396,
      "tokens_seen": 1817116672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004570909090909091,
      "loss": 2.8572,
      "theoretical_loss": 3.4581795891646,
      "tokens_seen": 1817182208
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 1001884,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.327115058898926,
      "objective/train/theoretical_loss": 3.4581689936368907,
      "objective/train/tokens_used": 176106976,
      "theoretical_loss": 3.4581689936368907,
      "tokens_seen": 1817247744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045707070707070706,
      "loss": 2.5736,
      "theoretical_loss": 3.4581689936368907,
      "tokens_seen": 1817247744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570505050505051,
      "loss": 2.8025,
      "theoretical_loss": 3.45815839859827,
      "tokens_seen": 1817313280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045703030303030303,
      "loss": 2.6386,
      "theoretical_loss": 3.458147804048699,
      "tokens_seen": 1817378816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045701010101010104,
      "loss": 2.719,
      "theoretical_loss": 3.458137209988138,
      "tokens_seen": 1817444352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456989898989899,
      "loss": 2.5823,
      "theoretical_loss": 3.4581266164165445,
      "tokens_seen": 1817509888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045696969696969696,
      "loss": 2.7439,
      "theoretical_loss": 3.45811602333388,
      "tokens_seen": 1817575424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045694949494949497,
      "loss": 2.4384,
      "theoretical_loss": 3.4581054307401047,
      "tokens_seen": 1817640960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569292929292929,
      "loss": 2.8483,
      "theoretical_loss": 3.458094838635177,
      "tokens_seen": 1817706496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045690909090909094,
      "loss": 2.6419,
      "theoretical_loss": 3.458084247019057,
      "tokens_seen": 1817772032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568888888888889,
      "loss": 2.7798,
      "theoretical_loss": 3.458073655891706,
      "tokens_seen": 1817837568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568686868686869,
      "loss": 2.5721,
      "theoretical_loss": 3.458063065253082,
      "tokens_seen": 1817903104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045684848484848486,
      "loss": 2.8557,
      "theoretical_loss": 3.4580524751031456,
      "tokens_seen": 1817968640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568282828282828,
      "loss": 2.6035,
      "theoretical_loss": 3.4580418854418573,
      "tokens_seen": 1818034176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045680808080808083,
      "loss": 2.5388,
      "theoretical_loss": 3.4580312962691755,
      "tokens_seen": 1818099712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567878787878788,
      "loss": 2.5895,
      "theoretical_loss": 3.458020707585061,
      "tokens_seen": 1818165248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567676767676768,
      "loss": 2.4687,
      "theoretical_loss": 3.4580101193894732,
      "tokens_seen": 1818230784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045674747474747475,
      "loss": 2.461,
      "theoretical_loss": 3.457999531682373,
      "tokens_seen": 1818296320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567272727272727,
      "loss": 2.5961,
      "theoretical_loss": 3.4579889444637186,
      "tokens_seen": 1818361856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567070707070707,
      "loss": 2.5384,
      "theoretical_loss": 3.457978357733471,
      "tokens_seen": 1818427392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045668686868686873,
      "loss": 2.729,
      "theoretical_loss": 3.4579677714915897,
      "tokens_seen": 1818492928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566666666666667,
      "loss": 2.8327,
      "theoretical_loss": 3.457957185738035,
      "tokens_seen": 1818558464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045664646464646465,
      "loss": 2.4631,
      "theoretical_loss": 3.4579466004727664,
      "tokens_seen": 1818624000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566262626262626,
      "loss": 2.7424,
      "theoretical_loss": 3.4579360156957435,
      "tokens_seen": 1818689536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566060606060606,
      "loss": 2.7302,
      "theoretical_loss": 3.4579254314069265,
      "tokens_seen": 1818755072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565858585858586,
      "loss": 2.8242,
      "theoretical_loss": 3.4579148476062755,
      "tokens_seen": 1818820608
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1002311,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1334288120269775,
      "objective/train/theoretical_loss": 3.4579042642937505,
      "objective/train/tokens_used": 177745376,
      "theoretical_loss": 3.4579042642937505,
      "tokens_seen": 1818886144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565656565656566,
      "loss": 2.9767,
      "theoretical_loss": 3.4579042642937505,
      "tokens_seen": 1818886144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045654545454545454,
      "loss": 2.3271,
      "theoretical_loss": 3.4578936814693106,
      "tokens_seen": 1818951680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565252525252525,
      "loss": 2.7679,
      "theoretical_loss": 3.4578830991329164,
      "tokens_seen": 1819017216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565050505050505,
      "loss": 2.4984,
      "theoretical_loss": 3.4578725172845273,
      "tokens_seen": 1819082752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564848484848485,
      "loss": 2.6605,
      "theoretical_loss": 3.457861935924104,
      "tokens_seen": 1819148288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564646464646465,
      "loss": 2.582,
      "theoretical_loss": 3.4578513550516057,
      "tokens_seen": 1819213824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045644444444444443,
      "loss": 2.6528,
      "theoretical_loss": 3.457840774666993,
      "tokens_seen": 1819279360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564242424242424,
      "loss": 2.6102,
      "theoretical_loss": 3.457830194770225,
      "tokens_seen": 1819344896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045640404040404045,
      "loss": 2.647,
      "theoretical_loss": 3.4578196153612617,
      "tokens_seen": 1819410432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563838383838384,
      "loss": 2.8564,
      "theoretical_loss": 3.457809036440064,
      "tokens_seen": 1819475968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045636363636363637,
      "loss": 2.581,
      "theoretical_loss": 3.4577984580065912,
      "tokens_seen": 1819541504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563434343434343,
      "loss": 2.8138,
      "theoretical_loss": 3.4577878800608026,
      "tokens_seen": 1819607040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563232323232323,
      "loss": 2.4587,
      "theoretical_loss": 3.4577773026026595,
      "tokens_seen": 1819672576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045630303030303035,
      "loss": 2.6125,
      "theoretical_loss": 3.457766725632121,
      "tokens_seen": 1819738112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562828282828283,
      "loss": 2.7351,
      "theoretical_loss": 3.457756149149147,
      "tokens_seen": 1819803648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045626262626262626,
      "loss": 2.8924,
      "theoretical_loss": 3.457745573153698,
      "tokens_seen": 1819869184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562424242424242,
      "loss": 2.6131,
      "theoretical_loss": 3.4577349976457334,
      "tokens_seen": 1819934720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045622222222222223,
      "loss": 2.7737,
      "theoretical_loss": 3.4577244226252137,
      "tokens_seen": 1820000256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045620202020202024,
      "loss": 2.6232,
      "theoretical_loss": 3.457713848092099,
      "tokens_seen": 1820065792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561818181818182,
      "loss": 2.7592,
      "theoretical_loss": 3.4577032740463483,
      "tokens_seen": 1820131328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045616161616161615,
      "loss": 2.7347,
      "theoretical_loss": 3.4576927004879225,
      "tokens_seen": 1820196864
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561414141414141,
      "loss": 2.6675,
      "theoretical_loss": 3.4576821274167813,
      "tokens_seen": 1820262400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561212121212121,
      "loss": 2.815,
      "theoretical_loss": 3.4576715548328845,
      "tokens_seen": 1820327936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045610101010101013,
      "loss": 2.8066,
      "theoretical_loss": 3.4576609827361926,
      "tokens_seen": 1820393472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560808080808081,
      "loss": 2.6566,
      "theoretical_loss": 3.4576504111266653,
      "tokens_seen": 1820459008
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1003206,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.652801990509033,
      "objective/train/theoretical_loss": 3.457639840004262,
      "objective/train/tokens_used": 179383776,
      "theoretical_loss": 3.457639840004262,
      "tokens_seen": 1820524544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045606060606060605,
      "loss": 2.7231,
      "theoretical_loss": 3.457639840004262,
      "tokens_seen": 1820524544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045604040404040406,
      "loss": 2.5342,
      "theoretical_loss": 3.4576292693689443,
      "tokens_seen": 1820590080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045602020202020207,
      "loss": 2.6447,
      "theoretical_loss": 3.4576186992206708,
      "tokens_seen": 1820655616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456,
      "loss": 2.5253,
      "theoretical_loss": 3.4576081295594023,
      "tokens_seen": 1820721152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455979797979798,
      "loss": 2.5932,
      "theoretical_loss": 3.457597560385098,
      "tokens_seen": 1820786688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045595959595959594,
      "loss": 2.4873,
      "theoretical_loss": 3.4575869916977195,
      "tokens_seen": 1820852224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045593939393939395,
      "loss": 2.7011,
      "theoretical_loss": 3.4575764234972253,
      "tokens_seen": 1820917760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045591919191919196,
      "loss": 2.6617,
      "theoretical_loss": 3.4575658557835753,
      "tokens_seen": 1820983296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558989898989899,
      "loss": 2.7763,
      "theoretical_loss": 3.457555288556731,
      "tokens_seen": 1821048832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558787878787879,
      "loss": 2.813,
      "theoretical_loss": 3.4575447218166517,
      "tokens_seen": 1821114368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558585858585859,
      "loss": 2.7815,
      "theoretical_loss": 3.4575341555632972,
      "tokens_seen": 1821179904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045583838383838384,
      "loss": 2.755,
      "theoretical_loss": 3.4575235897966277,
      "tokens_seen": 1821245440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045581818181818185,
      "loss": 2.4572,
      "theoretical_loss": 3.457513024516604,
      "tokens_seen": 1821310976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557979797979798,
      "loss": 2.9435,
      "theoretical_loss": 3.457502459723185,
      "tokens_seen": 1821376512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045577777777777777,
      "loss": 2.6106,
      "theoretical_loss": 3.4574918954163314,
      "tokens_seen": 1821442048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557575757575758,
      "loss": 2.4757,
      "theoretical_loss": 3.4574813315960036,
      "tokens_seen": 1821507584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045573737373737374,
      "loss": 2.5581,
      "theoretical_loss": 3.4574707682621613,
      "tokens_seen": 1821573120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045571717171717175,
      "loss": 2.8554,
      "theoretical_loss": 3.4574602054147645,
      "tokens_seen": 1821638656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556969696969697,
      "loss": 2.6493,
      "theoretical_loss": 3.4574496430537733,
      "tokens_seen": 1821704192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556767676767677,
      "loss": 2.9531,
      "theoretical_loss": 3.4574390811791478,
      "tokens_seen": 1821769728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045565656565656567,
      "loss": 2.811,
      "theoretical_loss": 3.4574285197908483,
      "tokens_seen": 1821835264
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045563636363636363,
      "loss": 2.8732,
      "theoretical_loss": 3.4574179588888345,
      "tokens_seen": 1821900800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045561616161616164,
      "loss": 2.7586,
      "theoretical_loss": 3.4574073984730678,
      "tokens_seen": 1821966336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555959595959596,
      "loss": 2.6515,
      "theoretical_loss": 3.4573968385435068,
      "tokens_seen": 1822031872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555757575757576,
      "loss": 3.0081,
      "theoretical_loss": 3.4573862791001124,
      "tokens_seen": 1822097408
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1004309,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1638598442077637,
      "objective/train/theoretical_loss": 3.4573757201428443,
      "objective/train/tokens_used": 181022176,
      "theoretical_loss": 3.4573757201428443,
      "tokens_seen": 1822162944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045555555555555556,
      "loss": 2.8326,
      "theoretical_loss": 3.4573757201428443,
      "tokens_seen": 1822162944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555353535353535,
      "loss": 2.7456,
      "theoretical_loss": 3.457365161671663,
      "tokens_seen": 1822228480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045551515151515153,
      "loss": 2.6156,
      "theoretical_loss": 3.457354603686529,
      "tokens_seen": 1822294016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045549494949494954,
      "loss": 2.7395,
      "theoretical_loss": 3.457344046187401,
      "tokens_seen": 1822359552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554747474747475,
      "loss": 2.901,
      "theoretical_loss": 3.457333489174241,
      "tokens_seen": 1822425088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045545454545454546,
      "loss": 2.5662,
      "theoretical_loss": 3.4573229326470076,
      "tokens_seen": 1822490624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554343434343434,
      "loss": 2.5802,
      "theoretical_loss": 3.457312376605662,
      "tokens_seen": 1822556160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554141414141414,
      "loss": 2.6943,
      "theoretical_loss": 3.457301821050164,
      "tokens_seen": 1822621696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045539393939393944,
      "loss": 2.5732,
      "theoretical_loss": 3.4572912659804738,
      "tokens_seen": 1822687232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553737373737374,
      "loss": 2.783,
      "theoretical_loss": 3.4572807113965514,
      "tokens_seen": 1822752768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045535353535353535,
      "loss": 2.7854,
      "theoretical_loss": 3.4572701572983573,
      "tokens_seen": 1822818304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553333333333333,
      "loss": 2.3922,
      "theoretical_loss": 3.457259603685851,
      "tokens_seen": 1822883840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045531313131313137,
      "loss": 2.6926,
      "theoretical_loss": 3.4572490505589943,
      "tokens_seen": 1822949376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045529292929292933,
      "loss": 2.8008,
      "theoretical_loss": 3.4572384979177455,
      "tokens_seen": 1823014912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552727272727273,
      "loss": 2.3679,
      "theoretical_loss": 3.4572279457620656,
      "tokens_seen": 1823080448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045525252525252524,
      "loss": 2.6327,
      "theoretical_loss": 3.457217394091915,
      "tokens_seen": 1823145984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552323232323232,
      "loss": 2.7428,
      "theoretical_loss": 3.4572068429072536,
      "tokens_seen": 1823211520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045521212121212126,
      "loss": 2.7718,
      "theoretical_loss": 3.457196292208042,
      "tokens_seen": 1823277056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551919191919192,
      "loss": 2.8718,
      "theoretical_loss": 3.4571857419942402,
      "tokens_seen": 1823342592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551717171717172,
      "loss": 2.5766,
      "theoretical_loss": 3.457175192265808,
      "tokens_seen": 1823408128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045515151515151513,
      "loss": 2.8171,
      "theoretical_loss": 3.457164643022706,
      "tokens_seen": 1823473664
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551313131313131,
      "loss": 2.6179,
      "theoretical_loss": 3.4571540942648946,
      "tokens_seen": 1823539200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045511111111111116,
      "loss": 2.6814,
      "theoretical_loss": 3.4571435459923334,
      "tokens_seen": 1823604736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550909090909091,
      "loss": 2.4929,
      "theoretical_loss": 3.457132998204984,
      "tokens_seen": 1823670272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045507070707070707,
      "loss": 2.7795,
      "theoretical_loss": 3.457122450902805,
      "tokens_seen": 1823735808
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1004868,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2328946590423584,
      "objective/train/theoretical_loss": 3.4571119040857576,
      "objective/train/tokens_used": 182660576,
      "theoretical_loss": 3.4571119040857576,
      "tokens_seen": 1823801344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045505050505050503,
      "loss": 2.337,
      "theoretical_loss": 3.4571119040857576,
      "tokens_seen": 1823801344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045503030303030304,
      "loss": 2.9013,
      "theoretical_loss": 3.4571013577538015,
      "tokens_seen": 1823866880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045501010101010105,
      "loss": 2.7025,
      "theoretical_loss": 3.457090811906898,
      "tokens_seen": 1823932416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454989898989899,
      "loss": 2.9696,
      "theoretical_loss": 3.457080266545006,
      "tokens_seen": 1823997952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045496969696969696,
      "loss": 2.6379,
      "theoretical_loss": 3.457069721668087,
      "tokens_seen": 1824063488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549494949494949,
      "loss": 2.7066,
      "theoretical_loss": 3.4570591772761006,
      "tokens_seen": 1824129024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045492929292929293,
      "loss": 2.756,
      "theoretical_loss": 3.4570486333690065,
      "tokens_seen": 1824194560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045490909090909094,
      "loss": 2.7721,
      "theoretical_loss": 3.4570380899467663,
      "tokens_seen": 1824260096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548888888888889,
      "loss": 2.7186,
      "theoretical_loss": 3.4570275470093397,
      "tokens_seen": 1824325632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045486868686868686,
      "loss": 2.7226,
      "theoretical_loss": 3.457017004556687,
      "tokens_seen": 1824391168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045484848484848487,
      "loss": 2.6073,
      "theoretical_loss": 3.457006462588768,
      "tokens_seen": 1824456704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548282828282828,
      "loss": 2.5764,
      "theoretical_loss": 3.456995921105544,
      "tokens_seen": 1824522240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045480808080808083,
      "loss": 2.7873,
      "theoretical_loss": 3.4569853801069743,
      "tokens_seen": 1824587776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547878787878788,
      "loss": 2.6969,
      "theoretical_loss": 3.4569748395930198,
      "tokens_seen": 1824653312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045476767676767675,
      "loss": 2.7108,
      "theoretical_loss": 3.4569642995636407,
      "tokens_seen": 1824718848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045474747474747476,
      "loss": 2.8133,
      "theoretical_loss": 3.4569537600187976,
      "tokens_seen": 1824784384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547272727272727,
      "loss": 2.5824,
      "theoretical_loss": 3.4569432209584496,
      "tokens_seen": 1824849920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045470707070707073,
      "loss": 2.7374,
      "theoretical_loss": 3.4569326823825586,
      "tokens_seen": 1824915456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546868686868687,
      "loss": 2.4365,
      "theoretical_loss": 3.4569221442910845,
      "tokens_seen": 1824980992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546666666666667,
      "loss": 2.9807,
      "theoretical_loss": 3.4569116066839873,
      "tokens_seen": 1825046528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045464646464646465,
      "loss": 2.7103,
      "theoretical_loss": 3.4569010695612272,
      "tokens_seen": 1825112064
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546262626262626,
      "loss": 2.7703,
      "theoretical_loss": 3.456890532922765,
      "tokens_seen": 1825177600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546060606060606,
      "loss": 2.5735,
      "theoretical_loss": 3.456879996768561,
      "tokens_seen": 1825243136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545858585858586,
      "loss": 2.5735,
      "theoretical_loss": 3.4568694610985755,
      "tokens_seen": 1825308672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545656565656566,
      "loss": 2.5422,
      "theoretical_loss": 3.456858925912768,
      "tokens_seen": 1825374208
    },
    {
      "debugging/Self-BLEU-5": 0.5101952161435601,
      "debugging/distinct-1-grams": 0.7564583571701602,
      "debugging/distinct-2-grams": 0.9365913780414189,
      "debugging/entropy-1-grams": 5.753082263870947,
      "debugging/entropy-2-grams": 6.573169152318852,
      "debugging/length": 571.4,
      "debugging/num_segments": 10,
      "epoch": 0.1,
      "objective/train/docs_used": 1006237,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5774552822113037,
      "objective/train/theoretical_loss": 3.4568483912111008,
      "objective/train/tokens_used": 184298976,
      "theoretical_loss": 3.4568483912111008,
      "tokens_seen": 1825439744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045454545454545455,
      "loss": 2.8626,
      "theoretical_loss": 3.4568483912111008,
      "tokens_seen": 1825439744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545252525252525,
      "loss": 2.7059,
      "theoretical_loss": 3.4568378569935323,
      "tokens_seen": 1825505280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545050505050505,
      "loss": 2.4361,
      "theoretical_loss": 3.4568273232600237,
      "tokens_seen": 1825570816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544848484848485,
      "loss": 2.6962,
      "theoretical_loss": 3.456816790010536,
      "tokens_seen": 1825636352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544646464646465,
      "loss": 2.8448,
      "theoretical_loss": 3.4568062572450287,
      "tokens_seen": 1825701888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045444444444444444,
      "loss": 2.7327,
      "theoretical_loss": 3.456795724963462,
      "tokens_seen": 1825767424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544242424242424,
      "loss": 2.7395,
      "theoretical_loss": 3.456785193165797,
      "tokens_seen": 1825832960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544040404040404,
      "loss": 2.7404,
      "theoretical_loss": 3.456774661851995,
      "tokens_seen": 1825898496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543838383838384,
      "loss": 2.6062,
      "theoretical_loss": 3.4567641310220134,
      "tokens_seen": 1825964032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543636363636364,
      "loss": 2.7898,
      "theoretical_loss": 3.4567536006758157,
      "tokens_seen": 1826029568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045434343434343433,
      "loss": 2.6872,
      "theoretical_loss": 3.456743070813361,
      "tokens_seen": 1826095104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543232323232323,
      "loss": 2.5771,
      "theoretical_loss": 3.4567325414346093,
      "tokens_seen": 1826160640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045430303030303035,
      "loss": 2.8704,
      "theoretical_loss": 3.4567220125395215,
      "tokens_seen": 1826226176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542828282828283,
      "loss": 2.7221,
      "theoretical_loss": 3.456711484128059,
      "tokens_seen": 1826291712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045426262626262627,
      "loss": 2.7092,
      "theoretical_loss": 3.45670095620018,
      "tokens_seen": 1826357248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542424242424242,
      "loss": 2.6278,
      "theoretical_loss": 3.456690428755847,
      "tokens_seen": 1826422784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045422222222222223,
      "loss": 2.6251,
      "theoretical_loss": 3.4566799017950194,
      "tokens_seen": 1826488320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045420202020202025,
      "loss": 2.5685,
      "theoretical_loss": 3.4566693753176576,
      "tokens_seen": 1826553856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541818181818182,
      "loss": 2.5096,
      "theoretical_loss": 3.4566588493237234,
      "tokens_seen": 1826619392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045416161616161616,
      "loss": 2.6342,
      "theoretical_loss": 3.4566483238131753,
      "tokens_seen": 1826684928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541414141414141,
      "loss": 2.8359,
      "theoretical_loss": 3.456637798785975,
      "tokens_seen": 1826750464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541212121212122,
      "loss": 2.5948,
      "theoretical_loss": 3.4566272742420825,
      "tokens_seen": 1826816000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045410101010101014,
      "loss": 2.586,
      "theoretical_loss": 3.4566167501814586,
      "tokens_seen": 1826881536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540808080808081,
      "loss": 2.5007,
      "theoretical_loss": 3.4566062266040634,
      "tokens_seen": 1826947072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045406060606060605,
      "loss": 2.7515,
      "theoretical_loss": 3.4565957035098576,
      "tokens_seen": 1827012608
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1006887,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7742507457733154,
      "objective/train/theoretical_loss": 3.4565851808988013,
      "objective/train/tokens_used": 185937376,
      "theoretical_loss": 3.4565851808988013,
      "tokens_seen": 1827078144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454040404040404,
      "loss": 2.6239,
      "theoretical_loss": 3.4565851808988013,
      "tokens_seen": 1827078144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540202020202021,
      "loss": 2.68,
      "theoretical_loss": 3.4565746587708563,
      "tokens_seen": 1827143680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045400000000000003,
      "loss": 2.5783,
      "theoretical_loss": 3.456564137125981,
      "tokens_seen": 1827209216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453979797979798,
      "loss": 2.852,
      "theoretical_loss": 3.456553615964138,
      "tokens_seen": 1827274752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045395959595959594,
      "loss": 2.5331,
      "theoretical_loss": 3.4565430952852862,
      "tokens_seen": 1827340288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045393939393939396,
      "loss": 2.4804,
      "theoretical_loss": 3.456532575089387,
      "tokens_seen": 1827405824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045391919191919197,
      "loss": 2.7907,
      "theoretical_loss": 3.4565220553764004,
      "tokens_seen": 1827471360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538989898989899,
      "loss": 2.5951,
      "theoretical_loss": 3.4565115361462873,
      "tokens_seen": 1827536896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538787878787879,
      "loss": 2.5612,
      "theoretical_loss": 3.456501017399008,
      "tokens_seen": 1827602432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045385858585858584,
      "loss": 2.5128,
      "theoretical_loss": 3.4564904991345236,
      "tokens_seen": 1827667968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045383838383838385,
      "loss": 2.7096,
      "theoretical_loss": 3.4564799813527936,
      "tokens_seen": 1827733504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045381818181818186,
      "loss": 2.6215,
      "theoretical_loss": 3.4564694640537788,
      "tokens_seen": 1827799040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537979797979798,
      "loss": 2.6144,
      "theoretical_loss": 3.4564589472374405,
      "tokens_seen": 1827864576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537777777777778,
      "loss": 2.4467,
      "theoretical_loss": 3.4564484309037384,
      "tokens_seen": 1827930112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537575757575758,
      "loss": 2.8435,
      "theoretical_loss": 3.456437915052634,
      "tokens_seen": 1827995648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045373737373737374,
      "loss": 2.7307,
      "theoretical_loss": 3.456427399684087,
      "tokens_seen": 1828061184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045371717171717175,
      "loss": 2.4665,
      "theoretical_loss": 3.456416884798058,
      "tokens_seen": 1828126720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536969696969697,
      "loss": 2.6755,
      "theoretical_loss": 3.456406370394508,
      "tokens_seen": 1828192256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045367676767676767,
      "loss": 2.6087,
      "theoretical_loss": 3.456395856473397,
      "tokens_seen": 1828257792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536565656565657,
      "loss": 2.4562,
      "theoretical_loss": 3.456385343034686,
      "tokens_seen": 1828323328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045363636363636363,
      "loss": 2.5869,
      "theoretical_loss": 3.456374830078336,
      "tokens_seen": 1828388864
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045361616161616164,
      "loss": 2.6877,
      "theoretical_loss": 3.456364317604306,
      "tokens_seen": 1828454400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535959595959596,
      "loss": 2.7657,
      "theoretical_loss": 3.456353805612559,
      "tokens_seen": 1828519936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535757575757576,
      "loss": 2.8022,
      "theoretical_loss": 3.4563432941030534,
      "tokens_seen": 1828585472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045355555555555557,
      "loss": 2.6839,
      "theoretical_loss": 3.4563327830757506,
      "tokens_seen": 1828651008
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1008261,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3774466514587402,
      "objective/train/theoretical_loss": 3.4563222725306115,
      "objective/train/tokens_used": 187575776,
      "theoretical_loss": 3.4563222725306115,
      "tokens_seen": 1828716544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535353535353535,
      "loss": 2.6512,
      "theoretical_loss": 3.4563222725306115,
      "tokens_seen": 1828716544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045351515151515154,
      "loss": 2.8056,
      "theoretical_loss": 3.4563117624675965,
      "tokens_seen": 1828782080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534949494949495,
      "loss": 2.4355,
      "theoretical_loss": 3.456301252886666,
      "tokens_seen": 1828847616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534747474747475,
      "loss": 2.7577,
      "theoretical_loss": 3.456290743787781,
      "tokens_seen": 1828913152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045345454545454546,
      "loss": 2.6925,
      "theoretical_loss": 3.456280235170902,
      "tokens_seen": 1828978688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534343434343434,
      "loss": 2.8243,
      "theoretical_loss": 3.4562697270359886,
      "tokens_seen": 1829044224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045341414141414143,
      "loss": 2.7169,
      "theoretical_loss": 3.456259219383003,
      "tokens_seen": 1829109760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045339393939393944,
      "loss": 2.5994,
      "theoretical_loss": 3.456248712211905,
      "tokens_seen": 1829175296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533737373737374,
      "loss": 2.7856,
      "theoretical_loss": 3.4562382055226557,
      "tokens_seen": 1829240832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045335353535353535,
      "loss": 2.584,
      "theoretical_loss": 3.456227699315215,
      "tokens_seen": 1829306368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533333333333333,
      "loss": 2.7333,
      "theoretical_loss": 3.456217193589544,
      "tokens_seen": 1829371904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533131313131313,
      "loss": 2.5758,
      "theoretical_loss": 3.4562066883456035,
      "tokens_seen": 1829437440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045329292929292933,
      "loss": 2.6453,
      "theoretical_loss": 3.456196183583354,
      "tokens_seen": 1829502976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532727272727273,
      "loss": 2.7702,
      "theoretical_loss": 3.4561856793027568,
      "tokens_seen": 1829568512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045325252525252525,
      "loss": 2.65,
      "theoretical_loss": 3.4561751755037706,
      "tokens_seen": 1829634048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532323232323232,
      "loss": 2.7699,
      "theoretical_loss": 3.4561646721863584,
      "tokens_seen": 1829699584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532121212121212,
      "loss": 2.6576,
      "theoretical_loss": 3.4561541693504796,
      "tokens_seen": 1829765120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531919191919192,
      "loss": 2.5966,
      "theoretical_loss": 3.4561436669960948,
      "tokens_seen": 1829830656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531717171717172,
      "loss": 2.5938,
      "theoretical_loss": 3.4561331651231653,
      "tokens_seen": 1829896192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045315151515151514,
      "loss": 2.7886,
      "theoretical_loss": 3.4561226637316516,
      "tokens_seen": 1829961728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531313131313131,
      "loss": 2.7663,
      "theoretical_loss": 3.456112162821514,
      "tokens_seen": 1830027264
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045311111111111116,
      "loss": 2.7715,
      "theoretical_loss": 3.456101662392714,
      "tokens_seen": 1830092800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530909090909091,
      "loss": 2.5694,
      "theoretical_loss": 3.4560911624452113,
      "tokens_seen": 1830158336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530707070707071,
      "loss": 2.5395,
      "theoretical_loss": 3.456080662978967,
      "tokens_seen": 1830223872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045305050505050503,
      "loss": 2.4467,
      "theoretical_loss": 3.4560701639939424,
      "tokens_seen": 1830289408
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1009482,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.900181770324707,
      "objective/train/theoretical_loss": 3.4560596654900975,
      "objective/train/tokens_used": 189214176,
      "theoretical_loss": 3.4560596654900975,
      "tokens_seen": 1830354944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453030303030303,
      "loss": 2.6643,
      "theoretical_loss": 3.4560596654900975,
      "tokens_seen": 1830354944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045301010101010106,
      "loss": 2.6989,
      "theoretical_loss": 3.456049167467393,
      "tokens_seen": 1830420480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452989898989899,
      "loss": 2.6867,
      "theoretical_loss": 3.45603866992579,
      "tokens_seen": 1830486016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045296969696969697,
      "loss": 2.452,
      "theoretical_loss": 3.456028172865249,
      "tokens_seen": 1830551552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529494949494949,
      "loss": 2.9193,
      "theoretical_loss": 3.4560176762857315,
      "tokens_seen": 1830617088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045292929292929294,
      "loss": 2.5696,
      "theoretical_loss": 3.456007180187197,
      "tokens_seen": 1830682624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045290909090909095,
      "loss": 2.7878,
      "theoretical_loss": 3.455996684569607,
      "tokens_seen": 1830748160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528888888888889,
      "loss": 2.4411,
      "theoretical_loss": 3.4559861894329216,
      "tokens_seen": 1830813696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045286868686868686,
      "loss": 2.7437,
      "theoretical_loss": 3.4559756947771025,
      "tokens_seen": 1830879232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528484848484848,
      "loss": 2.7311,
      "theoretical_loss": 3.45596520060211,
      "tokens_seen": 1830944768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045282828282828283,
      "loss": 2.6225,
      "theoretical_loss": 3.4559547069079044,
      "tokens_seen": 1831010304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045280808080808084,
      "loss": 2.6697,
      "theoretical_loss": 3.455944213694447,
      "tokens_seen": 1831075840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004527878787878788,
      "loss": 2.5926,
      "theoretical_loss": 3.4559337209616987,
      "tokens_seen": 1831141376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045276767676767675,
      "loss": 2.7966,
      "theoretical_loss": 3.45592322870962,
      "tokens_seen": 1831206912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045274747474747477,
      "loss": 2.5301,
      "theoretical_loss": 3.4559127369381715,
      "tokens_seen": 1831272448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004527272727272727,
      "loss": 2.6866,
      "theoretical_loss": 3.4559022456473145,
      "tokens_seen": 1831337984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045270707070707073,
      "loss": 2.681,
      "theoretical_loss": 3.455891754837009,
      "tokens_seen": 1831403520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526868686868687,
      "loss": 2.3233,
      "theoretical_loss": 3.4558812645072168,
      "tokens_seen": 1831469056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045266666666666665,
      "loss": 2.6848,
      "theoretical_loss": 3.455870774657898,
      "tokens_seen": 1831534592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045264646464646466,
      "loss": 2.6255,
      "theoretical_loss": 3.4558602852890132,
      "tokens_seen": 1831600128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526262626262626,
      "loss": 2.7053,
      "theoretical_loss": 3.4558497964005244,
      "tokens_seen": 1831665664
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526060606060606,
      "loss": 2.6355,
      "theoretical_loss": 3.455839307992391,
      "tokens_seen": 1831731200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525858585858586,
      "loss": 2.7105,
      "theoretical_loss": 3.455828820064575,
      "tokens_seen": 1831796736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525656565656566,
      "loss": 2.8606,
      "theoretical_loss": 3.455818332617036,
      "tokens_seen": 1831862272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045254545454545455,
      "loss": 2.9341,
      "theoretical_loss": 3.455807845649735,
      "tokens_seen": 1831927808
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1009995,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7280309200286865,
      "objective/train/theoretical_loss": 3.4557973591626343,
      "objective/train/tokens_used": 190852576,
      "theoretical_loss": 3.4557973591626343,
      "tokens_seen": 1831993344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045252525252525256,
      "loss": 2.636,
      "theoretical_loss": 3.4557973591626343,
      "tokens_seen": 1831993344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525050505050505,
      "loss": 2.6817,
      "theoretical_loss": 3.4557868731556933,
      "tokens_seen": 1832058880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524848484848485,
      "loss": 2.5711,
      "theoretical_loss": 3.455776387628873,
      "tokens_seen": 1832124416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524646464646465,
      "loss": 2.5362,
      "theoretical_loss": 3.4557659025821352,
      "tokens_seen": 1832189952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045244444444444444,
      "loss": 2.7667,
      "theoretical_loss": 3.45575541801544,
      "tokens_seen": 1832255488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045242424242424245,
      "loss": 2.78,
      "theoretical_loss": 3.455744933928748,
      "tokens_seen": 1832321024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524040404040404,
      "loss": 2.6392,
      "theoretical_loss": 3.45573445032202,
      "tokens_seen": 1832386560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523838383838384,
      "loss": 2.5771,
      "theoretical_loss": 3.455723967195218,
      "tokens_seen": 1832452096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523636363636364,
      "loss": 2.3533,
      "theoretical_loss": 3.4557134845483013,
      "tokens_seen": 1832517632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045234343434343434,
      "loss": 2.5183,
      "theoretical_loss": 3.4557030023812323,
      "tokens_seen": 1832583168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045232323232323235,
      "loss": 2.7558,
      "theoretical_loss": 3.4556925206939706,
      "tokens_seen": 1832648704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523030303030303,
      "loss": 2.6891,
      "theoretical_loss": 3.455682039486478,
      "tokens_seen": 1832714240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522828282828283,
      "loss": 2.5169,
      "theoretical_loss": 3.455671558758715,
      "tokens_seen": 1832779776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045226262626262627,
      "loss": 2.6283,
      "theoretical_loss": 3.4556610785106425,
      "tokens_seen": 1832845312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045224242424242423,
      "loss": 2.7893,
      "theoretical_loss": 3.455650598742221,
      "tokens_seen": 1832910848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045222222222222224,
      "loss": 2.8487,
      "theoretical_loss": 3.4556401194534123,
      "tokens_seen": 1832976384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045220202020202025,
      "loss": 2.5464,
      "theoretical_loss": 3.4556296406441764,
      "tokens_seen": 1833041920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521818181818182,
      "loss": 2.7207,
      "theoretical_loss": 3.4556191623144747,
      "tokens_seen": 1833107456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045216161616161616,
      "loss": 2.5345,
      "theoretical_loss": 3.455608684464268,
      "tokens_seen": 1833172992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521414141414141,
      "loss": 2.5418,
      "theoretical_loss": 3.4555982070935176,
      "tokens_seen": 1833238528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045212121212121213,
      "loss": 2.877,
      "theoretical_loss": 3.4555877302021836,
      "tokens_seen": 1833304064
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045210101010101014,
      "loss": 2.577,
      "theoretical_loss": 3.455577253790228,
      "tokens_seen": 1833369600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520808080808081,
      "loss": 2.6027,
      "theoretical_loss": 3.4555667778576105,
      "tokens_seen": 1833435136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045206060606060606,
      "loss": 2.8754,
      "theoretical_loss": 3.455556302404293,
      "tokens_seen": 1833500672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452040404040404,
      "loss": 2.8005,
      "theoretical_loss": 3.4555458274302353,
      "tokens_seen": 1833566208
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 1010832,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6558239459991455,
      "objective/train/theoretical_loss": 3.4555353529354003,
      "objective/train/tokens_used": 192490976,
      "theoretical_loss": 3.4555353529354003,
      "tokens_seen": 1833631744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520202020202021,
      "loss": 2.4985,
      "theoretical_loss": 3.4555353529354003,
      "tokens_seen": 1833631744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045200000000000004,
      "loss": 2.7386,
      "theoretical_loss": 3.4555248789197472,
      "tokens_seen": 1833697280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451979797979798,
      "loss": 2.6204,
      "theoretical_loss": 3.455514405383237,
      "tokens_seen": 1833762816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045195959595959595,
      "loss": 2.7539,
      "theoretical_loss": 3.455503932325832,
      "tokens_seen": 1833828352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519393939393939,
      "loss": 2.7639,
      "theoretical_loss": 3.4554934597474922,
      "tokens_seen": 1833893888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045191919191919197,
      "loss": 2.6223,
      "theoretical_loss": 3.4554829876481787,
      "tokens_seen": 1833959424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045189898989898993,
      "loss": 2.7164,
      "theoretical_loss": 3.4554725160278528,
      "tokens_seen": 1834024960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518787878787879,
      "loss": 2.3732,
      "theoretical_loss": 3.4554620448864743,
      "tokens_seen": 1834090496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045185858585858584,
      "loss": 2.6354,
      "theoretical_loss": 3.4554515742240057,
      "tokens_seen": 1834156032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518383838383838,
      "loss": 2.6698,
      "theoretical_loss": 3.4554411040404074,
      "tokens_seen": 1834221568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045181818181818186,
      "loss": 2.6226,
      "theoretical_loss": 3.45543063433564,
      "tokens_seen": 1834287104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004517979797979798,
      "loss": 2.6863,
      "theoretical_loss": 3.4554201651096648,
      "tokens_seen": 1834352640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004517777777777778,
      "loss": 2.6369,
      "theoretical_loss": 3.455409696362443,
      "tokens_seen": 1834418176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045175757575757574,
      "loss": 2.5681,
      "theoretical_loss": 3.4553992280939356,
      "tokens_seen": 1834483712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045173737373737375,
      "loss": 2.6862,
      "theoretical_loss": 3.4553887603041034,
      "tokens_seen": 1834549248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045171717171717176,
      "loss": 2.5591,
      "theoretical_loss": 3.4553782929929073,
      "tokens_seen": 1834614784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516969696969697,
      "loss": 2.6327,
      "theoretical_loss": 3.4553678261603085,
      "tokens_seen": 1834680320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045167676767676767,
      "loss": 2.6442,
      "theoretical_loss": 3.455357359806268,
      "tokens_seen": 1834745856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045165656565656563,
      "loss": 2.6768,
      "theoretical_loss": 3.455346893930747,
      "tokens_seen": 1834811392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045163636363636364,
      "loss": 2.4703,
      "theoretical_loss": 3.4553364285337063,
      "tokens_seen": 1834876928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045161616161616165,
      "loss": 2.7648,
      "theoretical_loss": 3.4553259636151066,
      "tokens_seen": 1834942464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515959595959596,
      "loss": 2.6158,
      "theoretical_loss": 3.45531549917491,
      "tokens_seen": 1835008000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045157575757575756,
      "loss": 2.8335,
      "theoretical_loss": 3.455305035213076,
      "tokens_seen": 1835073536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515555555555556,
      "loss": 2.7982,
      "theoretical_loss": 3.455294571729567,
      "tokens_seen": 1835139072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045153535353535353,
      "loss": 2.8067,
      "theoretical_loss": 3.455284108724344,
      "tokens_seen": 1835204608
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1011110,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.309302806854248,
      "objective/train/theoretical_loss": 3.4552736461973668,
      "objective/train/tokens_used": 194129376,
      "theoretical_loss": 3.4552736461973668,
      "tokens_seen": 1835270144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045151515151515154,
      "loss": 2.5007,
      "theoretical_loss": 3.4552736461973668,
      "tokens_seen": 1835270144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514949494949495,
      "loss": 2.5452,
      "theoretical_loss": 3.455263184148598,
      "tokens_seen": 1835335680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045147474747474746,
      "loss": 2.6689,
      "theoretical_loss": 3.455252722577997,
      "tokens_seen": 1835401216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045145454545454547,
      "loss": 2.6253,
      "theoretical_loss": 3.455242261485527,
      "tokens_seen": 1835466752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514343434343434,
      "loss": 2.4733,
      "theoretical_loss": 3.4552318008711476,
      "tokens_seen": 1835532288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045141414141414144,
      "loss": 2.4469,
      "theoretical_loss": 3.45522134073482,
      "tokens_seen": 1835597824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513939393939394,
      "loss": 2.7587,
      "theoretical_loss": 3.455210881076505,
      "tokens_seen": 1835663360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513737373737374,
      "loss": 2.5127,
      "theoretical_loss": 3.455200421896165,
      "tokens_seen": 1835728896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045135353535353536,
      "loss": 2.8253,
      "theoretical_loss": 3.45518996319376,
      "tokens_seen": 1835794432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513333333333333,
      "loss": 2.5292,
      "theoretical_loss": 3.4551795049692515,
      "tokens_seen": 1835859968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045131313131313133,
      "loss": 2.7594,
      "theoretical_loss": 3.4551690472226007,
      "tokens_seen": 1835925504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512929292929293,
      "loss": 2.6391,
      "theoretical_loss": 3.455158589953768,
      "tokens_seen": 1835991040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512727272727273,
      "loss": 2.6793,
      "theoretical_loss": 3.455148133162715,
      "tokens_seen": 1836056576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045125252525252525,
      "loss": 2.7743,
      "theoretical_loss": 3.455137676849403,
      "tokens_seen": 1836122112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512323232323232,
      "loss": 2.7776,
      "theoretical_loss": 3.455127221013793,
      "tokens_seen": 1836187648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512121212121212,
      "loss": 2.7312,
      "theoretical_loss": 3.4551167656558457,
      "tokens_seen": 1836253184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045119191919191923,
      "loss": 2.8942,
      "theoretical_loss": 3.4551063107755233,
      "tokens_seen": 1836318720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511717171717172,
      "loss": 2.6964,
      "theoretical_loss": 3.4550958563727856,
      "tokens_seen": 1836384256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045115151515151515,
      "loss": 2.4926,
      "theoretical_loss": 3.4550854024475948,
      "tokens_seen": 1836449792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511313131313131,
      "loss": 2.8051,
      "theoretical_loss": 3.4550749489999113,
      "tokens_seen": 1836515328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511111111111111,
      "loss": 2.7998,
      "theoretical_loss": 3.4550644960296966,
      "tokens_seen": 1836580864
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510909090909091,
      "loss": 2.4606,
      "theoretical_loss": 3.455054043536912,
      "tokens_seen": 1836646400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510707070707071,
      "loss": 2.5874,
      "theoretical_loss": 3.455043591521518,
      "tokens_seen": 1836711936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045105050505050504,
      "loss": 2.5794,
      "theoretical_loss": 3.4550331399834766,
      "tokens_seen": 1836777472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451030303030303,
      "loss": 2.7082,
      "theoretical_loss": 3.4550226889227487,
      "tokens_seen": 1836843008
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1012365,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6850903034210205,
      "objective/train/theoretical_loss": 3.455012238339296,
      "objective/train/tokens_used": 195767776,
      "theoretical_loss": 3.455012238339296,
      "tokens_seen": 1836908544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045101010101010106,
      "loss": 2.617,
      "theoretical_loss": 3.455012238339296,
      "tokens_seen": 1836908544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000450989898989899,
      "loss": 2.8187,
      "theoretical_loss": 3.455001788233078,
      "tokens_seen": 1836974080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000450969696969697,
      "loss": 2.7156,
      "theoretical_loss": 3.4549913386040574,
      "tokens_seen": 1837039616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045094949494949493,
      "loss": 2.5967,
      "theoretical_loss": 3.454980889452195,
      "tokens_seen": 1837105152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509292929292929,
      "loss": 2.4446,
      "theoretical_loss": 3.4549704407774517,
      "tokens_seen": 1837170688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045090909090909095,
      "loss": 2.6417,
      "theoretical_loss": 3.454959992579789,
      "tokens_seen": 1837236224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508888888888889,
      "loss": 2.592,
      "theoretical_loss": 3.4549495448591685,
      "tokens_seen": 1837301760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045086868686868687,
      "loss": 2.8509,
      "theoretical_loss": 3.4549390976155503,
      "tokens_seen": 1837367296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508484848484848,
      "loss": 2.6198,
      "theoretical_loss": 3.4549286508488963,
      "tokens_seen": 1837432832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508282828282829,
      "loss": 2.4963,
      "theoretical_loss": 3.454918204559168,
      "tokens_seen": 1837498368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045080808080808085,
      "loss": 2.5869,
      "theoretical_loss": 3.4549077587463257,
      "tokens_seen": 1837563904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507878787878788,
      "loss": 2.6174,
      "theoretical_loss": 3.4548973134103322,
      "tokens_seen": 1837629440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045076767676767676,
      "loss": 2.6871,
      "theoretical_loss": 3.4548868685511467,
      "tokens_seen": 1837694976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507474747474747,
      "loss": 2.8478,
      "theoretical_loss": 3.4548764241687318,
      "tokens_seen": 1837760512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507272727272728,
      "loss": 2.5321,
      "theoretical_loss": 3.4548659802630484,
      "tokens_seen": 1837826048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045070707070707074,
      "loss": 2.7471,
      "theoretical_loss": 3.454855536834058,
      "tokens_seen": 1837891584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506868686868687,
      "loss": 2.8312,
      "theoretical_loss": 3.4548450938817212,
      "tokens_seen": 1837957120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045066666666666665,
      "loss": 2.5958,
      "theoretical_loss": 3.454834651406,
      "tokens_seen": 1838022656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045064646464646466,
      "loss": 2.6896,
      "theoretical_loss": 3.454824209406855,
      "tokens_seen": 1838088192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506262626262627,
      "loss": 2.7956,
      "theoretical_loss": 3.4548137678842474,
      "tokens_seen": 1838153728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045060606060606063,
      "loss": 2.626,
      "theoretical_loss": 3.4548033268381397,
      "tokens_seen": 1838219264
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505858585858586,
      "loss": 2.7092,
      "theoretical_loss": 3.454792886268492,
      "tokens_seen": 1838284800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045056565656565655,
      "loss": 2.8204,
      "theoretical_loss": 3.4547824461752654,
      "tokens_seen": 1838350336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045054545454545456,
      "loss": 2.9183,
      "theoretical_loss": 3.4547720065584215,
      "tokens_seen": 1838415872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045052525252525257,
      "loss": 2.7341,
      "theoretical_loss": 3.454761567417922,
      "tokens_seen": 1838481408
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1013519,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.807917356491089,
      "objective/train/theoretical_loss": 3.454751128753728,
      "objective/train/tokens_used": 197406176,
      "theoretical_loss": 3.454751128753728,
      "tokens_seen": 1838546944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505050505050505,
      "loss": 2.8162,
      "theoretical_loss": 3.454751128753728,
      "tokens_seen": 1838546944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504848484848485,
      "loss": 2.6302,
      "theoretical_loss": 3.4547406905658002,
      "tokens_seen": 1838612480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504646464646465,
      "loss": 2.6506,
      "theoretical_loss": 3.454730252854101,
      "tokens_seen": 1838678016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045044444444444445,
      "loss": 2.7298,
      "theoretical_loss": 3.454719815618591,
      "tokens_seen": 1838743552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045042424242424246,
      "loss": 2.7674,
      "theoretical_loss": 3.4547093788592314,
      "tokens_seen": 1838809088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504040404040404,
      "loss": 2.5375,
      "theoretical_loss": 3.4546989425759835,
      "tokens_seen": 1838874624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503838383838384,
      "loss": 2.5675,
      "theoretical_loss": 3.454688506768809,
      "tokens_seen": 1838940160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503636363636364,
      "loss": 2.8516,
      "theoretical_loss": 3.4546780714376695,
      "tokens_seen": 1839005696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045034343434343434,
      "loss": 2.5366,
      "theoretical_loss": 3.454667636582525,
      "tokens_seen": 1839071232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045032323232323235,
      "loss": 2.511,
      "theoretical_loss": 3.4546572022033377,
      "tokens_seen": 1839136768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503030303030303,
      "loss": 2.8897,
      "theoretical_loss": 3.4546467683000692,
      "tokens_seen": 1839202304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502828282828283,
      "loss": 2.7794,
      "theoretical_loss": 3.4546363348726805,
      "tokens_seen": 1839267840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502626262626263,
      "loss": 2.832,
      "theoretical_loss": 3.4546259019211334,
      "tokens_seen": 1839333376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045024242424242423,
      "loss": 2.6955,
      "theoretical_loss": 3.4546154694453883,
      "tokens_seen": 1839398912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045022222222222225,
      "loss": 2.6539,
      "theoretical_loss": 3.454605037445407,
      "tokens_seen": 1839464448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502020202020202,
      "loss": 2.4324,
      "theoretical_loss": 3.454594605921151,
      "tokens_seen": 1839529984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501818181818182,
      "loss": 2.6089,
      "theoretical_loss": 3.454584174872582,
      "tokens_seen": 1839595520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045016161616161617,
      "loss": 2.7772,
      "theoretical_loss": 3.454573744299661,
      "tokens_seen": 1839661056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501414141414141,
      "loss": 2.7491,
      "theoretical_loss": 3.454563314202349,
      "tokens_seen": 1839726592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045012121212121214,
      "loss": 2.7735,
      "theoretical_loss": 3.454552884580607,
      "tokens_seen": 1839792128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045010101010101015,
      "loss": 2.666,
      "theoretical_loss": 3.4545424554343978,
      "tokens_seen": 1839857664
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500808080808081,
      "loss": 2.7511,
      "theoretical_loss": 3.454532026763682,
      "tokens_seen": 1839923200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045006060606060606,
      "loss": 2.4584,
      "theoretical_loss": 3.4545215985684212,
      "tokens_seen": 1839988736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000450040404040404,
      "loss": 2.7286,
      "theoretical_loss": 3.4545111708485763,
      "tokens_seen": 1840054272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045002020202020203,
      "loss": 2.8811,
      "theoretical_loss": 3.454500743604109,
      "tokens_seen": 1840119808
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1014020,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2581570148468018,
      "objective/train/theoretical_loss": 3.4544903168349808,
      "objective/train/tokens_used": 199044576,
      "theoretical_loss": 3.4544903168349808,
      "tokens_seen": 1840185344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045000000000000004,
      "loss": 2.6365,
      "theoretical_loss": 3.4544903168349808,
      "tokens_seen": 1840185344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449979797979798,
      "loss": 2.6191,
      "theoretical_loss": 3.454479890541153,
      "tokens_seen": 1840250880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044995959595959596,
      "loss": 2.6941,
      "theoretical_loss": 3.454469464722587,
      "tokens_seen": 1840316416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004499393939393939,
      "loss": 2.8638,
      "theoretical_loss": 3.4544590393792443,
      "tokens_seen": 1840381952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004499191919191919,
      "loss": 2.5423,
      "theoretical_loss": 3.4544486145110858,
      "tokens_seen": 1840447488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044989898989898993,
      "loss": 2.8338,
      "theoretical_loss": 3.454438190118074,
      "tokens_seen": 1840513024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498787878787879,
      "loss": 2.854,
      "theoretical_loss": 3.4544277662001694,
      "tokens_seen": 1840578560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044985858585858585,
      "loss": 2.5746,
      "theoretical_loss": 3.454417342757334,
      "tokens_seen": 1840644096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498383838383838,
      "loss": 2.5932,
      "theoretical_loss": 3.4544069197895286,
      "tokens_seen": 1840709632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044981818181818187,
      "loss": 2.7428,
      "theoretical_loss": 3.454396497296715,
      "tokens_seen": 1840775168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044979797979797983,
      "loss": 2.5032,
      "theoretical_loss": 3.454386075278855,
      "tokens_seen": 1840840704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497777777777778,
      "loss": 2.6914,
      "theoretical_loss": 3.4543756537359096,
      "tokens_seen": 1840906240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044975757575757574,
      "loss": 2.4781,
      "theoretical_loss": 3.4543652326678402,
      "tokens_seen": 1840971776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497373737373737,
      "loss": 2.8693,
      "theoretical_loss": 3.454354812074609,
      "tokens_seen": 1841037312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044971717171717176,
      "loss": 2.5425,
      "theoretical_loss": 3.454344391956176,
      "tokens_seen": 1841102848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496969696969697,
      "loss": 2.6198,
      "theoretical_loss": 3.4543339723125035,
      "tokens_seen": 1841168384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496767676767677,
      "loss": 2.6949,
      "theoretical_loss": 3.4543235531435537,
      "tokens_seen": 1841233920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044965656565656563,
      "loss": 2.6307,
      "theoretical_loss": 3.4543131344492872,
      "tokens_seen": 1841299456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044963636363636364,
      "loss": 2.632,
      "theoretical_loss": 3.454302716229665,
      "tokens_seen": 1841364992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044961616161616166,
      "loss": 2.7287,
      "theoretical_loss": 3.45429229848465,
      "tokens_seen": 1841430528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495959595959596,
      "loss": 2.521,
      "theoretical_loss": 3.454281881214203,
      "tokens_seen": 1841496064
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044957575757575757,
      "loss": 2.6339,
      "theoretical_loss": 3.454271464418285,
      "tokens_seen": 1841561600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495555555555555,
      "loss": 2.7453,
      "theoretical_loss": 3.454261048096858,
      "tokens_seen": 1841627136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044953535353535354,
      "loss": 2.5324,
      "theoretical_loss": 3.4542506322498836,
      "tokens_seen": 1841692672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044951515151515155,
      "loss": 2.5193,
      "theoretical_loss": 3.454240216877323,
      "tokens_seen": 1841758208
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1015263,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0066089630126953,
      "objective/train/theoretical_loss": 3.454229801979138,
      "objective/train/tokens_used": 200682976,
      "theoretical_loss": 3.454229801979138,
      "tokens_seen": 1841823744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494949494949495,
      "loss": 2.759,
      "theoretical_loss": 3.454229801979138,
      "tokens_seen": 1841823744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044947474747474746,
      "loss": 2.5906,
      "theoretical_loss": 3.45421938755529,
      "tokens_seen": 1841889280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494545454545455,
      "loss": 2.5523,
      "theoretical_loss": 3.4542089736057404,
      "tokens_seen": 1841954816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044943434343434343,
      "loss": 2.7433,
      "theoretical_loss": 3.4541985601304503,
      "tokens_seen": 1842020352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044941414141414144,
      "loss": 2.6943,
      "theoretical_loss": 3.4541881471293827,
      "tokens_seen": 1842085888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493939393939394,
      "loss": 2.7566,
      "theoretical_loss": 3.4541777346024976,
      "tokens_seen": 1842151424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044937373737373735,
      "loss": 2.6265,
      "theoretical_loss": 3.454167322549757,
      "tokens_seen": 1842216960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044935353535353537,
      "loss": 2.7595,
      "theoretical_loss": 3.4541569109711228,
      "tokens_seen": 1842282496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493333333333333,
      "loss": 2.3543,
      "theoretical_loss": 3.454146499866556,
      "tokens_seen": 1842348032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044931313131313133,
      "loss": 2.6629,
      "theoretical_loss": 3.4541360892360187,
      "tokens_seen": 1842413568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492929292929293,
      "loss": 2.6116,
      "theoretical_loss": 3.4541256790794725,
      "tokens_seen": 1842479104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492727272727273,
      "loss": 2.6179,
      "theoretical_loss": 3.4541152693968784,
      "tokens_seen": 1842544640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044925252525252526,
      "loss": 2.8316,
      "theoretical_loss": 3.454104860188198,
      "tokens_seen": 1842610176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492323232323232,
      "loss": 2.542,
      "theoretical_loss": 3.4540944514533933,
      "tokens_seen": 1842675712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492121212121212,
      "loss": 2.6138,
      "theoretical_loss": 3.454084043192426,
      "tokens_seen": 1842741248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491919191919192,
      "loss": 2.7504,
      "theoretical_loss": 3.4540736354052575,
      "tokens_seen": 1842806784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491717171717172,
      "loss": 2.663,
      "theoretical_loss": 3.454063228091848,
      "tokens_seen": 1842872320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044915151515151515,
      "loss": 2.6304,
      "theoretical_loss": 3.4540528212521617,
      "tokens_seen": 1842937856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044913131313131316,
      "loss": 2.5554,
      "theoretical_loss": 3.4540424148861586,
      "tokens_seen": 1843003392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491111111111111,
      "loss": 2.6124,
      "theoretical_loss": 3.4540320089938,
      "tokens_seen": 1843068928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044909090909090913,
      "loss": 2.8087,
      "theoretical_loss": 3.454021603575048,
      "tokens_seen": 1843134464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490707070707071,
      "loss": 2.9358,
      "theoretical_loss": 3.454011198629865,
      "tokens_seen": 1843200000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044905050505050504,
      "loss": 2.5626,
      "theoretical_loss": 3.454000794158211,
      "tokens_seen": 1843265536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044903030303030306,
      "loss": 2.4122,
      "theoretical_loss": 3.453990390160049,
      "tokens_seen": 1843331072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449010101010101,
      "loss": 2.8471,
      "theoretical_loss": 3.4539799866353396,
      "tokens_seen": 1843396608
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1015947,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9678730964660645,
      "objective/train/theoretical_loss": 3.453969583584045,
      "objective/train/tokens_used": 202321376,
      "theoretical_loss": 3.453969583584045,
      "tokens_seen": 1843462144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448989898989899,
      "loss": 2.7793,
      "theoretical_loss": 3.453969583584045,
      "tokens_seen": 1843462144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448969696969697,
      "loss": 2.5612,
      "theoretical_loss": 3.4539591810061268,
      "tokens_seen": 1843527680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044894949494949494,
      "loss": 2.6931,
      "theoretical_loss": 3.453948778901547,
      "tokens_seen": 1843593216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044892929292929295,
      "loss": 2.5578,
      "theoretical_loss": 3.4539383772702665,
      "tokens_seen": 1843658752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044890909090909096,
      "loss": 2.6837,
      "theoretical_loss": 3.4539279761122472,
      "tokens_seen": 1843724288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488888888888889,
      "loss": 2.615,
      "theoretical_loss": 3.4539175754274507,
      "tokens_seen": 1843789824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044886868686868687,
      "loss": 2.6294,
      "theoretical_loss": 3.453907175215839,
      "tokens_seen": 1843855360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044884848484848483,
      "loss": 2.5685,
      "theoretical_loss": 3.4538967754773733,
      "tokens_seen": 1843920896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044882828282828284,
      "loss": 2.6948,
      "theoretical_loss": 3.453886376212015,
      "tokens_seen": 1843986432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044880808080808085,
      "loss": 2.6051,
      "theoretical_loss": 3.4538759774197265,
      "tokens_seen": 1844051968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487878787878788,
      "loss": 2.7497,
      "theoretical_loss": 3.4538655791004693,
      "tokens_seen": 1844117504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044876767676767677,
      "loss": 2.9335,
      "theoretical_loss": 3.4538551812542053,
      "tokens_seen": 1844183040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487474747474747,
      "loss": 2.4906,
      "theoretical_loss": 3.4538447838808954,
      "tokens_seen": 1844248576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487272727272728,
      "loss": 2.6569,
      "theoretical_loss": 3.453834386980502,
      "tokens_seen": 1844314112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044870707070707074,
      "loss": 2.7522,
      "theoretical_loss": 3.453823990552986,
      "tokens_seen": 1844379648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486868686868687,
      "loss": 2.2554,
      "theoretical_loss": 3.4538135945983104,
      "tokens_seen": 1844445184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044866666666666666,
      "loss": 2.5178,
      "theoretical_loss": 3.453803199116435,
      "tokens_seen": 1844510720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486464646464646,
      "loss": 2.5563,
      "theoretical_loss": 3.453792804107324,
      "tokens_seen": 1844576256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486262626262627,
      "loss": 2.7275,
      "theoretical_loss": 3.4537824095709366,
      "tokens_seen": 1844641792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044860606060606064,
      "loss": 2.2752,
      "theoretical_loss": 3.4537720155072362,
      "tokens_seen": 1844707328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485858585858586,
      "loss": 2.7008,
      "theoretical_loss": 3.4537616219161835,
      "tokens_seen": 1844772864
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044856565656565655,
      "loss": 2.7969,
      "theoretical_loss": 3.4537512287977403,
      "tokens_seen": 1844838400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485454545454545,
      "loss": 2.7409,
      "theoretical_loss": 3.4537408361518693,
      "tokens_seen": 1844903936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485252525252526,
      "loss": 2.4061,
      "theoretical_loss": 3.4537304439785315,
      "tokens_seen": 1844969472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044850505050505053,
      "loss": 2.663,
      "theoretical_loss": 3.4537200522776885,
      "tokens_seen": 1845035008
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1017203,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1169590950012207,
      "objective/train/theoretical_loss": 3.4537096610493023,
      "objective/train/tokens_used": 203959776,
      "theoretical_loss": 3.4537096610493023,
      "tokens_seen": 1845100544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484848484848485,
      "loss": 2.6011,
      "theoretical_loss": 3.4537096610493023,
      "tokens_seen": 1845100544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044846464646464644,
      "loss": 2.5262,
      "theoretical_loss": 3.4536992702933347,
      "tokens_seen": 1845166080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044844444444444445,
      "loss": 2.5678,
      "theoretical_loss": 3.453688880009747,
      "tokens_seen": 1845231616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044842424242424247,
      "loss": 2.5909,
      "theoretical_loss": 3.4536784901985014,
      "tokens_seen": 1845297152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484040404040404,
      "loss": 2.6497,
      "theoretical_loss": 3.4536681008595593,
      "tokens_seen": 1845362688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483838383838384,
      "loss": 2.7736,
      "theoretical_loss": 3.4536577119928835,
      "tokens_seen": 1845428224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044836363636363634,
      "loss": 2.6749,
      "theoretical_loss": 3.4536473235984344,
      "tokens_seen": 1845493760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044834343434343435,
      "loss": 2.6772,
      "theoretical_loss": 3.4536369356761742,
      "tokens_seen": 1845559296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044832323232323236,
      "loss": 2.6142,
      "theoretical_loss": 3.4536265482260653,
      "tokens_seen": 1845624832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483030303030303,
      "loss": 2.6519,
      "theoretical_loss": 3.4536161612480685,
      "tokens_seen": 1845690368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044828282828282827,
      "loss": 2.735,
      "theoretical_loss": 3.453605774742146,
      "tokens_seen": 1845755904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482626262626263,
      "loss": 2.825,
      "theoretical_loss": 3.4535953887082598,
      "tokens_seen": 1845821440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044824242424242424,
      "loss": 2.8881,
      "theoretical_loss": 3.4535850031463715,
      "tokens_seen": 1845886976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044822222222222225,
      "loss": 2.6401,
      "theoretical_loss": 3.453574618056443,
      "tokens_seen": 1845952512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482020202020202,
      "loss": 2.6842,
      "theoretical_loss": 3.4535642334384358,
      "tokens_seen": 1846018048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044818181818181816,
      "loss": 2.7116,
      "theoretical_loss": 3.4535538492923123,
      "tokens_seen": 1846083584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481616161616162,
      "loss": 2.5188,
      "theoretical_loss": 3.4535434656180337,
      "tokens_seen": 1846149120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044814141414141413,
      "loss": 2.5895,
      "theoretical_loss": 3.4535330824155617,
      "tokens_seen": 1846214656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044812121212121214,
      "loss": 2.5826,
      "theoretical_loss": 3.453522699684859,
      "tokens_seen": 1846280192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481010101010101,
      "loss": 2.6438,
      "theoretical_loss": 3.4535123174258864,
      "tokens_seen": 1846345728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480808080808081,
      "loss": 2.6586,
      "theoretical_loss": 3.4535019356386067,
      "tokens_seen": 1846411264
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044806060606060607,
      "loss": 2.6629,
      "theoretical_loss": 3.4534915543229805,
      "tokens_seen": 1846476800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448040404040404,
      "loss": 2.7671,
      "theoretical_loss": 3.4534811734789708,
      "tokens_seen": 1846542336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044802020202020204,
      "loss": 2.7187,
      "theoretical_loss": 3.453470793106539,
      "tokens_seen": 1846607872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448,
      "loss": 2.6632,
      "theoretical_loss": 3.453460413205647,
      "tokens_seen": 1846673408
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1017618,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3263816833496094,
      "objective/train/theoretical_loss": 3.453450033776256,
      "objective/train/tokens_used": 205598176,
      "theoretical_loss": 3.453450033776256,
      "tokens_seen": 1846738944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447979797979798,
      "loss": 2.5091,
      "theoretical_loss": 3.453450033776256,
      "tokens_seen": 1846738944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044795959595959596,
      "loss": 2.5978,
      "theoretical_loss": 3.453439654818329,
      "tokens_seen": 1846804480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004479393939393939,
      "loss": 2.7581,
      "theoretical_loss": 3.4534292763318275,
      "tokens_seen": 1846870016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044791919191919193,
      "loss": 2.7108,
      "theoretical_loss": 3.4534188983167127,
      "tokens_seen": 1846935552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044789898989898994,
      "loss": 2.6206,
      "theoretical_loss": 3.453408520772947,
      "tokens_seen": 1847001088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478787878787879,
      "loss": 2.4462,
      "theoretical_loss": 3.453398143700492,
      "tokens_seen": 1847066624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044785858585858585,
      "loss": 2.6832,
      "theoretical_loss": 3.4533877670993096,
      "tokens_seen": 1847132160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478383838383838,
      "loss": 2.5257,
      "theoretical_loss": 3.4533773909693624,
      "tokens_seen": 1847197696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478181818181818,
      "loss": 2.721,
      "theoretical_loss": 3.4533670153106115,
      "tokens_seen": 1847263232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044779797979797983,
      "loss": 2.7503,
      "theoretical_loss": 3.453356640123019,
      "tokens_seen": 1847328768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477777777777778,
      "loss": 2.8506,
      "theoretical_loss": 3.4533462654065463,
      "tokens_seen": 1847394304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044775757575757575,
      "loss": 2.7111,
      "theoretical_loss": 3.4533358911611565,
      "tokens_seen": 1847459840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477373737373737,
      "loss": 2.7537,
      "theoretical_loss": 3.4533255173868103,
      "tokens_seen": 1847525376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044771717171717177,
      "loss": 2.7598,
      "theoretical_loss": 3.4533151440834704,
      "tokens_seen": 1847590912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476969696969697,
      "loss": 2.5996,
      "theoretical_loss": 3.4533047712510982,
      "tokens_seen": 1847656448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476767676767677,
      "loss": 2.5975,
      "theoretical_loss": 3.453294398889656,
      "tokens_seen": 1847721984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044765656565656564,
      "loss": 2.8062,
      "theoretical_loss": 3.453284026999105,
      "tokens_seen": 1847787520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476363636363636,
      "loss": 2.7529,
      "theoretical_loss": 3.4532736555794084,
      "tokens_seen": 1847853056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044761616161616166,
      "loss": 2.457,
      "theoretical_loss": 3.453263284630527,
      "tokens_seen": 1847918592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475959595959596,
      "loss": 2.3476,
      "theoretical_loss": 3.453252914152423,
      "tokens_seen": 1847984128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475757575757576,
      "loss": 2.5465,
      "theoretical_loss": 3.4532425441450587,
      "tokens_seen": 1848049664
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044755555555555553,
      "loss": 2.7043,
      "theoretical_loss": 3.4532321746083956,
      "tokens_seen": 1848115200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044753535353535354,
      "loss": 2.527,
      "theoretical_loss": 3.4532218055423956,
      "tokens_seen": 1848180736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044751515151515155,
      "loss": 2.6198,
      "theoretical_loss": 3.4532114369470213,
      "tokens_seen": 1848246272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474949494949495,
      "loss": 2.7901,
      "theoretical_loss": 3.453201068822234,
      "tokens_seen": 1848311808
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1018803,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.287393093109131,
      "objective/train/theoretical_loss": 3.453190701167996,
      "objective/train/tokens_used": 207236576,
      "theoretical_loss": 3.453190701167996,
      "tokens_seen": 1848377344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044747474747474747,
      "loss": 2.4596,
      "theoretical_loss": 3.453190701167996,
      "tokens_seen": 1848377344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474545454545454,
      "loss": 2.6234,
      "theoretical_loss": 3.4531803339842693,
      "tokens_seen": 1848442880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044743434343434344,
      "loss": 2.9293,
      "theoretical_loss": 3.453169967271015,
      "tokens_seen": 1848508416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044741414141414145,
      "loss": 2.6174,
      "theoretical_loss": 3.4531596010281964,
      "tokens_seen": 1848573952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473939393939394,
      "loss": 2.5937,
      "theoretical_loss": 3.4531492352557747,
      "tokens_seen": 1848639488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044737373737373736,
      "loss": 2.6619,
      "theoretical_loss": 3.453138869953712,
      "tokens_seen": 1848705024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044735353535353537,
      "loss": 2.7539,
      "theoretical_loss": 3.4531285051219704,
      "tokens_seen": 1848770560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473333333333334,
      "loss": 2.8116,
      "theoretical_loss": 3.4531181407605116,
      "tokens_seen": 1848836096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044731313131313134,
      "loss": 2.7512,
      "theoretical_loss": 3.453107776869298,
      "tokens_seen": 1848901632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472929292929293,
      "loss": 2.8565,
      "theoretical_loss": 3.4530974134482917,
      "tokens_seen": 1848967168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044727272727272725,
      "loss": 2.6797,
      "theoretical_loss": 3.4530870504974542,
      "tokens_seen": 1849032704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044725252525252526,
      "loss": 2.5755,
      "theoretical_loss": 3.4530766880167474,
      "tokens_seen": 1849098240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472323232323233,
      "loss": 2.5315,
      "theoretical_loss": 3.453066326006134,
      "tokens_seen": 1849163776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044721212121212123,
      "loss": 2.7439,
      "theoretical_loss": 3.4530559644655754,
      "tokens_seen": 1849229312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471919191919192,
      "loss": 2.7021,
      "theoretical_loss": 3.453045603395034,
      "tokens_seen": 1849294848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471717171717172,
      "loss": 2.4656,
      "theoretical_loss": 3.4530352427944715,
      "tokens_seen": 1849360384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044715151515151516,
      "loss": 2.5479,
      "theoretical_loss": 3.45302488266385,
      "tokens_seen": 1849425920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044713131313131317,
      "loss": 2.778,
      "theoretical_loss": 3.453014523003132,
      "tokens_seen": 1849491456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471111111111111,
      "loss": 2.9284,
      "theoretical_loss": 3.453004163812279,
      "tokens_seen": 1849556992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470909090909091,
      "loss": 2.6908,
      "theoretical_loss": 3.452993805091253,
      "tokens_seen": 1849622528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470707070707071,
      "loss": 2.711,
      "theoretical_loss": 3.4529834468400162,
      "tokens_seen": 1849688064
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044705050505050505,
      "loss": 2.7749,
      "theoretical_loss": 3.4529730890585313,
      "tokens_seen": 1849753600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044703030303030306,
      "loss": 2.4908,
      "theoretical_loss": 3.4529627317467595,
      "tokens_seen": 1849819136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447010101010101,
      "loss": 2.5087,
      "theoretical_loss": 3.4529523749046627,
      "tokens_seen": 1849884672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044698989898989903,
      "loss": 2.6053,
      "theoretical_loss": 3.4529420185322035,
      "tokens_seen": 1849950208
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 1019318,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6232709884643555,
      "objective/train/theoretical_loss": 3.4529316626293443,
      "objective/train/tokens_used": 208874976,
      "theoretical_loss": 3.4529316626293443,
      "tokens_seen": 1850015744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446969696969697,
      "loss": 2.5128,
      "theoretical_loss": 3.4529316626293443,
      "tokens_seen": 1850015744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044694949494949494,
      "loss": 2.6452,
      "theoretical_loss": 3.4529213071960463,
      "tokens_seen": 1850081280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044692929292929295,
      "loss": 2.74,
      "theoretical_loss": 3.4529109522322727,
      "tokens_seen": 1850146816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469090909090909,
      "loss": 2.6177,
      "theoretical_loss": 3.452900597737984,
      "tokens_seen": 1850212352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468888888888889,
      "loss": 2.609,
      "theoretical_loss": 3.452890243713144,
      "tokens_seen": 1850277888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468686868686869,
      "loss": 2.3597,
      "theoretical_loss": 3.452879890157713,
      "tokens_seen": 1850343424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044684848484848484,
      "loss": 2.6272,
      "theoretical_loss": 3.452869537071655,
      "tokens_seen": 1850408960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044682828282828285,
      "loss": 2.591,
      "theoretical_loss": 3.4528591844549306,
      "tokens_seen": 1850474496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468080808080808,
      "loss": 2.7254,
      "theoretical_loss": 3.4528488323075024,
      "tokens_seen": 1850540032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467878787878788,
      "loss": 2.5987,
      "theoretical_loss": 3.4528384806293326,
      "tokens_seen": 1850605568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044676767676767677,
      "loss": 2.7431,
      "theoretical_loss": 3.452828129420384,
      "tokens_seen": 1850671104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044674747474747473,
      "loss": 2.4767,
      "theoretical_loss": 3.452817778680617,
      "tokens_seen": 1850736640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044672727272727274,
      "loss": 2.5841,
      "theoretical_loss": 3.4528074284099954,
      "tokens_seen": 1850802176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044670707070707075,
      "loss": 2.5436,
      "theoretical_loss": 3.45279707860848,
      "tokens_seen": 1850867712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466868686868687,
      "loss": 2.7338,
      "theoretical_loss": 3.452786729276034,
      "tokens_seen": 1850933248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044666666666666666,
      "loss": 2.629,
      "theoretical_loss": 3.452776380412619,
      "tokens_seen": 1850998784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466464646464646,
      "loss": 2.5704,
      "theoretical_loss": 3.4527660320181974,
      "tokens_seen": 1851064320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044662626262626263,
      "loss": 2.6984,
      "theoretical_loss": 3.4527556840927307,
      "tokens_seen": 1851129856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044660606060606064,
      "loss": 2.5546,
      "theoretical_loss": 3.452745336636182,
      "tokens_seen": 1851195392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465858585858586,
      "loss": 2.7072,
      "theoretical_loss": 3.452734989648513,
      "tokens_seen": 1851260928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044656565656565656,
      "loss": 2.5535,
      "theoretical_loss": 3.4527246431296854,
      "tokens_seen": 1851326464
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465454545454545,
      "loss": 2.678,
      "theoretical_loss": 3.4527142970796625,
      "tokens_seen": 1851392000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465252525252526,
      "loss": 2.683,
      "theoretical_loss": 3.452703951498405,
      "tokens_seen": 1851457536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044650505050505054,
      "loss": 2.7782,
      "theoretical_loss": 3.452693606385876,
      "tokens_seen": 1851523072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464848484848485,
      "loss": 2.79,
      "theoretical_loss": 3.452683261742038,
      "tokens_seen": 1851588608
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1019593,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.985586643218994,
      "objective/train/theoretical_loss": 3.4526729175668525,
      "objective/train/tokens_used": 210513376,
      "theoretical_loss": 3.4526729175668525,
      "tokens_seen": 1851654144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044646464646464645,
      "loss": 2.7467,
      "theoretical_loss": 3.4526729175668525,
      "tokens_seen": 1851654144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464444444444444,
      "loss": 2.6811,
      "theoretical_loss": 3.452662573860281,
      "tokens_seen": 1851719680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044642424242424247,
      "loss": 2.4987,
      "theoretical_loss": 3.452652230622288,
      "tokens_seen": 1851785216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044640404040404043,
      "loss": 2.6987,
      "theoretical_loss": 3.4526418878528333,
      "tokens_seen": 1851850752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463838383838384,
      "loss": 2.4518,
      "theoretical_loss": 3.4526315455518803,
      "tokens_seen": 1851916288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044636363636363634,
      "loss": 2.7949,
      "theoretical_loss": 3.452621203719391,
      "tokens_seen": 1851981824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044634343434343435,
      "loss": 2.6532,
      "theoretical_loss": 3.452610862355327,
      "tokens_seen": 1852047360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044632323232323236,
      "loss": 2.8367,
      "theoretical_loss": 3.452600521459652,
      "tokens_seen": 1852112896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463030303030303,
      "loss": 2.6774,
      "theoretical_loss": 3.4525901810323263,
      "tokens_seen": 1852178432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462828282828283,
      "loss": 2.6519,
      "theoretical_loss": 3.4525798410733133,
      "tokens_seen": 1852243968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044626262626262623,
      "loss": 2.5936,
      "theoretical_loss": 3.4525695015825755,
      "tokens_seen": 1852309504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044624242424242425,
      "loss": 2.6858,
      "theoretical_loss": 3.4525591625600742,
      "tokens_seen": 1852375040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044622222222222226,
      "loss": 2.692,
      "theoretical_loss": 3.4525488240057722,
      "tokens_seen": 1852440576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462020202020202,
      "loss": 2.6024,
      "theoretical_loss": 3.4525384859196313,
      "tokens_seen": 1852506112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044618181818181817,
      "loss": 2.6869,
      "theoretical_loss": 3.4525281483016146,
      "tokens_seen": 1852571648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461616161616162,
      "loss": 2.7454,
      "theoretical_loss": 3.4525178111516834,
      "tokens_seen": 1852637184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044614141414141414,
      "loss": 2.6114,
      "theoretical_loss": 3.4525074744698,
      "tokens_seen": 1852702720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044612121212121215,
      "loss": 2.6078,
      "theoretical_loss": 3.4524971382559277,
      "tokens_seen": 1852768256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461010101010101,
      "loss": 2.7454,
      "theoretical_loss": 3.4524868025100277,
      "tokens_seen": 1852833792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044608080808080806,
      "loss": 2.6441,
      "theoretical_loss": 3.4524764672320627,
      "tokens_seen": 1852899328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460606060606061,
      "loss": 2.5881,
      "theoretical_loss": 3.4524661324219945,
      "tokens_seen": 1852964864
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044604040404040403,
      "loss": 2.6752,
      "theoretical_loss": 3.452455798079786,
      "tokens_seen": 1853030400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044602020202020204,
      "loss": 2.5053,
      "theoretical_loss": 3.4524454642053994,
      "tokens_seen": 1853095936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446,
      "loss": 2.603,
      "theoretical_loss": 3.4524351307987966,
      "tokens_seen": 1853161472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445979797979798,
      "loss": 2.8048,
      "theoretical_loss": 3.45242479785994,
      "tokens_seen": 1853227008
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1020707,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.561333656311035,
      "objective/train/theoretical_loss": 3.452414465388792,
      "objective/train/tokens_used": 212151776,
      "theoretical_loss": 3.452414465388792,
      "tokens_seen": 1853292544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044595959595959597,
      "loss": 2.6155,
      "theoretical_loss": 3.452414465388792,
      "tokens_seen": 1853292544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459393939393939,
      "loss": 2.8532,
      "theoretical_loss": 3.452404133385315,
      "tokens_seen": 1853358080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044591919191919193,
      "loss": 2.5471,
      "theoretical_loss": 3.4523938018494706,
      "tokens_seen": 1853423616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458989898989899,
      "loss": 2.6863,
      "theoretical_loss": 3.452383470781222,
      "tokens_seen": 1853489152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458787878787879,
      "loss": 2.5466,
      "theoretical_loss": 3.4523731401805313,
      "tokens_seen": 1853554688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044585858585858586,
      "loss": 2.7701,
      "theoretical_loss": 3.4523628100473607,
      "tokens_seen": 1853620224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458383838383838,
      "loss": 2.6979,
      "theoretical_loss": 3.4523524803816716,
      "tokens_seen": 1853685760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044581818181818183,
      "loss": 2.8536,
      "theoretical_loss": 3.4523421511834282,
      "tokens_seen": 1853751296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044579797979797984,
      "loss": 2.5863,
      "theoretical_loss": 3.4523318224525914,
      "tokens_seen": 1853816832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457777777777778,
      "loss": 2.8932,
      "theoretical_loss": 3.452321494189124,
      "tokens_seen": 1853882368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044575757575757575,
      "loss": 2.893,
      "theoretical_loss": 3.452311166392988,
      "tokens_seen": 1853947904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457373737373737,
      "loss": 2.5877,
      "theoretical_loss": 3.4523008390641463,
      "tokens_seen": 1854013440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457171717171717,
      "loss": 2.5738,
      "theoretical_loss": 3.452290512202561,
      "tokens_seen": 1854078976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044569696969696973,
      "loss": 2.6439,
      "theoretical_loss": 3.4522801858081946,
      "tokens_seen": 1854144512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456767676767677,
      "loss": 2.6565,
      "theoretical_loss": 3.4522698598810084,
      "tokens_seen": 1854210048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044565656565656564,
      "loss": 2.4392,
      "theoretical_loss": 3.452259534420967,
      "tokens_seen": 1854275584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044563636363636366,
      "loss": 2.4127,
      "theoretical_loss": 3.4522492094280297,
      "tokens_seen": 1854341120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044561616161616167,
      "loss": 2.7188,
      "theoretical_loss": 3.4522388849021617,
      "tokens_seen": 1854406656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455959595959596,
      "loss": 2.5797,
      "theoretical_loss": 3.4522285608433236,
      "tokens_seen": 1854472192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455757575757576,
      "loss": 2.8985,
      "theoretical_loss": 3.4522182372514783,
      "tokens_seen": 1854537728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044555555555555554,
      "loss": 2.5284,
      "theoretical_loss": 3.4522079141265882,
      "tokens_seen": 1854603264
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044553535353535355,
      "loss": 2.6146,
      "theoretical_loss": 3.4521975914686163,
      "tokens_seen": 1854668800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044551515151515156,
      "loss": 2.6305,
      "theoretical_loss": 3.4521872692775237,
      "tokens_seen": 1854734336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454949494949495,
      "loss": 2.566,
      "theoretical_loss": 3.4521769475532738,
      "tokens_seen": 1854799872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454747474747475,
      "loss": 2.7986,
      "theoretical_loss": 3.4521666262958286,
      "tokens_seen": 1854865408
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1021895,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6850180625915527,
      "objective/train/theoretical_loss": 3.4521563055051505,
      "objective/train/tokens_used": 213790176,
      "theoretical_loss": 3.4521563055051505,
      "tokens_seen": 1854930944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044545454545454543,
      "loss": 2.5554,
      "theoretical_loss": 3.4521563055051505,
      "tokens_seen": 1854930944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454343434343435,
      "loss": 2.6896,
      "theoretical_loss": 3.452145985181202,
      "tokens_seen": 1854996480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044541414141414145,
      "loss": 2.8255,
      "theoretical_loss": 3.4521356653239454,
      "tokens_seen": 1855062016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453939393939394,
      "loss": 2.5766,
      "theoretical_loss": 3.452125345933343,
      "tokens_seen": 1855127552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044537373737373737,
      "loss": 2.8282,
      "theoretical_loss": 3.452115027009358,
      "tokens_seen": 1855193088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453535353535353,
      "loss": 2.7015,
      "theoretical_loss": 3.4521047085519516,
      "tokens_seen": 1855258624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453333333333334,
      "loss": 2.7837,
      "theoretical_loss": 3.4520943905610872,
      "tokens_seen": 1855324160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044531313131313135,
      "loss": 2.5069,
      "theoretical_loss": 3.4520840730367266,
      "tokens_seen": 1855389696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452929292929293,
      "loss": 2.5115,
      "theoretical_loss": 3.452073755978833,
      "tokens_seen": 1855455232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044527272727272726,
      "loss": 2.906,
      "theoretical_loss": 3.4520634393873673,
      "tokens_seen": 1855520768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452525252525252,
      "loss": 2.5688,
      "theoretical_loss": 3.452053123262294,
      "tokens_seen": 1855586304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452323232323233,
      "loss": 2.5582,
      "theoretical_loss": 3.452042807603574,
      "tokens_seen": 1855651840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044521212121212124,
      "loss": 2.632,
      "theoretical_loss": 3.45203249241117,
      "tokens_seen": 1855717376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451919191919192,
      "loss": 2.7484,
      "theoretical_loss": 3.452022177685045,
      "tokens_seen": 1855782912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044517171717171715,
      "loss": 2.7794,
      "theoretical_loss": 3.4520118634251613,
      "tokens_seen": 1855848448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044515151515151516,
      "loss": 2.4757,
      "theoretical_loss": 3.452001549631481,
      "tokens_seen": 1855913984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451313131313132,
      "loss": 2.7872,
      "theoretical_loss": 3.451991236303967,
      "tokens_seen": 1855979520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044511111111111113,
      "loss": 2.7724,
      "theoretical_loss": 3.4519809234425813,
      "tokens_seen": 1856045056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450909090909091,
      "loss": 2.5745,
      "theoretical_loss": 3.4519706110472868,
      "tokens_seen": 1856110592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044507070707070704,
      "loss": 2.6076,
      "theoretical_loss": 3.451960299118046,
      "tokens_seen": 1856176128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044505050505050506,
      "loss": 2.7614,
      "theoretical_loss": 3.451949987654821,
      "tokens_seen": 1856241664
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044503030303030307,
      "loss": 2.6492,
      "theoretical_loss": 3.4519396766575747,
      "tokens_seen": 1856307200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445010101010101,
      "loss": 2.5089,
      "theoretical_loss": 3.451929366126269,
      "tokens_seen": 1856372736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444989898989899,
      "loss": 2.8286,
      "theoretical_loss": 3.4519190560608672,
      "tokens_seen": 1856438272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444969696969697,
      "loss": 2.7507,
      "theoretical_loss": 3.451908746461331,
      "tokens_seen": 1856503808
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1022391,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0687434673309326,
      "objective/train/theoretical_loss": 3.4518984373276234,
      "objective/train/tokens_used": 215428576,
      "theoretical_loss": 3.4518984373276234,
      "tokens_seen": 1856569344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044494949494949495,
      "loss": 2.7867,
      "theoretical_loss": 3.4518984373276234,
      "tokens_seen": 1856569344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044492929292929296,
      "loss": 2.713,
      "theoretical_loss": 3.4518881286597067,
      "tokens_seen": 1856634880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449090909090909,
      "loss": 2.8357,
      "theoretical_loss": 3.4518778204575433,
      "tokens_seen": 1856700416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044488888888888887,
      "loss": 2.7542,
      "theoretical_loss": 3.4518675127210967,
      "tokens_seen": 1856765952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448686868686869,
      "loss": 2.4508,
      "theoretical_loss": 3.4518572054503283,
      "tokens_seen": 1856831488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044484848484848484,
      "loss": 2.7486,
      "theoretical_loss": 3.4518468986452007,
      "tokens_seen": 1856897024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044482828282828285,
      "loss": 2.5418,
      "theoretical_loss": 3.4518365923056766,
      "tokens_seen": 1856962560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448080808080808,
      "loss": 2.4934,
      "theoretical_loss": 3.4518262864317193,
      "tokens_seen": 1857028096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447878787878788,
      "loss": 2.6511,
      "theoretical_loss": 3.45181598102329,
      "tokens_seen": 1857093632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447676767676768,
      "loss": 2.6574,
      "theoretical_loss": 3.4518056760803524,
      "tokens_seen": 1857159168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044474747474747473,
      "loss": 2.5949,
      "theoretical_loss": 3.451795371602868,
      "tokens_seen": 1857224704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044472727272727274,
      "loss": 2.5193,
      "theoretical_loss": 3.4517850675908006,
      "tokens_seen": 1857290240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447070707070707,
      "loss": 2.7867,
      "theoretical_loss": 3.451774764044112,
      "tokens_seen": 1857355776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446868686868687,
      "loss": 2.755,
      "theoretical_loss": 3.4517644609627647,
      "tokens_seen": 1857421312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044466666666666667,
      "loss": 2.6344,
      "theoretical_loss": 3.4517541583467213,
      "tokens_seen": 1857486848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446464646464646,
      "loss": 2.667,
      "theoretical_loss": 3.451743856195945,
      "tokens_seen": 1857552384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044462626262626264,
      "loss": 2.8186,
      "theoretical_loss": 3.4517335545103967,
      "tokens_seen": 1857617920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044460606060606065,
      "loss": 2.3931,
      "theoretical_loss": 3.4517232532900413,
      "tokens_seen": 1857683456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445858585858586,
      "loss": 2.6651,
      "theoretical_loss": 3.45171295253484,
      "tokens_seen": 1857748992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044456565656565656,
      "loss": 2.6855,
      "theoretical_loss": 3.451702652244755,
      "tokens_seen": 1857814528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445454545454545,
      "loss": 2.609,
      "theoretical_loss": 3.45169235241975,
      "tokens_seen": 1857880064
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044452525252525253,
      "loss": 2.716,
      "theoretical_loss": 3.4516820530597867,
      "tokens_seen": 1857945600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044450505050505054,
      "loss": 2.572,
      "theoretical_loss": 3.4516717541648285,
      "tokens_seen": 1858011136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444848484848485,
      "loss": 2.4947,
      "theoretical_loss": 3.4516614557348375,
      "tokens_seen": 1858076672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044446464646464645,
      "loss": 2.7446,
      "theoretical_loss": 3.451651157769777,
      "tokens_seen": 1858142208
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1023648,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7927091121673584,
      "objective/train/theoretical_loss": 3.451640860269608,
      "objective/train/tokens_used": 217066976,
      "theoretical_loss": 3.451640860269608,
      "tokens_seen": 1858207744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444444444444444,
      "loss": 2.4124,
      "theoretical_loss": 3.451640860269608,
      "tokens_seen": 1858207744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444242424242425,
      "loss": 2.5427,
      "theoretical_loss": 3.4516305632342945,
      "tokens_seen": 1858273280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044440404040404043,
      "loss": 2.5259,
      "theoretical_loss": 3.4516202666637987,
      "tokens_seen": 1858338816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443838383838384,
      "loss": 2.6627,
      "theoretical_loss": 3.451609970558083,
      "tokens_seen": 1858404352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044436363636363635,
      "loss": 2.7357,
      "theoretical_loss": 3.4515996749171114,
      "tokens_seen": 1858469888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443434343434343,
      "loss": 2.7667,
      "theoretical_loss": 3.4515893797408443,
      "tokens_seen": 1858535424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044432323232323237,
      "loss": 2.5378,
      "theoretical_loss": 3.4515790850292465,
      "tokens_seen": 1858600960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443030303030303,
      "loss": 2.5758,
      "theoretical_loss": 3.4515687907822787,
      "tokens_seen": 1858666496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442828282828283,
      "loss": 2.536,
      "theoretical_loss": 3.451558496999905,
      "tokens_seen": 1858732032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044426262626262624,
      "loss": 2.6107,
      "theoretical_loss": 3.451548203682088,
      "tokens_seen": 1858797568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044424242424242425,
      "loss": 2.7001,
      "theoretical_loss": 3.451537910828789,
      "tokens_seen": 1858863104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044422222222222226,
      "loss": 2.6327,
      "theoretical_loss": 3.4515276184399717,
      "tokens_seen": 1858928640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442020202020202,
      "loss": 2.5813,
      "theoretical_loss": 3.451517326515599,
      "tokens_seen": 1858994176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441818181818182,
      "loss": 2.7582,
      "theoretical_loss": 3.4515070350556334,
      "tokens_seen": 1859059712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044416161616161613,
      "loss": 2.8071,
      "theoretical_loss": 3.4514967440600364,
      "tokens_seen": 1859125248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044414141414141414,
      "loss": 2.7393,
      "theoretical_loss": 3.451486453528773,
      "tokens_seen": 1859190784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044412121212121215,
      "loss": 2.5984,
      "theoretical_loss": 3.4514761634618036,
      "tokens_seen": 1859256320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441010101010101,
      "loss": 2.734,
      "theoretical_loss": 3.4514658738590915,
      "tokens_seen": 1859321856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044408080808080807,
      "loss": 2.7378,
      "theoretical_loss": 3.4514555847206,
      "tokens_seen": 1859387392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440606060606061,
      "loss": 2.7765,
      "theoretical_loss": 3.451445296046292,
      "tokens_seen": 1859452928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044404040404040404,
      "loss": 2.5779,
      "theoretical_loss": 3.4514350078361296,
      "tokens_seen": 1859518464
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044402020202020205,
      "loss": 2.4927,
      "theoretical_loss": 3.4514247200900754,
      "tokens_seen": 1859584000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444,
      "loss": 2.6562,
      "theoretical_loss": 3.4514144328080922,
      "tokens_seen": 1859649536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044397979797979796,
      "loss": 2.9958,
      "theoretical_loss": 3.4514041459901432,
      "tokens_seen": 1859715072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044395959595959597,
      "loss": 2.7743,
      "theoretical_loss": 3.4513938596361906,
      "tokens_seen": 1859780608
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1024242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3011066913604736,
      "objective/train/theoretical_loss": 3.451383573746197,
      "objective/train/tokens_used": 218705376,
      "theoretical_loss": 3.451383573746197,
      "tokens_seen": 1859846144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443939393939394,
      "loss": 2.4715,
      "theoretical_loss": 3.451383573746197,
      "tokens_seen": 1859846144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044391919191919194,
      "loss": 2.6938,
      "theoretical_loss": 3.4513732883201254,
      "tokens_seen": 1859911680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438989898989899,
      "loss": 2.6091,
      "theoretical_loss": 3.451363003357939,
      "tokens_seen": 1859977216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438787878787879,
      "loss": 2.6714,
      "theoretical_loss": 3.4513527188596,
      "tokens_seen": 1860042752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044385858585858587,
      "loss": 2.6668,
      "theoretical_loss": 3.4513424348250705,
      "tokens_seen": 1860108288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438383838383839,
      "loss": 2.8937,
      "theoretical_loss": 3.4513321512543147,
      "tokens_seen": 1860173824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044381818181818183,
      "loss": 2.8199,
      "theoretical_loss": 3.451321868147294,
      "tokens_seen": 1860239360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437979797979798,
      "loss": 2.6694,
      "theoretical_loss": 3.451311585503972,
      "tokens_seen": 1860304896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437777777777778,
      "loss": 2.7428,
      "theoretical_loss": 3.4513013033243114,
      "tokens_seen": 1860370432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044375757575757576,
      "loss": 2.5192,
      "theoretical_loss": 3.4512910216082746,
      "tokens_seen": 1860435968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044373737373737377,
      "loss": 2.5591,
      "theoretical_loss": 3.4512807403558243,
      "tokens_seen": 1860501504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437171717171717,
      "loss": 2.5545,
      "theoretical_loss": 3.451270459566924,
      "tokens_seen": 1860567040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044369696969696974,
      "loss": 2.7749,
      "theoretical_loss": 3.451260179241536,
      "tokens_seen": 1860632576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436767676767677,
      "loss": 2.2815,
      "theoretical_loss": 3.451249899379622,
      "tokens_seen": 1860698112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044365656565656565,
      "loss": 2.8408,
      "theoretical_loss": 3.4512396199811466,
      "tokens_seen": 1860763648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044363636363636366,
      "loss": 2.5795,
      "theoretical_loss": 3.451229341046072,
      "tokens_seen": 1860829184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436161616161616,
      "loss": 2.7172,
      "theoretical_loss": 3.4512190625743604,
      "tokens_seen": 1860894720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044359595959595963,
      "loss": 2.7811,
      "theoretical_loss": 3.451208784565975,
      "tokens_seen": 1860960256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435757575757576,
      "loss": 2.5311,
      "theoretical_loss": 3.4511985070208793,
      "tokens_seen": 1861025792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044355555555555554,
      "loss": 2.4705,
      "theoretical_loss": 3.4511882299390346,
      "tokens_seen": 1861091328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044353535353535355,
      "loss": 2.8382,
      "theoretical_loss": 3.4511779533204043,
      "tokens_seen": 1861156864
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435151515151515,
      "loss": 2.5743,
      "theoretical_loss": 3.4511676771649524,
      "tokens_seen": 1861222400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434949494949495,
      "loss": 2.7183,
      "theoretical_loss": 3.4511574014726403,
      "tokens_seen": 1861287936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434747474747475,
      "loss": 2.7241,
      "theoretical_loss": 3.4511471262434306,
      "tokens_seen": 1861353472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044345454545454544,
      "loss": 2.7197,
      "theoretical_loss": 3.4511368514772878,
      "tokens_seen": 1861419008
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1025485,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8361258506774902,
      "objective/train/theoretical_loss": 3.451126577174173,
      "objective/train/tokens_used": 220343776,
      "theoretical_loss": 3.451126577174173,
      "tokens_seen": 1861484544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044343434343434345,
      "loss": 2.6971,
      "theoretical_loss": 3.451126577174173,
      "tokens_seen": 1861484544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044341414141414146,
      "loss": 2.5431,
      "theoretical_loss": 3.4511163033340497,
      "tokens_seen": 1861550080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004433939393939394,
      "loss": 2.3856,
      "theoretical_loss": 3.451106029956881,
      "tokens_seen": 1861615616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044337373737373737,
      "loss": 2.4283,
      "theoretical_loss": 3.45109575704263,
      "tokens_seen": 1861681152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044335353535353533,
      "loss": 2.6594,
      "theoretical_loss": 3.451085484591258,
      "tokens_seen": 1861746688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044333333333333334,
      "loss": 2.663,
      "theoretical_loss": 3.4510752126027295,
      "tokens_seen": 1861812224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044331313131313135,
      "loss": 2.6829,
      "theoretical_loss": 3.451064941077007,
      "tokens_seen": 1861877760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432929292929293,
      "loss": 2.6757,
      "theoretical_loss": 3.451054670014053,
      "tokens_seen": 1861943296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044327272727272726,
      "loss": 2.665,
      "theoretical_loss": 3.4510443994138305,
      "tokens_seen": 1862008832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432525252525252,
      "loss": 2.6872,
      "theoretical_loss": 3.451034129276302,
      "tokens_seen": 1862074368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432323232323233,
      "loss": 2.6956,
      "theoretical_loss": 3.451023859601431,
      "tokens_seen": 1862139904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044321212121212124,
      "loss": 2.6309,
      "theoretical_loss": 3.45101359038918,
      "tokens_seen": 1862205440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431919191919192,
      "loss": 2.7008,
      "theoretical_loss": 3.4510033216395124,
      "tokens_seen": 1862270976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044317171717171716,
      "loss": 2.6083,
      "theoretical_loss": 3.4509930533523905,
      "tokens_seen": 1862336512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431515151515151,
      "loss": 2.6467,
      "theoretical_loss": 3.450982785527777,
      "tokens_seen": 1862402048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431313131313132,
      "loss": 2.9646,
      "theoretical_loss": 3.4509725181656354,
      "tokens_seen": 1862467584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044311111111111114,
      "loss": 2.6964,
      "theoretical_loss": 3.4509622512659286,
      "tokens_seen": 1862533120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004430909090909091,
      "loss": 2.5257,
      "theoretical_loss": 3.4509519848286185,
      "tokens_seen": 1862598656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044307070707070705,
      "loss": 2.5954,
      "theoretical_loss": 3.4509417188536693,
      "tokens_seen": 1862664192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044305050505050506,
      "loss": 2.6717,
      "theoretical_loss": 3.4509314533410436,
      "tokens_seen": 1862729728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044303030303030307,
      "loss": 2.5898,
      "theoretical_loss": 3.450921188290704,
      "tokens_seen": 1862795264
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044301010101010103,
      "loss": 2.6666,
      "theoretical_loss": 3.450910923702613,
      "tokens_seen": 1862860800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442989898989899,
      "loss": 2.7131,
      "theoretical_loss": 3.4509006595767344,
      "tokens_seen": 1862926336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044296969696969694,
      "loss": 2.4907,
      "theoretical_loss": 3.4508903959130306,
      "tokens_seen": 1862991872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044294949494949495,
      "loss": 2.5954,
      "theoretical_loss": 3.4508801327114647,
      "tokens_seen": 1863057408
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1026014,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.185640573501587,
      "objective/train/theoretical_loss": 3.450869869972,
      "objective/train/tokens_used": 221982176,
      "theoretical_loss": 3.450869869972,
      "tokens_seen": 1863122944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044292929292929296,
      "loss": 2.6933,
      "theoretical_loss": 3.450869869972,
      "tokens_seen": 1863122944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429090909090909,
      "loss": 2.906,
      "theoretical_loss": 3.450859607694599,
      "tokens_seen": 1863188480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428888888888889,
      "loss": 2.5886,
      "theoretical_loss": 3.4508493458792246,
      "tokens_seen": 1863254016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428686868686869,
      "loss": 2.6838,
      "theoretical_loss": 3.4508390845258394,
      "tokens_seen": 1863319552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044284848484848485,
      "loss": 2.7344,
      "theoretical_loss": 3.4508288236344073,
      "tokens_seen": 1863385088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044282828282828286,
      "loss": 2.7642,
      "theoretical_loss": 3.4508185632048907,
      "tokens_seen": 1863450624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428080808080808,
      "loss": 2.5856,
      "theoretical_loss": 3.4508083032372525,
      "tokens_seen": 1863516160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044278787878787877,
      "loss": 2.4,
      "theoretical_loss": 3.4507980437314556,
      "tokens_seen": 1863581696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427676767676768,
      "loss": 2.8011,
      "theoretical_loss": 3.4507877846874635,
      "tokens_seen": 1863647232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044274747474747474,
      "loss": 2.6666,
      "theoretical_loss": 3.450777526105239,
      "tokens_seen": 1863712768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044272727272727275,
      "loss": 2.6065,
      "theoretical_loss": 3.4507672679847445,
      "tokens_seen": 1863778304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427070707070707,
      "loss": 2.651,
      "theoretical_loss": 3.450757010325944,
      "tokens_seen": 1863843840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426868686868687,
      "loss": 2.8708,
      "theoretical_loss": 3.450746753128799,
      "tokens_seen": 1863909376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426666666666667,
      "loss": 2.6971,
      "theoretical_loss": 3.450736496393274,
      "tokens_seen": 1863974912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044264646464646463,
      "loss": 2.8966,
      "theoretical_loss": 3.4507262401193315,
      "tokens_seen": 1864040448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044262626262626264,
      "loss": 2.7598,
      "theoretical_loss": 3.4507159843069344,
      "tokens_seen": 1864105984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426060606060606,
      "loss": 2.6357,
      "theoretical_loss": 3.450705728956045,
      "tokens_seen": 1864171520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425858585858586,
      "loss": 2.6469,
      "theoretical_loss": 3.4506954740666274,
      "tokens_seen": 1864237056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044256565656565657,
      "loss": 2.7603,
      "theoretical_loss": 3.450685219638644,
      "tokens_seen": 1864302592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425454545454545,
      "loss": 2.5525,
      "theoretical_loss": 3.450674965672059,
      "tokens_seen": 1864368128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044252525252525254,
      "loss": 2.4534,
      "theoretical_loss": 3.4506647121668332,
      "tokens_seen": 1864433664
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044250505050505055,
      "loss": 2.7848,
      "theoretical_loss": 3.450654459122932,
      "tokens_seen": 1864499200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424848484848485,
      "loss": 2.7757,
      "theoretical_loss": 3.450644206540316,
      "tokens_seen": 1864564736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044246464646464646,
      "loss": 2.3675,
      "theoretical_loss": 3.4506339544189504,
      "tokens_seen": 1864630272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424444444444444,
      "loss": 2.9488,
      "theoretical_loss": 3.450623702758797,
      "tokens_seen": 1864695808
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1027424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.689354419708252,
      "objective/train/theoretical_loss": 3.4506134515598195,
      "objective/train/tokens_used": 223620576,
      "theoretical_loss": 3.4506134515598195,
      "tokens_seen": 1864761344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044242424242424243,
      "loss": 2.8144,
      "theoretical_loss": 3.4506134515598195,
      "tokens_seen": 1864761344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044240404040404044,
      "loss": 2.7622,
      "theoretical_loss": 3.450603200821981,
      "tokens_seen": 1864826880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423838383838384,
      "loss": 2.4991,
      "theoretical_loss": 3.4505929505452437,
      "tokens_seen": 1864892416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044236363636363635,
      "loss": 2.7188,
      "theoretical_loss": 3.450582700729571,
      "tokens_seen": 1864957952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423434343434343,
      "loss": 2.4162,
      "theoretical_loss": 3.4505724513749265,
      "tokens_seen": 1865023488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423232323232324,
      "loss": 2.6481,
      "theoretical_loss": 3.4505622024812728,
      "tokens_seen": 1865089024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044230303030303033,
      "loss": 2.8747,
      "theoretical_loss": 3.450551954048573,
      "tokens_seen": 1865154560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422828282828283,
      "loss": 2.652,
      "theoretical_loss": 3.4505417060767902,
      "tokens_seen": 1865220096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044226262626262625,
      "loss": 2.5749,
      "theoretical_loss": 3.4505314585658877,
      "tokens_seen": 1865285632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422424242424242,
      "loss": 2.5441,
      "theoretical_loss": 3.450521211515828,
      "tokens_seen": 1865351168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044222222222222227,
      "loss": 2.6244,
      "theoretical_loss": 3.4505109649265755,
      "tokens_seen": 1865416704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422020202020202,
      "loss": 2.8392,
      "theoretical_loss": 3.4505007187980916,
      "tokens_seen": 1865482240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421818181818182,
      "loss": 2.7035,
      "theoretical_loss": 3.4504904731303405,
      "tokens_seen": 1865547776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044216161616161614,
      "loss": 2.767,
      "theoretical_loss": 3.4504802279232845,
      "tokens_seen": 1865613312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421414141414142,
      "loss": 2.6047,
      "theoretical_loss": 3.4504699831768875,
      "tokens_seen": 1865678848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044212121212121216,
      "loss": 2.3905,
      "theoretical_loss": 3.4504597388911122,
      "tokens_seen": 1865744384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421010101010101,
      "loss": 2.7433,
      "theoretical_loss": 3.4504494950659215,
      "tokens_seen": 1865809920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420808080808081,
      "loss": 2.7318,
      "theoretical_loss": 3.4504392517012796,
      "tokens_seen": 1865875456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044206060606060603,
      "loss": 2.7193,
      "theoretical_loss": 3.450429008797148,
      "tokens_seen": 1865940992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420404040404041,
      "loss": 2.6802,
      "theoretical_loss": 3.450418766353491,
      "tokens_seen": 1866006528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044202020202020205,
      "loss": 2.5807,
      "theoretical_loss": 3.450408524370271,
      "tokens_seen": 1866072064
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442,
      "loss": 2.5472,
      "theoretical_loss": 3.450398282847452,
      "tokens_seen": 1866137600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044197979797979797,
      "loss": 2.7356,
      "theoretical_loss": 3.4503880417849966,
      "tokens_seen": 1866203136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004419595959595959,
      "loss": 2.7341,
      "theoretical_loss": 3.450377801182868,
      "tokens_seen": 1866268672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000441939393939394,
      "loss": 2.8592,
      "theoretical_loss": 3.4503675610410287,
      "tokens_seen": 1866334208
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 1028516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.410780429840088,
      "objective/train/theoretical_loss": 3.4503573213594434,
      "objective/train/tokens_used": 225258976,
      "theoretical_loss": 3.4503573213594434,
      "tokens_seen": 1866399744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044191919191919195,
      "loss": 2.5768,
      "theoretical_loss": 3.4503573213594434,
      "tokens_seen": 1866399744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418989898989899,
      "loss": 2.6665,
      "theoretical_loss": 3.450347082138074,
      "tokens_seen": 1866465280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044187878787878786,
      "loss": 2.7986,
      "theoretical_loss": 3.4503368433768835,
      "tokens_seen": 1866530816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044185858585858587,
      "loss": 2.6747,
      "theoretical_loss": 3.450326605075836,
      "tokens_seen": 1866596352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418383838383839,
      "loss": 2.6103,
      "theoretical_loss": 3.450316367234894,
      "tokens_seen": 1866661888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044181818181818184,
      "loss": 2.3946,
      "theoretical_loss": 3.450306129854021,
      "tokens_seen": 1866727424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417979797979798,
      "loss": 2.7839,
      "theoretical_loss": 3.45029589293318,
      "tokens_seen": 1866792960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044177777777777775,
      "loss": 2.5326,
      "theoretical_loss": 3.4502856564723343,
      "tokens_seen": 1866858496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044175757575757576,
      "loss": 2.6544,
      "theoretical_loss": 3.450275420471447,
      "tokens_seen": 1866924032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417373737373738,
      "loss": 2.5638,
      "theoretical_loss": 3.4502651849304815,
      "tokens_seen": 1866989568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044171717171717173,
      "loss": 2.8576,
      "theoretical_loss": 3.4502549498494,
      "tokens_seen": 1867055104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416969696969697,
      "loss": 2.5272,
      "theoretical_loss": 3.4502447152281674,
      "tokens_seen": 1867120640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416767676767677,
      "loss": 2.7956,
      "theoretical_loss": 3.4502344810667456,
      "tokens_seen": 1867186176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044165656565656566,
      "loss": 2.7912,
      "theoretical_loss": 3.4502242473650986,
      "tokens_seen": 1867251712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044163636363636367,
      "loss": 2.6048,
      "theoretical_loss": 3.4502140141231887,
      "tokens_seen": 1867317248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416161616161616,
      "loss": 2.652,
      "theoretical_loss": 3.4502037813409796,
      "tokens_seen": 1867382784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415959595959596,
      "loss": 3.0165,
      "theoretical_loss": 3.450193549018435,
      "tokens_seen": 1867448320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415757575757576,
      "loss": 2.4882,
      "theoretical_loss": 3.4501833171555174,
      "tokens_seen": 1867513856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044155555555555555,
      "loss": 2.6933,
      "theoretical_loss": 3.4501730857521906,
      "tokens_seen": 1867579392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044153535353535356,
      "loss": 2.5284,
      "theoretical_loss": 3.4501628548084167,
      "tokens_seen": 1867644928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415151515151515,
      "loss": 2.6121,
      "theoretical_loss": 3.4501526243241605,
      "tokens_seen": 1867710464
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044149494949494953,
      "loss": 2.6236,
      "theoretical_loss": 3.4501423942993843,
      "tokens_seen": 1867776000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414747474747475,
      "loss": 2.724,
      "theoretical_loss": 3.4501321647340513,
      "tokens_seen": 1867841536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044145454545454544,
      "loss": 2.5005,
      "theoretical_loss": 3.4501219356281254,
      "tokens_seen": 1867907072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044143434343434345,
      "loss": 2.7259,
      "theoretical_loss": 3.450111706981569,
      "tokens_seen": 1867972608
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1029117,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.526183605194092,
      "objective/train/theoretical_loss": 3.4501014787943456,
      "objective/train/tokens_used": 226897376,
      "theoretical_loss": 3.4501014787943456,
      "tokens_seen": 1868038144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414141414141414,
      "loss": 2.4437,
      "theoretical_loss": 3.4501014787943456,
      "tokens_seen": 1868038144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413939393939394,
      "loss": 2.4661,
      "theoretical_loss": 3.4500912510664192,
      "tokens_seen": 1868103680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413737373737374,
      "loss": 2.5983,
      "theoretical_loss": 3.450081023797752,
      "tokens_seen": 1868169216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044135353535353533,
      "loss": 2.4983,
      "theoretical_loss": 3.4500707969883084,
      "tokens_seen": 1868234752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044133333333333335,
      "loss": 2.5441,
      "theoretical_loss": 3.4500605706380503,
      "tokens_seen": 1868300288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044131313131313136,
      "loss": 2.7335,
      "theoretical_loss": 3.450050344746942,
      "tokens_seen": 1868365824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412929292929293,
      "loss": 2.6975,
      "theoretical_loss": 3.450040119314947,
      "tokens_seen": 1868431360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044127272727272727,
      "loss": 2.7856,
      "theoretical_loss": 3.450029894342028,
      "tokens_seen": 1868496896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412525252525252,
      "loss": 2.7327,
      "theoretical_loss": 3.450019669828148,
      "tokens_seen": 1868562432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044123232323232324,
      "loss": 2.394,
      "theoretical_loss": 3.4500094457732704,
      "tokens_seen": 1868627968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044121212121212125,
      "loss": 2.8116,
      "theoretical_loss": 3.449999222177359,
      "tokens_seen": 1868693504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411919191919192,
      "loss": 2.6899,
      "theoretical_loss": 3.449988999040377,
      "tokens_seen": 1868759040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044117171717171716,
      "loss": 2.7514,
      "theoretical_loss": 3.449978776362288,
      "tokens_seen": 1868824576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411515151515151,
      "loss": 2.7282,
      "theoretical_loss": 3.4499685541430543,
      "tokens_seen": 1868890112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411313131313132,
      "loss": 2.7994,
      "theoretical_loss": 3.4499583323826397,
      "tokens_seen": 1868955648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044111111111111114,
      "loss": 2.7855,
      "theoretical_loss": 3.4499481110810075,
      "tokens_seen": 1869021184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410909090909091,
      "loss": 2.6568,
      "theoretical_loss": 3.4499378902381217,
      "tokens_seen": 1869086720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044107070707070706,
      "loss": 2.6813,
      "theoretical_loss": 3.449927669853945,
      "tokens_seen": 1869152256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441050505050505,
      "loss": 2.5984,
      "theoretical_loss": 3.4499174499284404,
      "tokens_seen": 1869217792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410303030303031,
      "loss": 2.7613,
      "theoretical_loss": 3.449907230461572,
      "tokens_seen": 1869283328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044101010101010103,
      "loss": 2.708,
      "theoretical_loss": 3.4498970114533023,
      "tokens_seen": 1869348864
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440989898989899,
      "loss": 2.5817,
      "theoretical_loss": 3.4498867929035955,
      "tokens_seen": 1869414400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044096969696969695,
      "loss": 2.7155,
      "theoretical_loss": 3.4498765748124143,
      "tokens_seen": 1869479936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044094949494949496,
      "loss": 2.931,
      "theoretical_loss": 3.449866357179723,
      "tokens_seen": 1869545472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044092929292929297,
      "loss": 2.7079,
      "theoretical_loss": 3.4498561400054832,
      "tokens_seen": 1869611008
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1029470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.821355104446411,
      "objective/train/theoretical_loss": 3.44984592328966,
      "objective/train/tokens_used": 228535776,
      "theoretical_loss": 3.44984592328966,
      "tokens_seen": 1869676544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044090909090909093,
      "loss": 2.7606,
      "theoretical_loss": 3.44984592328966,
      "tokens_seen": 1869676544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408888888888889,
      "loss": 2.6028,
      "theoretical_loss": 3.449835707032216,
      "tokens_seen": 1869742080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044086868686868684,
      "loss": 2.73,
      "theoretical_loss": 3.4498254912331148,
      "tokens_seen": 1869807616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044084848484848485,
      "loss": 2.5199,
      "theoretical_loss": 3.4498152758923197,
      "tokens_seen": 1869873152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044082828282828286,
      "loss": 2.6589,
      "theoretical_loss": 3.4498050610097932,
      "tokens_seen": 1869938688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408080808080808,
      "loss": 2.7433,
      "theoretical_loss": 3.4497948465855,
      "tokens_seen": 1870004224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407878787878788,
      "loss": 2.7544,
      "theoretical_loss": 3.4497846326194033,
      "tokens_seen": 1870069760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407676767676768,
      "loss": 2.585,
      "theoretical_loss": 3.449774419111466,
      "tokens_seen": 1870135296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044074747474747474,
      "loss": 2.6629,
      "theoretical_loss": 3.449764206061652,
      "tokens_seen": 1870200832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044072727272727276,
      "loss": 2.7832,
      "theoretical_loss": 3.4497539934699235,
      "tokens_seen": 1870266368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407070707070707,
      "loss": 2.7545,
      "theoretical_loss": 3.4497437813362453,
      "tokens_seen": 1870331904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044068686868686867,
      "loss": 2.8311,
      "theoretical_loss": 3.44973356966058,
      "tokens_seen": 1870397440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406666666666667,
      "loss": 3.011,
      "theoretical_loss": 3.449723358442892,
      "tokens_seen": 1870462976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044064646464646464,
      "loss": 2.5545,
      "theoretical_loss": 3.449713147683143,
      "tokens_seen": 1870528512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044062626262626265,
      "loss": 2.8115,
      "theoretical_loss": 3.449702937381298,
      "tokens_seen": 1870594048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406060606060606,
      "loss": 2.8322,
      "theoretical_loss": 3.44969272753732,
      "tokens_seen": 1870659584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405858585858586,
      "loss": 2.8499,
      "theoretical_loss": 3.449682518151172,
      "tokens_seen": 1870725120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405656565656566,
      "loss": 2.9033,
      "theoretical_loss": 3.4496723092228176,
      "tokens_seen": 1870790656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044054545454545453,
      "loss": 2.5401,
      "theoretical_loss": 3.449662100752221,
      "tokens_seen": 1870856192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044052525252525254,
      "loss": 2.7093,
      "theoretical_loss": 3.449651892739344,
      "tokens_seen": 1870921728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405050505050505,
      "loss": 2.4943,
      "theoretical_loss": 3.449641685184152,
      "tokens_seen": 1870987264
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404848484848485,
      "loss": 2.7457,
      "theoretical_loss": 3.449631478086607,
      "tokens_seen": 1871052800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044046464646464647,
      "loss": 2.6909,
      "theoretical_loss": 3.4496212714466727,
      "tokens_seen": 1871118336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404444444444445,
      "loss": 2.8722,
      "theoretical_loss": 3.449611065264313,
      "tokens_seen": 1871183872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044042424242424243,
      "loss": 2.7971,
      "theoretical_loss": 3.449600859539491,
      "tokens_seen": 1871249408
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1030757,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.707038402557373,
      "objective/train/theoretical_loss": 3.4495906542721704,
      "objective/train/tokens_used": 230174176,
      "theoretical_loss": 3.4495906542721704,
      "tokens_seen": 1871314944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044040404040404044,
      "loss": 2.6459,
      "theoretical_loss": 3.4495906542721704,
      "tokens_seen": 1871314944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403838383838384,
      "loss": 2.5098,
      "theoretical_loss": 3.4495804494623146,
      "tokens_seen": 1871380480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044036363636363636,
      "loss": 2.7132,
      "theoretical_loss": 3.4495702451098875,
      "tokens_seen": 1871446016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044034343434343437,
      "loss": 2.6227,
      "theoretical_loss": 3.449560041214851,
      "tokens_seen": 1871511552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403232323232323,
      "loss": 2.7804,
      "theoretical_loss": 3.4495498377771705,
      "tokens_seen": 1871577088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044030303030303034,
      "loss": 2.7748,
      "theoretical_loss": 3.4495396347968086,
      "tokens_seen": 1871642624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402828282828283,
      "loss": 2.6117,
      "theoretical_loss": 3.4495294322737293,
      "tokens_seen": 1871708160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044026262626262625,
      "loss": 2.6279,
      "theoretical_loss": 3.449519230207895,
      "tokens_seen": 1871773696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044024242424242426,
      "loss": 2.7321,
      "theoretical_loss": 3.4495090285992704,
      "tokens_seen": 1871839232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402222222222222,
      "loss": 2.7325,
      "theoretical_loss": 3.4494988274478184,
      "tokens_seen": 1871904768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044020202020202023,
      "loss": 2.7222,
      "theoretical_loss": 3.4494886267535025,
      "tokens_seen": 1871970304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401818181818182,
      "loss": 2.5767,
      "theoretical_loss": 3.4494784265162863,
      "tokens_seen": 1872035840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044016161616161614,
      "loss": 2.425,
      "theoretical_loss": 3.4494682267361334,
      "tokens_seen": 1872101376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044014141414141416,
      "loss": 2.5256,
      "theoretical_loss": 3.449458027413007,
      "tokens_seen": 1872166912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044012121212121217,
      "loss": 2.5125,
      "theoretical_loss": 3.4494478285468713,
      "tokens_seen": 1872232448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401010101010101,
      "loss": 2.6882,
      "theoretical_loss": 3.4494376301376892,
      "tokens_seen": 1872297984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400808080808081,
      "loss": 2.6736,
      "theoretical_loss": 3.4494274321854244,
      "tokens_seen": 1872363520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044006060606060604,
      "loss": 2.6651,
      "theoretical_loss": 3.4494172346900402,
      "tokens_seen": 1872429056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044004040404040405,
      "loss": 2.7091,
      "theoretical_loss": 3.449407037651501,
      "tokens_seen": 1872494592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044002020202020206,
      "loss": 2.6583,
      "theoretical_loss": 3.4493968410697695,
      "tokens_seen": 1872560128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044,
      "loss": 2.7233,
      "theoretical_loss": 3.449386644944809,
      "tokens_seen": 1872625664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043997979797979797,
      "loss": 2.6792,
      "theoretical_loss": 3.449376449276584,
      "tokens_seen": 1872691200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043995959595959593,
      "loss": 2.6617,
      "theoretical_loss": 3.449366254065058,
      "tokens_seen": 1872756736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439939393939394,
      "loss": 2.8347,
      "theoretical_loss": 3.449356059310194,
      "tokens_seen": 1872822272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043991919191919195,
      "loss": 2.6349,
      "theoretical_loss": 3.4493458650119555,
      "tokens_seen": 1872887808
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1031383,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7423930168151855,
      "objective/train/theoretical_loss": 3.4493356711703065,
      "objective/train/tokens_used": 231812576,
      "theoretical_loss": 3.4493356711703065,
      "tokens_seen": 1872953344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398989898989899,
      "loss": 2.6487,
      "theoretical_loss": 3.4493356711703065,
      "tokens_seen": 1872953344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043987878787878787,
      "loss": 2.6297,
      "theoretical_loss": 3.4493254777852096,
      "tokens_seen": 1873018880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398585858585858,
      "loss": 2.742,
      "theoretical_loss": 3.44931528485663,
      "tokens_seen": 1873084416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398383838383839,
      "loss": 2.7249,
      "theoretical_loss": 3.4493050923845305,
      "tokens_seen": 1873149952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043981818181818184,
      "loss": 2.4449,
      "theoretical_loss": 3.4492949003688747,
      "tokens_seen": 1873215488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397979797979798,
      "loss": 2.6473,
      "theoretical_loss": 3.4492847088096257,
      "tokens_seen": 1873281024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043977777777777776,
      "loss": 2.7582,
      "theoretical_loss": 3.4492745177067476,
      "tokens_seen": 1873346560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043975757575757577,
      "loss": 2.7019,
      "theoretical_loss": 3.449264327060204,
      "tokens_seen": 1873412096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397373737373738,
      "loss": 2.5911,
      "theoretical_loss": 3.4492541368699587,
      "tokens_seen": 1873477632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043971717171717174,
      "loss": 2.5838,
      "theoretical_loss": 3.449243947135974,
      "tokens_seen": 1873543168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396969696969697,
      "loss": 2.7736,
      "theoretical_loss": 3.4492337578582157,
      "tokens_seen": 1873608704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043967676767676765,
      "loss": 2.7215,
      "theoretical_loss": 3.449223569036646,
      "tokens_seen": 1873674240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043965656565656566,
      "loss": 2.6328,
      "theoretical_loss": 3.449213380671229,
      "tokens_seen": 1873739776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043963636363636367,
      "loss": 3.0429,
      "theoretical_loss": 3.4492031927619275,
      "tokens_seen": 1873805312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043961616161616163,
      "loss": 2.6799,
      "theoretical_loss": 3.4491930053087065,
      "tokens_seen": 1873870848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395959595959596,
      "loss": 2.6578,
      "theoretical_loss": 3.4491828183115283,
      "tokens_seen": 1873936384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395757575757576,
      "loss": 2.5755,
      "theoretical_loss": 3.449172631770357,
      "tokens_seen": 1874001920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043955555555555555,
      "loss": 2.5907,
      "theoretical_loss": 3.4491624456851566,
      "tokens_seen": 1874067456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043953535353535357,
      "loss": 2.6247,
      "theoretical_loss": 3.4491522600558904,
      "tokens_seen": 1874132992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395151515151515,
      "loss": 2.7296,
      "theoretical_loss": 3.4491420748825226,
      "tokens_seen": 1874198528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394949494949495,
      "loss": 2.5939,
      "theoretical_loss": 3.449131890165016,
      "tokens_seen": 1874264064
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394747474747475,
      "loss": 2.6861,
      "theoretical_loss": 3.449121705903335,
      "tokens_seen": 1874329600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043945454545454545,
      "loss": 2.6701,
      "theoretical_loss": 3.4491115220974424,
      "tokens_seen": 1874395136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043943434343434346,
      "loss": 2.4645,
      "theoretical_loss": 3.449101338747303,
      "tokens_seen": 1874460672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394141414141414,
      "loss": 2.5077,
      "theoretical_loss": 3.449091155852879,
      "tokens_seen": 1874526208
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1032514,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0022151470184326,
      "objective/train/theoretical_loss": 3.4490809734141354,
      "objective/train/tokens_used": 233450976,
      "theoretical_loss": 3.4490809734141354,
      "tokens_seen": 1874591744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393939393939394,
      "loss": 2.7027,
      "theoretical_loss": 3.4490809734141354,
      "tokens_seen": 1874591744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393737373737374,
      "loss": 2.6109,
      "theoretical_loss": 3.4490707914310352,
      "tokens_seen": 1874657280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043935353535353534,
      "loss": 2.7608,
      "theoretical_loss": 3.449060609903543,
      "tokens_seen": 1874722816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043933333333333335,
      "loss": 2.5745,
      "theoretical_loss": 3.4490504288316215,
      "tokens_seen": 1874788352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393131313131313,
      "loss": 2.6831,
      "theoretical_loss": 3.4490402482152343,
      "tokens_seen": 1874853888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392929292929293,
      "loss": 2.8604,
      "theoretical_loss": 3.4490300680543458,
      "tokens_seen": 1874919424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392727272727273,
      "loss": 2.5409,
      "theoretical_loss": 3.449019888348919,
      "tokens_seen": 1874984960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043925252525252523,
      "loss": 2.5154,
      "theoretical_loss": 3.4490097090989185,
      "tokens_seen": 1875050496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043923232323232324,
      "loss": 2.6237,
      "theoretical_loss": 3.4489995303043073,
      "tokens_seen": 1875116032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043921212121212125,
      "loss": 2.4967,
      "theoretical_loss": 3.448989351965049,
      "tokens_seen": 1875181568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391919191919192,
      "loss": 2.6085,
      "theoretical_loss": 3.448979174081108,
      "tokens_seen": 1875247104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043917171717171717,
      "loss": 2.7374,
      "theoretical_loss": 3.4489689966524475,
      "tokens_seen": 1875312640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391515151515151,
      "loss": 2.3794,
      "theoretical_loss": 3.4489588196790315,
      "tokens_seen": 1875378176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043913131313131314,
      "loss": 2.6384,
      "theoretical_loss": 3.448948643160823,
      "tokens_seen": 1875443712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043911111111111115,
      "loss": 2.7276,
      "theoretical_loss": 3.448938467097787,
      "tokens_seen": 1875509248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390909090909091,
      "loss": 2.5831,
      "theoretical_loss": 3.448928291489886,
      "tokens_seen": 1875574784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043907070707070706,
      "loss": 2.431,
      "theoretical_loss": 3.448918116337085,
      "tokens_seen": 1875640320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439050505050505,
      "loss": 2.662,
      "theoretical_loss": 3.448907941639346,
      "tokens_seen": 1875705856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390303030303031,
      "loss": 2.4266,
      "theoretical_loss": 3.448897767396635,
      "tokens_seen": 1875771392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043901010101010104,
      "loss": 2.7484,
      "theoretical_loss": 3.448887593608914,
      "tokens_seen": 1875836928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438989898989899,
      "loss": 2.6149,
      "theoretical_loss": 3.448877420276147,
      "tokens_seen": 1875902464
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043896969696969695,
      "loss": 2.6773,
      "theoretical_loss": 3.448867247398298,
      "tokens_seen": 1875968000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389494949494949,
      "loss": 2.6709,
      "theoretical_loss": 3.448857074975332,
      "tokens_seen": 1876033536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438929292929293,
      "loss": 2.669,
      "theoretical_loss": 3.448846903007211,
      "tokens_seen": 1876099072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043890909090909093,
      "loss": 2.5903,
      "theoretical_loss": 3.448836731493899,
      "tokens_seen": 1876164608
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1033016,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4885294437408447,
      "objective/train/theoretical_loss": 3.44882656043536,
      "objective/train/tokens_used": 235089376,
      "theoretical_loss": 3.44882656043536,
      "tokens_seen": 1876230144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388888888888889,
      "loss": 2.6458,
      "theoretical_loss": 3.44882656043536,
      "tokens_seen": 1876230144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043886868686868685,
      "loss": 2.7158,
      "theoretical_loss": 3.4488163898315585,
      "tokens_seen": 1876295680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388484848484848,
      "loss": 2.576,
      "theoretical_loss": 3.4488062196824574,
      "tokens_seen": 1876361216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043882828282828287,
      "loss": 2.5016,
      "theoretical_loss": 3.448796049988021,
      "tokens_seen": 1876426752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388080808080808,
      "loss": 2.5476,
      "theoretical_loss": 3.448785880748213,
      "tokens_seen": 1876492288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387878787878788,
      "loss": 2.7925,
      "theoretical_loss": 3.4487757119629965,
      "tokens_seen": 1876557824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043876767676767674,
      "loss": 2.7359,
      "theoretical_loss": 3.4487655436323363,
      "tokens_seen": 1876623360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387474747474748,
      "loss": 2.5205,
      "theoretical_loss": 3.448755375756196,
      "tokens_seen": 1876688896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043872727272727276,
      "loss": 2.7716,
      "theoretical_loss": 3.4487452083345396,
      "tokens_seen": 1876754432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387070707070707,
      "loss": 2.6872,
      "theoretical_loss": 3.44873504136733,
      "tokens_seen": 1876819968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386868686868687,
      "loss": 2.5491,
      "theoretical_loss": 3.4487248748545314,
      "tokens_seen": 1876885504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043866666666666663,
      "loss": 2.7096,
      "theoretical_loss": 3.448714708796108,
      "tokens_seen": 1876951040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386464646464647,
      "loss": 2.6948,
      "theoretical_loss": 3.4487045431920236,
      "tokens_seen": 1877016576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043862626262626265,
      "loss": 2.4952,
      "theoretical_loss": 3.4486943780422417,
      "tokens_seen": 1877082112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386060606060606,
      "loss": 2.6549,
      "theoretical_loss": 3.4486842133467266,
      "tokens_seen": 1877147648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043858585858585857,
      "loss": 2.4912,
      "theoretical_loss": 3.4486740491054415,
      "tokens_seen": 1877213184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385656565656566,
      "loss": 2.7272,
      "theoretical_loss": 3.448663885318351,
      "tokens_seen": 1877278720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385454545454546,
      "loss": 2.9654,
      "theoretical_loss": 3.448653721985418,
      "tokens_seen": 1877344256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043852525252525255,
      "loss": 2.7963,
      "theoretical_loss": 3.448643559106607,
      "tokens_seen": 1877409792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385050505050505,
      "loss": 2.6008,
      "theoretical_loss": 3.448633396681882,
      "tokens_seen": 1877475328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043848484848484846,
      "loss": 2.6553,
      "theoretical_loss": 3.4486232347112065,
      "tokens_seen": 1877540864
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043846464646464647,
      "loss": 2.7418,
      "theoretical_loss": 3.448613073194544,
      "tokens_seen": 1877606400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384444444444445,
      "loss": 2.9507,
      "theoretical_loss": 3.4486029121318595,
      "tokens_seen": 1877671936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043842424242424244,
      "loss": 2.7075,
      "theoretical_loss": 3.448592751523116,
      "tokens_seen": 1877737472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384040404040404,
      "loss": 2.6087,
      "theoretical_loss": 3.4485825913682775,
      "tokens_seen": 1877803008
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1034169,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.757983922958374,
      "objective/train/theoretical_loss": 3.448572431667308,
      "objective/train/tokens_used": 236727776,
      "theoretical_loss": 3.448572431667308,
      "tokens_seen": 1877868544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383838383838384,
      "loss": 2.4029,
      "theoretical_loss": 3.448572431667308,
      "tokens_seen": 1877868544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043836363636363636,
      "loss": 2.6623,
      "theoretical_loss": 3.448562272420171,
      "tokens_seen": 1877934080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383434343434344,
      "loss": 2.6641,
      "theoretical_loss": 3.4485521136268313,
      "tokens_seen": 1877999616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043832323232323233,
      "loss": 2.6791,
      "theoretical_loss": 3.4485419552872525,
      "tokens_seen": 1878065152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383030303030303,
      "loss": 2.5486,
      "theoretical_loss": 3.4485317974013974,
      "tokens_seen": 1878130688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382828282828283,
      "loss": 2.7649,
      "theoretical_loss": 3.4485216399692313,
      "tokens_seen": 1878196224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043826262626262626,
      "loss": 2.7451,
      "theoretical_loss": 3.4485114829907175,
      "tokens_seen": 1878261760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043824242424242427,
      "loss": 2.7333,
      "theoretical_loss": 3.44850132646582,
      "tokens_seen": 1878327296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382222222222222,
      "loss": 2.6926,
      "theoretical_loss": 3.4484911703945027,
      "tokens_seen": 1878392832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043820202020202024,
      "loss": 2.4957,
      "theoretical_loss": 3.4484810147767293,
      "tokens_seen": 1878458368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381818181818182,
      "loss": 2.5296,
      "theoretical_loss": 3.448470859612464,
      "tokens_seen": 1878523904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043816161616161615,
      "loss": 2.9351,
      "theoretical_loss": 3.4484607049016702,
      "tokens_seen": 1878589440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043814141414141416,
      "loss": 2.7707,
      "theoretical_loss": 3.448450550644313,
      "tokens_seen": 1878654976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381212121212121,
      "loss": 2.5848,
      "theoretical_loss": 3.4484403968403554,
      "tokens_seen": 1878720512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043810101010101013,
      "loss": 2.7145,
      "theoretical_loss": 3.448430243489762,
      "tokens_seen": 1878786048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380808080808081,
      "loss": 2.7403,
      "theoretical_loss": 3.4484200905924953,
      "tokens_seen": 1878851584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043806060606060604,
      "loss": 2.778,
      "theoretical_loss": 3.448409938148521,
      "tokens_seen": 1878917120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043804040404040405,
      "loss": 2.4632,
      "theoretical_loss": 3.448399786157802,
      "tokens_seen": 1878982656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043802020202020206,
      "loss": 2.5384,
      "theoretical_loss": 3.4483896346203027,
      "tokens_seen": 1879048192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438,
      "loss": 2.6639,
      "theoretical_loss": 3.448379483535987,
      "tokens_seen": 1879113728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437979797979798,
      "loss": 2.6435,
      "theoretical_loss": 3.4483693329048184,
      "tokens_seen": 1879179264
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043795959595959593,
      "loss": 2.61,
      "theoretical_loss": 3.4483591827267617,
      "tokens_seen": 1879244800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043793939393939395,
      "loss": 2.7294,
      "theoretical_loss": 3.4483490330017803,
      "tokens_seen": 1879310336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043791919191919196,
      "loss": 2.4736,
      "theoretical_loss": 3.4483388837298383,
      "tokens_seen": 1879375872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378989898989899,
      "loss": 2.7058,
      "theoretical_loss": 3.4483287349108993,
      "tokens_seen": 1879441408
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1034876,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5990076065063477,
      "objective/train/theoretical_loss": 3.448318586544928,
      "objective/train/tokens_used": 238366176,
      "theoretical_loss": 3.448318586544928,
      "tokens_seen": 1879506944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043787878787878787,
      "loss": 2.8493,
      "theoretical_loss": 3.448318586544928,
      "tokens_seen": 1879506944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043785858585858583,
      "loss": 2.5927,
      "theoretical_loss": 3.448308438631888,
      "tokens_seen": 1879572480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378383838383839,
      "loss": 2.7262,
      "theoretical_loss": 3.4482982911717435,
      "tokens_seen": 1879638016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043781818181818185,
      "loss": 2.8021,
      "theoretical_loss": 3.4482881441644584,
      "tokens_seen": 1879703552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377979797979798,
      "loss": 2.8252,
      "theoretical_loss": 3.4482779976099964,
      "tokens_seen": 1879769088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043777777777777776,
      "loss": 2.7085,
      "theoretical_loss": 3.448267851508322,
      "tokens_seen": 1879834624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377575757575757,
      "loss": 2.636,
      "theoretical_loss": 3.448257705859399,
      "tokens_seen": 1879900160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377373737373738,
      "loss": 2.7353,
      "theoretical_loss": 3.4482475606631913,
      "tokens_seen": 1879965696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043771717171717174,
      "loss": 2.6714,
      "theoretical_loss": 3.448237415919663,
      "tokens_seen": 1880031232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376969696969697,
      "loss": 2.5129,
      "theoretical_loss": 3.448227271628778,
      "tokens_seen": 1880096768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043767676767676766,
      "loss": 2.812,
      "theoretical_loss": 3.4482171277905005,
      "tokens_seen": 1880162304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043765656565656567,
      "loss": 2.6062,
      "theoretical_loss": 3.448206984404795,
      "tokens_seen": 1880227840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376363636363637,
      "loss": 2.7784,
      "theoretical_loss": 3.448196841471624,
      "tokens_seen": 1880293376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043761616161616164,
      "loss": 2.6684,
      "theoretical_loss": 3.4481866989909533,
      "tokens_seen": 1880358912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375959595959596,
      "loss": 2.5526,
      "theoretical_loss": 3.4481765569627463,
      "tokens_seen": 1880424448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043757575757575755,
      "loss": 2.494,
      "theoretical_loss": 3.448166415386966,
      "tokens_seen": 1880489984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043755555555555556,
      "loss": 2.8102,
      "theoretical_loss": 3.448156274263578,
      "tokens_seen": 1880555520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043753535353535357,
      "loss": 2.7983,
      "theoretical_loss": 3.4481461335925463,
      "tokens_seen": 1880621056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043751515151515153,
      "loss": 2.645,
      "theoretical_loss": 3.4481359933738336,
      "tokens_seen": 1880686592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374949494949495,
      "loss": 2.5337,
      "theoretical_loss": 3.4481258536074053,
      "tokens_seen": 1880752128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374747474747475,
      "loss": 2.8128,
      "theoretical_loss": 3.448115714293225,
      "tokens_seen": 1880817664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043745454545454545,
      "loss": 2.5904,
      "theoretical_loss": 3.4481055754312564,
      "tokens_seen": 1880883200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043743434343434346,
      "loss": 2.8262,
      "theoretical_loss": 3.448095437021464,
      "tokens_seen": 1880948736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374141414141414,
      "loss": 2.548,
      "theoretical_loss": 3.448085299063812,
      "tokens_seen": 1881014272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373939393939394,
      "loss": 2.2551,
      "theoretical_loss": 3.448075161558264,
      "tokens_seen": 1881079808
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1035977,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.632066488265991,
      "objective/train/theoretical_loss": 3.448065024504784,
      "objective/train/tokens_used": 240004576,
      "theoretical_loss": 3.448065024504784,
      "tokens_seen": 1881145344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373737373737374,
      "loss": 2.8682,
      "theoretical_loss": 3.448065024504784,
      "tokens_seen": 1881145344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043735353535353535,
      "loss": 2.567,
      "theoretical_loss": 3.448054887903337,
      "tokens_seen": 1881210880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043733333333333336,
      "loss": 2.6562,
      "theoretical_loss": 3.448044751753886,
      "tokens_seen": 1881276416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373131313131313,
      "loss": 2.5353,
      "theoretical_loss": 3.448034616056396,
      "tokens_seen": 1881341952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372929292929293,
      "loss": 2.6606,
      "theoretical_loss": 3.448024480810831,
      "tokens_seen": 1881407488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372727272727273,
      "loss": 2.7143,
      "theoretical_loss": 3.4480143460171546,
      "tokens_seen": 1881473024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043725252525252524,
      "loss": 2.723,
      "theoretical_loss": 3.448004211675331,
      "tokens_seen": 1881538560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043723232323232325,
      "loss": 2.8467,
      "theoretical_loss": 3.447994077785325,
      "tokens_seen": 1881604096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372121212121212,
      "loss": 2.727,
      "theoretical_loss": 3.4479839443471,
      "tokens_seen": 1881669632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371919191919192,
      "loss": 2.7398,
      "theoretical_loss": 3.4479738113606198,
      "tokens_seen": 1881735168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371717171717172,
      "loss": 2.4879,
      "theoretical_loss": 3.4479636788258494,
      "tokens_seen": 1881800704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043715151515151513,
      "loss": 2.8289,
      "theoretical_loss": 3.4479535467427525,
      "tokens_seen": 1881866240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043713131313131314,
      "loss": 2.7452,
      "theoretical_loss": 3.4479434151112933,
      "tokens_seen": 1881931776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043711111111111115,
      "loss": 2.546,
      "theoretical_loss": 3.4479332839314365,
      "tokens_seen": 1881997312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370909090909091,
      "loss": 2.6797,
      "theoretical_loss": 3.447923153203145,
      "tokens_seen": 1882062848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043707070707070707,
      "loss": 2.7142,
      "theoretical_loss": 3.447913022926384,
      "tokens_seen": 1882128384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370505050505051,
      "loss": 2.75,
      "theoretical_loss": 3.4479028931011175,
      "tokens_seen": 1882193920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043703030303030303,
      "loss": 2.8057,
      "theoretical_loss": 3.4478927637273094,
      "tokens_seen": 1882259456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043701010101010105,
      "loss": 2.827,
      "theoretical_loss": 3.447882634804923,
      "tokens_seen": 1882324992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000436989898989899,
      "loss": 2.4703,
      "theoretical_loss": 3.4478725063339244,
      "tokens_seen": 1882390528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043696969696969696,
      "loss": 2.749,
      "theoretical_loss": 3.4478623783142766,
      "tokens_seen": 1882456064
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043694949494949497,
      "loss": 2.6755,
      "theoretical_loss": 3.447852250745944,
      "tokens_seen": 1882521600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043692929292929293,
      "loss": 2.6679,
      "theoretical_loss": 3.4478421236288908,
      "tokens_seen": 1882587136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043690909090909094,
      "loss": 2.7283,
      "theoretical_loss": 3.447831996963081,
      "tokens_seen": 1882652672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004368888888888889,
      "loss": 2.8893,
      "theoretical_loss": 3.447821870748479,
      "tokens_seen": 1882718208
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 1036495,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.311638116836548,
      "objective/train/theoretical_loss": 3.4478117449850485,
      "objective/train/tokens_used": 241642976,
      "theoretical_loss": 3.4478117449850485,
      "tokens_seen": 1882783744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043686868686868685,
      "loss": 2.602,
      "theoretical_loss": 3.4478117449850485,
      "tokens_seen": 1882783744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043684848484848486,
      "loss": 2.7245,
      "theoretical_loss": 3.447801619672754,
      "tokens_seen": 1882849280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004368282828282829,
      "loss": 2.6415,
      "theoretical_loss": 3.44779149481156,
      "tokens_seen": 1882914816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043680808080808083,
      "loss": 2.8219,
      "theoretical_loss": 3.4477813704014313,
      "tokens_seen": 1882980352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367878787878788,
      "loss": 2.4011,
      "theoretical_loss": 3.44777124644233,
      "tokens_seen": 1883045888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043676767676767674,
      "loss": 2.6543,
      "theoretical_loss": 3.4477611229342227,
      "tokens_seen": 1883111424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043674747474747476,
      "loss": 2.66,
      "theoretical_loss": 3.447750999877072,
      "tokens_seen": 1883176960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043672727272727277,
      "loss": 3.0482,
      "theoretical_loss": 3.447740877270842,
      "tokens_seen": 1883242496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367070707070707,
      "loss": 2.9646,
      "theoretical_loss": 3.4477307551154985,
      "tokens_seen": 1883308032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366868686868687,
      "loss": 2.713,
      "theoretical_loss": 3.4477206334110044,
      "tokens_seen": 1883373568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043666666666666664,
      "loss": 2.5537,
      "theoretical_loss": 3.4477105121573244,
      "tokens_seen": 1883439104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366464646464647,
      "loss": 2.7314,
      "theoretical_loss": 3.447700391354423,
      "tokens_seen": 1883504640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043662626262626266,
      "loss": 2.5515,
      "theoretical_loss": 3.4476902710022634,
      "tokens_seen": 1883570176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366060606060606,
      "loss": 2.6342,
      "theoretical_loss": 3.447680151100811,
      "tokens_seen": 1883635712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365858585858586,
      "loss": 2.5765,
      "theoretical_loss": 3.4476700316500293,
      "tokens_seen": 1883701248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043656565656565653,
      "loss": 2.624,
      "theoretical_loss": 3.447659912649883,
      "tokens_seen": 1883766784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365454545454546,
      "loss": 2.7815,
      "theoretical_loss": 3.447649794100336,
      "tokens_seen": 1883832320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043652525252525255,
      "loss": 2.4913,
      "theoretical_loss": 3.4476396760013532,
      "tokens_seen": 1883897856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365050505050505,
      "loss": 2.7279,
      "theoretical_loss": 3.447629558352898,
      "tokens_seen": 1883963392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043648484848484847,
      "loss": 2.5147,
      "theoretical_loss": 3.4476194411549352,
      "tokens_seen": 1884028928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364646464646465,
      "loss": 2.669,
      "theoretical_loss": 3.447609324407429,
      "tokens_seen": 1884094464
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364444444444445,
      "loss": 2.5508,
      "theoretical_loss": 3.4475992081103435,
      "tokens_seen": 1884160000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043642424242424244,
      "loss": 2.7432,
      "theoretical_loss": 3.4475890922636436,
      "tokens_seen": 1884225536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364040404040404,
      "loss": 2.6209,
      "theoretical_loss": 3.447578976867293,
      "tokens_seen": 1884291072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043638383838383836,
      "loss": 2.4768,
      "theoretical_loss": 3.4475688619212557,
      "tokens_seen": 1884356608
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1037793,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5640203952789307,
      "objective/train/theoretical_loss": 3.4475587474254965,
      "objective/train/tokens_used": 243281376,
      "theoretical_loss": 3.4475587474254965,
      "tokens_seen": 1884422144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043636363636363637,
      "loss": 2.8154,
      "theoretical_loss": 3.4475587474254965,
      "tokens_seen": 1884422144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363434343434344,
      "loss": 2.7487,
      "theoretical_loss": 3.4475486333799794,
      "tokens_seen": 1884487680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043632323232323234,
      "loss": 2.6674,
      "theoretical_loss": 3.4475385197846693,
      "tokens_seen": 1884553216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363030303030303,
      "loss": 2.6732,
      "theoretical_loss": 3.4475284066395293,
      "tokens_seen": 1884618752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362828282828283,
      "loss": 2.7621,
      "theoretical_loss": 3.4475182939445252,
      "tokens_seen": 1884684288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043626262626262626,
      "loss": 2.6823,
      "theoretical_loss": 3.4475081816996207,
      "tokens_seen": 1884749824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362424242424243,
      "loss": 2.9771,
      "theoretical_loss": 3.4474980699047792,
      "tokens_seen": 1884815360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043622222222222223,
      "loss": 2.6468,
      "theoretical_loss": 3.4474879585599667,
      "tokens_seen": 1884880896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362020202020202,
      "loss": 2.6748,
      "theoretical_loss": 3.447477847665146,
      "tokens_seen": 1884946432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361818181818182,
      "loss": 2.6497,
      "theoretical_loss": 3.4474677372202827,
      "tokens_seen": 1885011968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043616161616161616,
      "loss": 2.4553,
      "theoretical_loss": 3.4474576272253405,
      "tokens_seen": 1885077504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043614141414141417,
      "loss": 2.7674,
      "theoretical_loss": 3.4474475176802835,
      "tokens_seen": 1885143040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361212121212121,
      "loss": 2.7152,
      "theoretical_loss": 3.4474374085850763,
      "tokens_seen": 1885208576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043610101010101013,
      "loss": 2.5882,
      "theoretical_loss": 3.447427299939683,
      "tokens_seen": 1885274112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360808080808081,
      "loss": 2.7256,
      "theoretical_loss": 3.447417191744069,
      "tokens_seen": 1885339648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043606060606060605,
      "loss": 2.6481,
      "theoretical_loss": 3.4474070839981974,
      "tokens_seen": 1885405184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043604040404040406,
      "loss": 2.6377,
      "theoretical_loss": 3.447396976702033,
      "tokens_seen": 1885470720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436020202020202,
      "loss": 2.8219,
      "theoretical_loss": 3.4473868698555403,
      "tokens_seen": 1885536256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436,
      "loss": 2.7823,
      "theoretical_loss": 3.447376763458683,
      "tokens_seen": 1885601792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435979797979798,
      "loss": 2.7027,
      "theoretical_loss": 3.4473666575114263,
      "tokens_seen": 1885667328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043595959595959594,
      "loss": 2.7249,
      "theoretical_loss": 3.447356552013735,
      "tokens_seen": 1885732864
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043593939393939395,
      "loss": 2.7471,
      "theoretical_loss": 3.4473464469655717,
      "tokens_seen": 1885798400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043591919191919196,
      "loss": 2.6487,
      "theoretical_loss": 3.447336342366903,
      "tokens_seen": 1885863936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358989898989899,
      "loss": 2.4869,
      "theoretical_loss": 3.4473262382176912,
      "tokens_seen": 1885929472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358787878787879,
      "loss": 2.766,
      "theoretical_loss": 3.447316134517902,
      "tokens_seen": 1885995008
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1038946,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.652374267578125,
      "objective/train/theoretical_loss": 3.447306031267499,
      "objective/train/tokens_used": 244919776,
      "theoretical_loss": 3.447306031267499,
      "tokens_seen": 1886060544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043585858585858583,
      "loss": 2.7403,
      "theoretical_loss": 3.447306031267499,
      "tokens_seen": 1886060544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043583838383838384,
      "loss": 2.7421,
      "theoretical_loss": 3.4472959284664473,
      "tokens_seen": 1886126080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043581818181818186,
      "loss": 2.7329,
      "theoretical_loss": 3.4472858261147112,
      "tokens_seen": 1886191616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357979797979798,
      "loss": 2.6731,
      "theoretical_loss": 3.447275724212255,
      "tokens_seen": 1886257152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043577777777777777,
      "loss": 2.684,
      "theoretical_loss": 3.4472656227590424,
      "tokens_seen": 1886322688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357575757575757,
      "loss": 2.4135,
      "theoretical_loss": 3.4472555217550394,
      "tokens_seen": 1886388224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357373737373738,
      "loss": 2.6845,
      "theoretical_loss": 3.4472454212002086,
      "tokens_seen": 1886453760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043571717171717175,
      "loss": 2.8128,
      "theoretical_loss": 3.4472353210945155,
      "tokens_seen": 1886519296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356969696969697,
      "loss": 2.6487,
      "theoretical_loss": 3.4472252214379244,
      "tokens_seen": 1886584832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043567676767676766,
      "loss": 2.8778,
      "theoretical_loss": 3.4472151222304,
      "tokens_seen": 1886650368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356565656565656,
      "loss": 2.8099,
      "theoretical_loss": 3.447205023471906,
      "tokens_seen": 1886715904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356363636363637,
      "loss": 2.6793,
      "theoretical_loss": 3.4471949251624077,
      "tokens_seen": 1886781440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043561616161616164,
      "loss": 2.8376,
      "theoretical_loss": 3.4471848273018684,
      "tokens_seen": 1886846976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355959595959596,
      "loss": 2.816,
      "theoretical_loss": 3.4471747298902535,
      "tokens_seen": 1886912512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043557575757575755,
      "loss": 2.5422,
      "theoretical_loss": 3.447164632927527,
      "tokens_seen": 1886978048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355555555555555,
      "loss": 2.7872,
      "theoretical_loss": 3.447154536413654,
      "tokens_seen": 1887043584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355353535353536,
      "loss": 2.5499,
      "theoretical_loss": 3.4471444403485982,
      "tokens_seen": 1887109120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043551515151515153,
      "loss": 2.8599,
      "theoretical_loss": 3.447134344732324,
      "tokens_seen": 1887174656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354949494949495,
      "loss": 2.5081,
      "theoretical_loss": 3.447124249564797,
      "tokens_seen": 1887240192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043547474747474745,
      "loss": 2.5106,
      "theoretical_loss": 3.44711415484598,
      "tokens_seen": 1887305728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043545454545454546,
      "loss": 2.8259,
      "theoretical_loss": 3.447104060575839,
      "tokens_seen": 1887371264
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043543434343434347,
      "loss": 2.6445,
      "theoretical_loss": 3.4470939667543377,
      "tokens_seen": 1887436800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354141414141414,
      "loss": 2.6661,
      "theoretical_loss": 3.44708387338144,
      "tokens_seen": 1887502336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353939393939394,
      "loss": 2.4127,
      "theoretical_loss": 3.447073780457112,
      "tokens_seen": 1887567872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043537373737373734,
      "loss": 2.5971,
      "theoretical_loss": 3.4470636879813172,
      "tokens_seen": 1887633408
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1039523,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.781736135482788,
      "objective/train/theoretical_loss": 3.44705359595402,
      "objective/train/tokens_used": 246558176,
      "theoretical_loss": 3.44705359595402,
      "tokens_seen": 1887698944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043535353535353535,
      "loss": 2.7614,
      "theoretical_loss": 3.44705359595402,
      "tokens_seen": 1887698944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043533333333333336,
      "loss": 2.8807,
      "theoretical_loss": 3.447043504375185,
      "tokens_seen": 1887764480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353131313131313,
      "loss": 2.7627,
      "theoretical_loss": 3.447033413244777,
      "tokens_seen": 1887830016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352929292929293,
      "loss": 2.763,
      "theoretical_loss": 3.4470233225627602,
      "tokens_seen": 1887895552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352727272727273,
      "loss": 2.7779,
      "theoretical_loss": 3.447013232329099,
      "tokens_seen": 1887961088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352525252525253,
      "loss": 2.7414,
      "theoretical_loss": 3.4470031425437586,
      "tokens_seen": 1888026624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043523232323232325,
      "loss": 2.8604,
      "theoretical_loss": 3.446993053206703,
      "tokens_seen": 1888092160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352121212121212,
      "loss": 2.4655,
      "theoretical_loss": 3.4469829643178964,
      "tokens_seen": 1888157696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043519191919191917,
      "loss": 2.7379,
      "theoretical_loss": 3.4469728758773037,
      "tokens_seen": 1888223232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351717171717172,
      "loss": 2.8034,
      "theoretical_loss": 3.44696278788489,
      "tokens_seen": 1888288768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351515151515152,
      "loss": 2.7015,
      "theoretical_loss": 3.446952700340619,
      "tokens_seen": 1888354304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043513131313131315,
      "loss": 2.6724,
      "theoretical_loss": 3.446942613244455,
      "tokens_seen": 1888419840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351111111111111,
      "loss": 2.9569,
      "theoretical_loss": 3.4469325265963637,
      "tokens_seen": 1888485376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350909090909091,
      "loss": 2.6612,
      "theoretical_loss": 3.4469224403963086,
      "tokens_seen": 1888550912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043507070707070707,
      "loss": 2.868,
      "theoretical_loss": 3.4469123546442546,
      "tokens_seen": 1888616448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350505050505051,
      "loss": 2.6853,
      "theoretical_loss": 3.4469022693401667,
      "tokens_seen": 1888681984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043503030303030304,
      "loss": 2.698,
      "theoretical_loss": 3.446892184484009,
      "tokens_seen": 1888747520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435010101010101,
      "loss": 2.7729,
      "theoretical_loss": 3.4468821000757464,
      "tokens_seen": 1888813056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434989898989899,
      "loss": 2.7598,
      "theoretical_loss": 3.4468720161153428,
      "tokens_seen": 1888878592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043496969696969696,
      "loss": 2.5709,
      "theoretical_loss": 3.446861932602763,
      "tokens_seen": 1888944128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434949494949495,
      "loss": 2.7712,
      "theoretical_loss": 3.446851849537972,
      "tokens_seen": 1889009664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043492929292929293,
      "loss": 2.7276,
      "theoretical_loss": 3.446841766920934,
      "tokens_seen": 1889075200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043490909090909094,
      "loss": 2.6929,
      "theoretical_loss": 3.446831684751614,
      "tokens_seen": 1889140736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348888888888889,
      "loss": 2.6453,
      "theoretical_loss": 3.446821603029976,
      "tokens_seen": 1889206272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043486868686868686,
      "loss": 2.5445,
      "theoretical_loss": 3.4468115217559854,
      "tokens_seen": 1889271808
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1040960,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6786797046661377,
      "objective/train/theoretical_loss": 3.446801440929606,
      "objective/train/tokens_used": 248196576,
      "theoretical_loss": 3.446801440929606,
      "tokens_seen": 1889337344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043484848484848487,
      "loss": 2.8896,
      "theoretical_loss": 3.446801440929606,
      "tokens_seen": 1889337344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348282828282828,
      "loss": 2.7437,
      "theoretical_loss": 3.4467913605508027,
      "tokens_seen": 1889402880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043480808080808084,
      "loss": 2.7,
      "theoretical_loss": 3.4467812806195397,
      "tokens_seen": 1889468416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347878787878788,
      "loss": 2.7376,
      "theoretical_loss": 3.4467712011357827,
      "tokens_seen": 1889533952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043476767676767675,
      "loss": 2.7908,
      "theoretical_loss": 3.4467611220994954,
      "tokens_seen": 1889599488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043474747474747476,
      "loss": 2.7504,
      "theoretical_loss": 3.446751043510642,
      "tokens_seen": 1889665024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043472727272727277,
      "loss": 2.4802,
      "theoretical_loss": 3.4467409653691887,
      "tokens_seen": 1889730560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043470707070707073,
      "loss": 2.7347,
      "theoretical_loss": 3.446730887675099,
      "tokens_seen": 1889796096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346868686868687,
      "loss": 2.6914,
      "theoretical_loss": 3.4467208104283373,
      "tokens_seen": 1889861632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043466666666666664,
      "loss": 2.8039,
      "theoretical_loss": 3.446710733628869,
      "tokens_seen": 1889927168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043464646464646465,
      "loss": 2.9309,
      "theoretical_loss": 3.446700657276658,
      "tokens_seen": 1889992704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043462626262626267,
      "loss": 2.7041,
      "theoretical_loss": 3.4466905813716697,
      "tokens_seen": 1890058240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346060606060606,
      "loss": 2.8169,
      "theoretical_loss": 3.446680505913868,
      "tokens_seen": 1890123776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345858585858586,
      "loss": 2.67,
      "theoretical_loss": 3.446670430903218,
      "tokens_seen": 1890189312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043456565656565654,
      "loss": 2.7819,
      "theoretical_loss": 3.4466603563396845,
      "tokens_seen": 1890254848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345454545454546,
      "loss": 2.817,
      "theoretical_loss": 3.4466502822232323,
      "tokens_seen": 1890320384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043452525252525256,
      "loss": 2.8273,
      "theoretical_loss": 3.4466402085538252,
      "tokens_seen": 1890385920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345050505050505,
      "loss": 2.863,
      "theoretical_loss": 3.446630135331428,
      "tokens_seen": 1890451456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043448484848484847,
      "loss": 2.6555,
      "theoretical_loss": 3.4466200625560064,
      "tokens_seen": 1890516992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043446464646464643,
      "loss": 2.7843,
      "theoretical_loss": 3.446609990227524,
      "tokens_seen": 1890582528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344444444444445,
      "loss": 2.6445,
      "theoretical_loss": 3.446599918345946,
      "tokens_seen": 1890648064
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043442424242424245,
      "loss": 2.6006,
      "theoretical_loss": 3.4465898469112366,
      "tokens_seen": 1890713600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344040404040404,
      "loss": 2.928,
      "theoretical_loss": 3.446579775923362,
      "tokens_seen": 1890779136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043438383838383836,
      "loss": 2.6915,
      "theoretical_loss": 3.4465697053822844,
      "tokens_seen": 1890844672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343636363636364,
      "loss": 2.9646,
      "theoretical_loss": 3.4465596352879704,
      "tokens_seen": 1890910208
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1041567,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.472703456878662,
      "objective/train/theoretical_loss": 3.446549565640384,
      "objective/train/tokens_used": 249834976,
      "theoretical_loss": 3.446549565640384,
      "tokens_seen": 1890975744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343434343434344,
      "loss": 2.969,
      "theoretical_loss": 3.446549565640384,
      "tokens_seen": 1890975744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043432323232323234,
      "loss": 2.5333,
      "theoretical_loss": 3.44653949643949,
      "tokens_seen": 1891041280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343030303030303,
      "loss": 2.7445,
      "theoretical_loss": 3.4465294276852534,
      "tokens_seen": 1891106816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043428282828282826,
      "loss": 2.5441,
      "theoretical_loss": 3.446519359377638,
      "tokens_seen": 1891172352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043426262626262627,
      "loss": 2.5551,
      "theoretical_loss": 3.44650929151661,
      "tokens_seen": 1891237888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342424242424243,
      "loss": 2.7024,
      "theoretical_loss": 3.4464992241021326,
      "tokens_seen": 1891303424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043422222222222224,
      "loss": 2.6171,
      "theoretical_loss": 3.446489157134171,
      "tokens_seen": 1891368960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342020202020202,
      "loss": 2.7406,
      "theoretical_loss": 3.4464790906126908,
      "tokens_seen": 1891434496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341818181818182,
      "loss": 2.6309,
      "theoretical_loss": 3.446469024537656,
      "tokens_seen": 1891500032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043416161616161616,
      "loss": 2.6551,
      "theoretical_loss": 3.446458958909031,
      "tokens_seen": 1891565568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043414141414141417,
      "loss": 2.6236,
      "theoretical_loss": 3.4464488937267808,
      "tokens_seen": 1891631104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043412121212121213,
      "loss": 2.6514,
      "theoretical_loss": 3.4464388289908703,
      "tokens_seen": 1891696640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341010101010101,
      "loss": 2.6576,
      "theoretical_loss": 3.4464287647012646,
      "tokens_seen": 1891762176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340808080808081,
      "loss": 2.6109,
      "theoretical_loss": 3.4464187008579277,
      "tokens_seen": 1891827712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043406060606060605,
      "loss": 2.5924,
      "theoretical_loss": 3.446408637460825,
      "tokens_seen": 1891893248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043404040404040406,
      "loss": 2.4555,
      "theoretical_loss": 3.446398574509921,
      "tokens_seen": 1891958784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434020202020202,
      "loss": 2.7578,
      "theoretical_loss": 3.44638851200518,
      "tokens_seen": 1892024320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043400000000000003,
      "loss": 2.4286,
      "theoretical_loss": 3.4463784499465673,
      "tokens_seen": 1892089856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433979797979798,
      "loss": 2.4674,
      "theoretical_loss": 3.4463683883340472,
      "tokens_seen": 1892155392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043395959595959595,
      "loss": 2.7483,
      "theoretical_loss": 3.4463583271675855,
      "tokens_seen": 1892220928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043393939393939396,
      "loss": 2.8933,
      "theoretical_loss": 3.446348266447146,
      "tokens_seen": 1892286464
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339191919191919,
      "loss": 2.6796,
      "theoretical_loss": 3.446338206172694,
      "tokens_seen": 1892352000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338989898989899,
      "loss": 2.6755,
      "theoretical_loss": 3.446328146344194,
      "tokens_seen": 1892417536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338787878787879,
      "loss": 2.6965,
      "theoretical_loss": 3.44631808696161,
      "tokens_seen": 1892483072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043385858585858584,
      "loss": 2.6102,
      "theoretical_loss": 3.4463080280249088,
      "tokens_seen": 1892548608
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1042675,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5914690494537354,
      "objective/train/theoretical_loss": 3.4462979695340534,
      "objective/train/tokens_used": 251473376,
      "theoretical_loss": 3.4462979695340534,
      "tokens_seen": 1892614144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043383838383838385,
      "loss": 2.5992,
      "theoretical_loss": 3.4462979695340534,
      "tokens_seen": 1892614144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043381818181818186,
      "loss": 2.7988,
      "theoretical_loss": 3.4462879114890095,
      "tokens_seen": 1892679680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337979797979798,
      "loss": 2.7278,
      "theoretical_loss": 3.446277853889742,
      "tokens_seen": 1892745216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337777777777778,
      "loss": 2.7914,
      "theoretical_loss": 3.4462677967362145,
      "tokens_seen": 1892810752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043375757575757573,
      "loss": 2.5553,
      "theoretical_loss": 3.4462577400283934,
      "tokens_seen": 1892876288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043373737373737374,
      "loss": 2.8135,
      "theoretical_loss": 3.4462476837662424,
      "tokens_seen": 1892941824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043371717171717175,
      "loss": 2.7649,
      "theoretical_loss": 3.446237627949727,
      "tokens_seen": 1893007360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336969696969697,
      "loss": 2.454,
      "theoretical_loss": 3.4462275725788114,
      "tokens_seen": 1893072896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043367676767676767,
      "loss": 2.9034,
      "theoretical_loss": 3.4462175176534613,
      "tokens_seen": 1893138432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336565656565656,
      "loss": 2.8346,
      "theoretical_loss": 3.44620746317364,
      "tokens_seen": 1893203968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043363636363636364,
      "loss": 2.3368,
      "theoretical_loss": 3.446197409139314,
      "tokens_seen": 1893269504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043361616161616165,
      "loss": 2.8816,
      "theoretical_loss": 3.446187355550448,
      "tokens_seen": 1893335040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335959595959596,
      "loss": 2.591,
      "theoretical_loss": 3.4461773024070057,
      "tokens_seen": 1893400576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043357575757575756,
      "loss": 2.8308,
      "theoretical_loss": 3.4461672497089526,
      "tokens_seen": 1893466112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043355555555555557,
      "loss": 2.7431,
      "theoretical_loss": 3.4461571974562535,
      "tokens_seen": 1893531648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335353535353536,
      "loss": 2.6631,
      "theoretical_loss": 3.4461471456488733,
      "tokens_seen": 1893597184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043351515151515154,
      "loss": 2.6796,
      "theoretical_loss": 3.446137094286777,
      "tokens_seen": 1893662720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334949494949495,
      "loss": 2.5672,
      "theoretical_loss": 3.4461270433699296,
      "tokens_seen": 1893728256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043347474747474745,
      "loss": 2.5745,
      "theoretical_loss": 3.446116992898295,
      "tokens_seen": 1893793792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043345454545454546,
      "loss": 2.7892,
      "theoretical_loss": 3.4461069428718387,
      "tokens_seen": 1893859328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334343434343435,
      "loss": 2.7141,
      "theoretical_loss": 3.4460968932905263,
      "tokens_seen": 1893924864
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043341414141414143,
      "loss": 2.7473,
      "theoretical_loss": 3.4460868441543218,
      "tokens_seen": 1893990400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333939393939394,
      "loss": 2.5743,
      "theoretical_loss": 3.44607679546319,
      "tokens_seen": 1894055936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043337373737373735,
      "loss": 2.7776,
      "theoretical_loss": 3.4460667472170963,
      "tokens_seen": 1894121472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333535353535354,
      "loss": 2.7774,
      "theoretical_loss": 3.446056699416005,
      "tokens_seen": 1894187008
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1043835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.616513967514038,
      "objective/train/theoretical_loss": 3.446046652059882,
      "objective/train/tokens_used": 253111776,
      "theoretical_loss": 3.446046652059882,
      "tokens_seen": 1894252544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043333333333333337,
      "loss": 2.6525,
      "theoretical_loss": 3.446046652059882,
      "tokens_seen": 1894252544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333131313131313,
      "loss": 2.6293,
      "theoretical_loss": 3.4460366051486915,
      "tokens_seen": 1894318080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332929292929293,
      "loss": 2.7239,
      "theoretical_loss": 3.446026558682398,
      "tokens_seen": 1894383616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043327272727272724,
      "loss": 2.7973,
      "theoretical_loss": 3.446016512660967,
      "tokens_seen": 1894449152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332525252525253,
      "loss": 2.6836,
      "theoretical_loss": 3.4460064670843638,
      "tokens_seen": 1894514688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043323232323232326,
      "loss": 2.7825,
      "theoretical_loss": 3.445996421952552,
      "tokens_seen": 1894580224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332121212121212,
      "loss": 2.7624,
      "theoretical_loss": 3.445986377265498,
      "tokens_seen": 1894645760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331919191919192,
      "loss": 2.5716,
      "theoretical_loss": 3.445976333023166,
      "tokens_seen": 1894711296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331717171717172,
      "loss": 2.7031,
      "theoretical_loss": 3.4459662892255203,
      "tokens_seen": 1894776832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331515151515152,
      "loss": 2.5949,
      "theoretical_loss": 3.4459562458725275,
      "tokens_seen": 1894842368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043313131313131315,
      "loss": 2.5145,
      "theoretical_loss": 3.445946202964151,
      "tokens_seen": 1894907904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331111111111111,
      "loss": 2.6517,
      "theoretical_loss": 3.4459361605003567,
      "tokens_seen": 1894973440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043309090909090907,
      "loss": 2.5671,
      "theoretical_loss": 3.4459261184811085,
      "tokens_seen": 1895038976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330707070707071,
      "loss": 2.6993,
      "theoretical_loss": 3.4459160769063724,
      "tokens_seen": 1895104512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330505050505051,
      "loss": 2.6661,
      "theoretical_loss": 3.445906035776113,
      "tokens_seen": 1895170048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043303030303030305,
      "loss": 2.8961,
      "theoretical_loss": 3.4458959950902948,
      "tokens_seen": 1895235584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433010101010101,
      "loss": 2.7048,
      "theoretical_loss": 3.445885954848884,
      "tokens_seen": 1895301120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432989898989899,
      "loss": 2.6647,
      "theoretical_loss": 3.445875915051844,
      "tokens_seen": 1895366656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043296969696969697,
      "loss": 2.5512,
      "theoretical_loss": 3.4458658756991407,
      "tokens_seen": 1895432192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432949494949495,
      "loss": 2.8466,
      "theoretical_loss": 3.445855836790739,
      "tokens_seen": 1895497728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043292929292929294,
      "loss": 2.7774,
      "theoretical_loss": 3.4458457983266033,
      "tokens_seen": 1895563264
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329090909090909,
      "loss": 2.8034,
      "theoretical_loss": 3.4458357603066996,
      "tokens_seen": 1895628800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328888888888889,
      "loss": 2.6688,
      "theoretical_loss": 3.445825722730992,
      "tokens_seen": 1895694336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043286868686868686,
      "loss": 2.2792,
      "theoretical_loss": 3.4458156855994457,
      "tokens_seen": 1895759872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328484848484849,
      "loss": 2.6926,
      "theoretical_loss": 3.4458056489120263,
      "tokens_seen": 1895825408
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1044447,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.802863836288452,
      "objective/train/theoretical_loss": 3.4457956126686975,
      "objective/train/tokens_used": 254750176,
      "theoretical_loss": 3.4457956126686975,
      "tokens_seen": 1895890944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043282828282828283,
      "loss": 2.6796,
      "theoretical_loss": 3.4457956126686975,
      "tokens_seen": 1895890944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043280808080808084,
      "loss": 2.7162,
      "theoretical_loss": 3.445785576869426,
      "tokens_seen": 1895956480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327878787878788,
      "loss": 2.6145,
      "theoretical_loss": 3.445775541514175,
      "tokens_seen": 1896022016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043276767676767676,
      "loss": 2.8484,
      "theoretical_loss": 3.445765506602911,
      "tokens_seen": 1896087552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043274747474747477,
      "loss": 2.6471,
      "theoretical_loss": 3.4457554721355983,
      "tokens_seen": 1896153088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327272727272727,
      "loss": 2.6083,
      "theoretical_loss": 3.445745438112202,
      "tokens_seen": 1896218624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043270707070707073,
      "loss": 2.5806,
      "theoretical_loss": 3.4457354045326873,
      "tokens_seen": 1896284160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326868686868687,
      "loss": 2.8685,
      "theoretical_loss": 3.4457253713970184,
      "tokens_seen": 1896349696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043266666666666665,
      "loss": 2.6143,
      "theoretical_loss": 3.4457153387051616,
      "tokens_seen": 1896415232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043264646464646466,
      "loss": 2.7779,
      "theoretical_loss": 3.4457053064570813,
      "tokens_seen": 1896480768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043262626262626267,
      "loss": 2.8286,
      "theoretical_loss": 3.4456952746527425,
      "tokens_seen": 1896546304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043260606060606063,
      "loss": 2.8636,
      "theoretical_loss": 3.44568524329211,
      "tokens_seen": 1896611840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325858585858586,
      "loss": 2.7182,
      "theoretical_loss": 3.44567521237515,
      "tokens_seen": 1896677376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043256565656565654,
      "loss": 2.7157,
      "theoretical_loss": 3.445665181901826,
      "tokens_seen": 1896742912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043254545454545455,
      "loss": 2.5607,
      "theoretical_loss": 3.4456551518721037,
      "tokens_seen": 1896808448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043252525252525256,
      "loss": 2.6054,
      "theoretical_loss": 3.4456451222859483,
      "tokens_seen": 1896873984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325050505050505,
      "loss": 2.5935,
      "theoretical_loss": 3.4456350931433253,
      "tokens_seen": 1896939520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324848484848485,
      "loss": 2.6931,
      "theoretical_loss": 3.445625064444199,
      "tokens_seen": 1897005056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043246464646464643,
      "loss": 2.7926,
      "theoretical_loss": 3.445615036188534,
      "tokens_seen": 1897070592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324444444444445,
      "loss": 2.783,
      "theoretical_loss": 3.445605008376297,
      "tokens_seen": 1897136128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043242424242424246,
      "loss": 2.4811,
      "theoretical_loss": 3.4455949810074515,
      "tokens_seen": 1897201664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324040404040404,
      "loss": 2.6182,
      "theoretical_loss": 3.445584954081964,
      "tokens_seen": 1897267200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043238383838383837,
      "loss": 2.4329,
      "theoretical_loss": 3.4455749275997976,
      "tokens_seen": 1897332736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323636363636363,
      "loss": 2.8432,
      "theoretical_loss": 3.4455649015609193,
      "tokens_seen": 1897398272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323434343434344,
      "loss": 2.5067,
      "theoretical_loss": 3.4455548759652936,
      "tokens_seen": 1897463808
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1045650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5606865882873535,
      "objective/train/theoretical_loss": 3.4455448508128854,
      "objective/train/tokens_used": 256388576,
      "theoretical_loss": 3.4455448508128854,
      "tokens_seen": 1897529344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043232323232323235,
      "loss": 2.4806,
      "theoretical_loss": 3.4455448508128854,
      "tokens_seen": 1897529344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323030303030303,
      "loss": 2.5365,
      "theoretical_loss": 3.4455348261036596,
      "tokens_seen": 1897594880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043228282828282826,
      "loss": 2.4085,
      "theoretical_loss": 3.445524801837582,
      "tokens_seen": 1897660416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322626262626262,
      "loss": 2.7238,
      "theoretical_loss": 3.4455147780146174,
      "tokens_seen": 1897725952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322424242424243,
      "loss": 2.5818,
      "theoretical_loss": 3.4455047546347304,
      "tokens_seen": 1897791488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043222222222222224,
      "loss": 2.8515,
      "theoretical_loss": 3.4454947316978863,
      "tokens_seen": 1897857024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322020202020202,
      "loss": 2.5814,
      "theoretical_loss": 3.445484709204051,
      "tokens_seen": 1897922560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043218181818181816,
      "loss": 3.0203,
      "theoretical_loss": 3.445474687153189,
      "tokens_seen": 1897988096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043216161616161617,
      "loss": 3.0408,
      "theoretical_loss": 3.4454646655452654,
      "tokens_seen": 1898053632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321414141414142,
      "loss": 2.7143,
      "theoretical_loss": 3.4454546443802454,
      "tokens_seen": 1898119168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043212121212121213,
      "loss": 2.8371,
      "theoretical_loss": 3.4454446236580942,
      "tokens_seen": 1898184704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321010101010101,
      "loss": 2.7068,
      "theoretical_loss": 3.445434603378777,
      "tokens_seen": 1898250240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043208080808080805,
      "loss": 2.6067,
      "theoretical_loss": 3.445424583542258,
      "tokens_seen": 1898315776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043206060606060606,
      "loss": 2.607,
      "theoretical_loss": 3.4454145641485043,
      "tokens_seen": 1898381312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043204040404040407,
      "loss": 2.4594,
      "theoretical_loss": 3.445404545197479,
      "tokens_seen": 1898446848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432020202020202,
      "loss": 2.6032,
      "theoretical_loss": 3.4453945266891486,
      "tokens_seen": 1898512384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432,
      "loss": 2.5731,
      "theoretical_loss": 3.4453845086234782,
      "tokens_seen": 1898577920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000431979797979798,
      "loss": 2.7252,
      "theoretical_loss": 3.4453744910004325,
      "tokens_seen": 1898643456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043195959595959595,
      "loss": 2.5493,
      "theoretical_loss": 3.4453644738199767,
      "tokens_seen": 1898708992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043193939393939396,
      "loss": 2.7383,
      "theoretical_loss": 3.4453544570820758,
      "tokens_seen": 1898774528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319191919191919,
      "loss": 2.5631,
      "theoretical_loss": 3.4453444407866956,
      "tokens_seen": 1898840064
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318989898989899,
      "loss": 2.5864,
      "theoretical_loss": 3.4453344249338005,
      "tokens_seen": 1898905600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318787878787879,
      "loss": 2.5916,
      "theoretical_loss": 3.445324409523356,
      "tokens_seen": 1898971136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318585858585859,
      "loss": 2.7156,
      "theoretical_loss": 3.4453143945553277,
      "tokens_seen": 1899036672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043183838383838386,
      "loss": 2.8104,
      "theoretical_loss": 3.4453043800296808,
      "tokens_seen": 1899102208
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 1046191,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9813382625579834,
      "objective/train/theoretical_loss": 3.4452943659463795,
      "objective/train/tokens_used": 258026976,
      "theoretical_loss": 3.4452943659463795,
      "tokens_seen": 1899167744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318181818181818,
      "loss": 2.7552,
      "theoretical_loss": 3.4452943659463795,
      "tokens_seen": 1899167744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317979797979798,
      "loss": 2.7213,
      "theoretical_loss": 3.4452843523053898,
      "tokens_seen": 1899233280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317777777777778,
      "loss": 2.5726,
      "theoretical_loss": 3.445274339106677,
      "tokens_seen": 1899298816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317575757575758,
      "loss": 2.6529,
      "theoretical_loss": 3.445264326350206,
      "tokens_seen": 1899364352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043173737373737375,
      "loss": 2.6968,
      "theoretical_loss": 3.4452543140359415,
      "tokens_seen": 1899429888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317171717171717,
      "loss": 2.6248,
      "theoretical_loss": 3.44524430216385,
      "tokens_seen": 1899495424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316969696969697,
      "loss": 2.8309,
      "theoretical_loss": 3.4452342907338958,
      "tokens_seen": 1899560960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316767676767677,
      "loss": 2.5161,
      "theoretical_loss": 3.445224279746044,
      "tokens_seen": 1899626496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316565656565657,
      "loss": 2.6626,
      "theoretical_loss": 3.4452142692002603,
      "tokens_seen": 1899692032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043163636363636364,
      "loss": 2.6857,
      "theoretical_loss": 3.44520425909651,
      "tokens_seen": 1899757568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043161616161616165,
      "loss": 2.7009,
      "theoretical_loss": 3.445194249434758,
      "tokens_seen": 1899823104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315959595959596,
      "loss": 2.4982,
      "theoretical_loss": 3.44518424021497,
      "tokens_seen": 1899888640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043157575757575757,
      "loss": 2.698,
      "theoretical_loss": 3.4451742314371105,
      "tokens_seen": 1899954176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315555555555556,
      "loss": 2.7291,
      "theoretical_loss": 3.4451642231011452,
      "tokens_seen": 1900019712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043153535353535353,
      "loss": 2.6778,
      "theoretical_loss": 3.4451542152070393,
      "tokens_seen": 1900085248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043151515151515154,
      "loss": 2.676,
      "theoretical_loss": 3.4451442077547583,
      "tokens_seen": 1900150784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314949494949495,
      "loss": 2.7419,
      "theoretical_loss": 3.445134200744267,
      "tokens_seen": 1900216320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043147474747474746,
      "loss": 2.5326,
      "theoretical_loss": 3.44512419417553,
      "tokens_seen": 1900281856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043145454545454547,
      "loss": 2.5492,
      "theoretical_loss": 3.4451141880485148,
      "tokens_seen": 1900347392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314343434343435,
      "loss": 2.4373,
      "theoretical_loss": 3.4451041823631843,
      "tokens_seen": 1900412928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043141414141414144,
      "loss": 2.5494,
      "theoretical_loss": 3.445094177119505,
      "tokens_seen": 1900478464
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313939393939394,
      "loss": 2.5809,
      "theoretical_loss": 3.4450841723174426,
      "tokens_seen": 1900544000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043137373737373735,
      "loss": 2.4989,
      "theoretical_loss": 3.4450741679569608,
      "tokens_seen": 1900609536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043135353535353536,
      "loss": 2.677,
      "theoretical_loss": 3.445064164038026,
      "tokens_seen": 1900675072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313333333333334,
      "loss": 2.6094,
      "theoretical_loss": 3.4450541605606038,
      "tokens_seen": 1900740608
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1047224,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5529582500457764,
      "objective/train/theoretical_loss": 3.4450441575246584,
      "objective/train/tokens_used": 259665376,
      "theoretical_loss": 3.4450441575246584,
      "tokens_seen": 1900806144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043131313131313133,
      "loss": 2.4374,
      "theoretical_loss": 3.4450441575246584,
      "tokens_seen": 1900806144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312929292929293,
      "loss": 2.5186,
      "theoretical_loss": 3.4450341549301564,
      "tokens_seen": 1900871680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043127272727272724,
      "loss": 2.7014,
      "theoretical_loss": 3.4450241527770613,
      "tokens_seen": 1900937216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312525252525253,
      "loss": 2.8584,
      "theoretical_loss": 3.4450141510653403,
      "tokens_seen": 1901002752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043123232323232327,
      "loss": 2.5654,
      "theoretical_loss": 3.4450041497949577,
      "tokens_seen": 1901068288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312121212121212,
      "loss": 2.6917,
      "theoretical_loss": 3.444994148965879,
      "tokens_seen": 1901133824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311919191919192,
      "loss": 2.5157,
      "theoretical_loss": 3.4449841485780697,
      "tokens_seen": 1901199360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043117171717171714,
      "loss": 2.457,
      "theoretical_loss": 3.444974148631495,
      "tokens_seen": 1901264896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311515151515152,
      "loss": 2.7993,
      "theoretical_loss": 3.4449641491261196,
      "tokens_seen": 1901330432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043113131313131316,
      "loss": 2.3877,
      "theoretical_loss": 3.4449541500619096,
      "tokens_seen": 1901395968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311111111111111,
      "loss": 2.762,
      "theoretical_loss": 3.4449441514388304,
      "tokens_seen": 1901461504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043109090909090907,
      "loss": 2.5943,
      "theoretical_loss": 3.4449341532568467,
      "tokens_seen": 1901527040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310707070707071,
      "loss": 2.8475,
      "theoretical_loss": 3.4449241555159245,
      "tokens_seen": 1901592576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310505050505051,
      "loss": 2.5425,
      "theoretical_loss": 3.4449141582160285,
      "tokens_seen": 1901658112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043103030303030305,
      "loss": 2.6034,
      "theoretical_loss": 3.4449041613571247,
      "tokens_seen": 1901723648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000431010101010101,
      "loss": 2.6465,
      "theoretical_loss": 3.4448941649391784,
      "tokens_seen": 1901789184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043098989898989897,
      "loss": 2.8249,
      "theoretical_loss": 3.444884168962154,
      "tokens_seen": 1901854720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430969696969697,
      "loss": 2.6515,
      "theoretical_loss": 3.444874173426018,
      "tokens_seen": 1901920256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430949494949495,
      "loss": 2.7534,
      "theoretical_loss": 3.444864178330735,
      "tokens_seen": 1901985792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043092929292929294,
      "loss": 2.5677,
      "theoretical_loss": 3.444854183676271,
      "tokens_seen": 1902051328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309090909090909,
      "loss": 2.5041,
      "theoretical_loss": 3.444844189462591,
      "tokens_seen": 1902116864
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308888888888889,
      "loss": 2.4229,
      "theoretical_loss": 3.44483419568966,
      "tokens_seen": 1902182400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043086868686868687,
      "loss": 2.8729,
      "theoretical_loss": 3.4448242023574447,
      "tokens_seen": 1902247936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308484848484849,
      "loss": 2.726,
      "theoretical_loss": 3.444814209465909,
      "tokens_seen": 1902313472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043082828282828284,
      "loss": 2.6824,
      "theoretical_loss": 3.444804217015019,
      "tokens_seen": 1902379008
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1047755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.809091329574585,
      "objective/train/theoretical_loss": 3.44479422500474,
      "objective/train/tokens_used": 261303776,
      "theoretical_loss": 3.44479422500474,
      "tokens_seen": 1902444544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308080808080808,
      "loss": 2.5754,
      "theoretical_loss": 3.44479422500474,
      "tokens_seen": 1902444544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307878787878788,
      "loss": 2.6735,
      "theoretical_loss": 3.4447842334350374,
      "tokens_seen": 1902510080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043076767676767676,
      "loss": 2.7893,
      "theoretical_loss": 3.4447742423058765,
      "tokens_seen": 1902575616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043074747474747477,
      "loss": 2.7023,
      "theoretical_loss": 3.4447642516172223,
      "tokens_seen": 1902641152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043072727272727273,
      "loss": 2.6068,
      "theoretical_loss": 3.4447542613690416,
      "tokens_seen": 1902706688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043070707070707074,
      "loss": 2.6704,
      "theoretical_loss": 3.444744271561298,
      "tokens_seen": 1902772224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306868686868687,
      "loss": 2.5841,
      "theoretical_loss": 3.4447342821939584,
      "tokens_seen": 1902837760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043066666666666665,
      "loss": 2.5968,
      "theoretical_loss": 3.4447242932669875,
      "tokens_seen": 1902903296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043064646464646467,
      "loss": 2.4953,
      "theoretical_loss": 3.4447143047803506,
      "tokens_seen": 1902968832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306262626262626,
      "loss": 2.7003,
      "theoretical_loss": 3.444704316734013,
      "tokens_seen": 1903034368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043060606060606063,
      "loss": 2.6531,
      "theoretical_loss": 3.444694329127941,
      "tokens_seen": 1903099904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305858585858586,
      "loss": 2.6429,
      "theoretical_loss": 3.4446843419621,
      "tokens_seen": 1903165440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043056565656565655,
      "loss": 2.6939,
      "theoretical_loss": 3.4446743552364545,
      "tokens_seen": 1903230976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043054545454545456,
      "loss": 2.5613,
      "theoretical_loss": 3.4446643689509697,
      "tokens_seen": 1903296512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043052525252525257,
      "loss": 2.8203,
      "theoretical_loss": 3.444654383105613,
      "tokens_seen": 1903362048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305050505050505,
      "loss": 2.6728,
      "theoretical_loss": 3.4446443977003476,
      "tokens_seen": 1903427584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304848484848485,
      "loss": 2.6016,
      "theoretical_loss": 3.44463441273514,
      "tokens_seen": 1903493120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043046464646464644,
      "loss": 2.7691,
      "theoretical_loss": 3.444624428209956,
      "tokens_seen": 1903558656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043044444444444445,
      "loss": 2.7279,
      "theoretical_loss": 3.4446144441247606,
      "tokens_seen": 1903624192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043042424242424246,
      "loss": 2.7133,
      "theoretical_loss": 3.4446044604795194,
      "tokens_seen": 1903689728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304040404040404,
      "loss": 2.7337,
      "theoretical_loss": 3.4445944772741974,
      "tokens_seen": 1903755264
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303838383838384,
      "loss": 2.7528,
      "theoretical_loss": 3.444584494508761,
      "tokens_seen": 1903820800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043036363636363633,
      "loss": 2.6488,
      "theoretical_loss": 3.444574512183175,
      "tokens_seen": 1903886336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043034343434343434,
      "loss": 2.686,
      "theoretical_loss": 3.444564530297405,
      "tokens_seen": 1903951872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043032323232323235,
      "loss": 2.7676,
      "theoretical_loss": 3.444554548851416,
      "tokens_seen": 1904017408
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1048883,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5599634647369385,
      "objective/train/theoretical_loss": 3.4445445678451745,
      "objective/train/tokens_used": 262942176,
      "theoretical_loss": 3.4445445678451745,
      "tokens_seen": 1904082944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303030303030303,
      "loss": 2.8532,
      "theoretical_loss": 3.4445445678451745,
      "tokens_seen": 1904082944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043028282828282827,
      "loss": 2.8837,
      "theoretical_loss": 3.4445345872786453,
      "tokens_seen": 1904148480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302626262626262,
      "loss": 2.6033,
      "theoretical_loss": 3.444524607151794,
      "tokens_seen": 1904214016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302424242424243,
      "loss": 2.7305,
      "theoretical_loss": 3.444514627464587,
      "tokens_seen": 1904279552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043022222222222225,
      "loss": 2.7995,
      "theoretical_loss": 3.444504648216988,
      "tokens_seen": 1904345088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302020202020202,
      "loss": 2.8029,
      "theoretical_loss": 3.4444946694089635,
      "tokens_seen": 1904410624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043018181818181816,
      "loss": 2.6914,
      "theoretical_loss": 3.4444846910404796,
      "tokens_seen": 1904476160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043016161616161617,
      "loss": 2.6082,
      "theoretical_loss": 3.444474713111501,
      "tokens_seen": 1904541696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301414141414142,
      "loss": 2.8255,
      "theoretical_loss": 3.4444647356219935,
      "tokens_seen": 1904607232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043012121212121214,
      "loss": 2.5401,
      "theoretical_loss": 3.444454758571922,
      "tokens_seen": 1904672768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301010101010101,
      "loss": 2.7175,
      "theoretical_loss": 3.4444447819612534,
      "tokens_seen": 1904738304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043008080808080805,
      "loss": 2.5902,
      "theoretical_loss": 3.444434805789952,
      "tokens_seen": 1904803840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300606060606061,
      "loss": 2.8099,
      "theoretical_loss": 3.444424830057984,
      "tokens_seen": 1904869376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300404040404041,
      "loss": 2.6037,
      "theoretical_loss": 3.444414854765314,
      "tokens_seen": 1904934912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043002020202020203,
      "loss": 2.7313,
      "theoretical_loss": 3.444404879911909,
      "tokens_seen": 1905000448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043,
      "loss": 2.8642,
      "theoretical_loss": 3.4443949054977336,
      "tokens_seen": 1905065984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042997979797979795,
      "loss": 2.777,
      "theoretical_loss": 3.4443849315227535,
      "tokens_seen": 1905131520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429959595959596,
      "loss": 2.5088,
      "theoretical_loss": 3.444374957986934,
      "tokens_seen": 1905197056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042993939393939397,
      "loss": 2.6678,
      "theoretical_loss": 3.444364984890241,
      "tokens_seen": 1905262592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299191919191919,
      "loss": 2.7726,
      "theoretical_loss": 3.4443550122326405,
      "tokens_seen": 1905328128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298989898989899,
      "loss": 2.6621,
      "theoretical_loss": 3.444345040014097,
      "tokens_seen": 1905393664
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298787878787879,
      "loss": 2.8235,
      "theoretical_loss": 3.444335068234577,
      "tokens_seen": 1905459200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298585858585859,
      "loss": 2.807,
      "theoretical_loss": 3.4443250968940458,
      "tokens_seen": 1905524736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042983838383838386,
      "loss": 2.785,
      "theoretical_loss": 3.4443151259924685,
      "tokens_seen": 1905590272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298181818181818,
      "loss": 2.69,
      "theoretical_loss": 3.444305155529811,
      "tokens_seen": 1905655808
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1049481,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5614824295043945,
      "objective/train/theoretical_loss": 3.4442951855060393,
      "objective/train/tokens_used": 264580576,
      "theoretical_loss": 3.4442951855060393,
      "tokens_seen": 1905721344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297979797979798,
      "loss": 2.5091,
      "theoretical_loss": 3.4442951855060393,
      "tokens_seen": 1905721344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297777777777778,
      "loss": 2.7097,
      "theoretical_loss": 3.4442852159211186,
      "tokens_seen": 1905786880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297575757575758,
      "loss": 2.5279,
      "theoretical_loss": 3.444275246775015,
      "tokens_seen": 1905852416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042973737373737375,
      "loss": 2.6529,
      "theoretical_loss": 3.4442652780676926,
      "tokens_seen": 1905917952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297171717171717,
      "loss": 2.6415,
      "theoretical_loss": 3.444255309799119,
      "tokens_seen": 1905983488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296969696969697,
      "loss": 2.7135,
      "theoretical_loss": 3.444245341969258,
      "tokens_seen": 1906049024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296767676767677,
      "loss": 2.7133,
      "theoretical_loss": 3.4442353745780765,
      "tokens_seen": 1906114560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296565656565657,
      "loss": 2.5431,
      "theoretical_loss": 3.44422540762554,
      "tokens_seen": 1906180096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042963636363636365,
      "loss": 2.6903,
      "theoretical_loss": 3.444215441111613,
      "tokens_seen": 1906245632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296161616161616,
      "loss": 2.6747,
      "theoretical_loss": 3.444205475036262,
      "tokens_seen": 1906311168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295959595959596,
      "loss": 2.6883,
      "theoretical_loss": 3.444195509399453,
      "tokens_seen": 1906376704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042957575757575757,
      "loss": 2.5868,
      "theoretical_loss": 3.444185544201151,
      "tokens_seen": 1906442240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295555555555556,
      "loss": 2.8323,
      "theoretical_loss": 3.444175579441322,
      "tokens_seen": 1906507776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042953535353535354,
      "loss": 2.766,
      "theoretical_loss": 3.444165615119931,
      "tokens_seen": 1906573312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042951515151515155,
      "loss": 2.665,
      "theoretical_loss": 3.4441556512369442,
      "tokens_seen": 1906638848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294949494949495,
      "loss": 2.7014,
      "theoretical_loss": 3.4441456877923273,
      "tokens_seen": 1906704384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042947474747474746,
      "loss": 2.6946,
      "theoretical_loss": 3.4441357247860456,
      "tokens_seen": 1906769920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294545454545455,
      "loss": 2.4496,
      "theoretical_loss": 3.4441257622180648,
      "tokens_seen": 1906835456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042943434343434343,
      "loss": 2.6594,
      "theoretical_loss": 3.4441158000883507,
      "tokens_seen": 1906900992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042941414141414144,
      "loss": 2.3002,
      "theoretical_loss": 3.444105838396869,
      "tokens_seen": 1906966528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293939393939394,
      "loss": 2.5308,
      "theoretical_loss": 3.4440958771435852,
      "tokens_seen": 1907032064
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042937373737373736,
      "loss": 2.5182,
      "theoretical_loss": 3.444085916328465,
      "tokens_seen": 1907097600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042935353535353537,
      "loss": 2.5564,
      "theoretical_loss": 3.444075955951474,
      "tokens_seen": 1907163136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293333333333334,
      "loss": 2.5707,
      "theoretical_loss": 3.4440659960125783,
      "tokens_seen": 1907228672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042931313131313134,
      "loss": 2.6122,
      "theoretical_loss": 3.4440560365117427,
      "tokens_seen": 1907294208
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1050597,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.497748374938965,
      "objective/train/theoretical_loss": 3.444046077448934,
      "objective/train/tokens_used": 266218976,
      "theoretical_loss": 3.444046077448934,
      "tokens_seen": 1907359744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292929292929293,
      "loss": 2.6615,
      "theoretical_loss": 3.444046077448934,
      "tokens_seen": 1907359744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042927272727272725,
      "loss": 2.7089,
      "theoretical_loss": 3.4440361188241173,
      "tokens_seen": 1907425280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042925252525252526,
      "loss": 2.6821,
      "theoretical_loss": 3.444026160637258,
      "tokens_seen": 1907490816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042923232323232327,
      "loss": 2.6433,
      "theoretical_loss": 3.4440162028883226,
      "tokens_seen": 1907556352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042921212121212123,
      "loss": 2.7893,
      "theoretical_loss": 3.444006245577276,
      "tokens_seen": 1907621888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291919191919192,
      "loss": 2.4906,
      "theoretical_loss": 3.443996288704084,
      "tokens_seen": 1907687424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042917171717171714,
      "loss": 2.745,
      "theoretical_loss": 3.443986332268713,
      "tokens_seen": 1907752960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291515151515152,
      "loss": 2.4249,
      "theoretical_loss": 3.443976376271128,
      "tokens_seen": 1907818496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042913131313131316,
      "loss": 2.7227,
      "theoretical_loss": 3.4439664207112948,
      "tokens_seen": 1907884032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291111111111111,
      "loss": 2.762,
      "theoretical_loss": 3.4439564655891797,
      "tokens_seen": 1907949568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290909090909091,
      "loss": 2.7586,
      "theoretical_loss": 3.4439465109047473,
      "tokens_seen": 1908015104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042907070707070703,
      "loss": 2.616,
      "theoretical_loss": 3.4439365566579645,
      "tokens_seen": 1908080640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290505050505051,
      "loss": 2.6937,
      "theoretical_loss": 3.4439266028487965,
      "tokens_seen": 1908146176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042903030303030306,
      "loss": 2.6258,
      "theoretical_loss": 3.4439166494772087,
      "tokens_seen": 1908211712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429010101010101,
      "loss": 2.7516,
      "theoretical_loss": 3.4439066965431677,
      "tokens_seen": 1908277248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042898989898989897,
      "loss": 2.7227,
      "theoretical_loss": 3.4438967440466386,
      "tokens_seen": 1908342784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042896969696969693,
      "loss": 2.4991,
      "theoretical_loss": 3.443886791987587,
      "tokens_seen": 1908408320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428949494949495,
      "loss": 2.5172,
      "theoretical_loss": 3.443876840365979,
      "tokens_seen": 1908473856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042892929292929295,
      "loss": 2.6511,
      "theoretical_loss": 3.4438668891817805,
      "tokens_seen": 1908539392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289090909090909,
      "loss": 2.6616,
      "theoretical_loss": 3.443856938434957,
      "tokens_seen": 1908604928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042888888888888886,
      "loss": 2.7852,
      "theoretical_loss": 3.4438469881254745,
      "tokens_seen": 1908670464
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288686868686869,
      "loss": 2.6391,
      "theoretical_loss": 3.443837038253298,
      "tokens_seen": 1908736000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288484848484849,
      "loss": 2.6146,
      "theoretical_loss": 3.443827088818394,
      "tokens_seen": 1908801536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042882828282828284,
      "loss": 2.6307,
      "theoretical_loss": 3.4438171398207285,
      "tokens_seen": 1908867072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288080808080808,
      "loss": 2.4644,
      "theoretical_loss": 3.443807191260267,
      "tokens_seen": 1908932608
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1051477,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.080481767654419,
      "objective/train/theoretical_loss": 3.4437972431369746,
      "objective/train/tokens_used": 267857376,
      "theoretical_loss": 3.4437972431369746,
      "tokens_seen": 1908998144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042878787878787876,
      "loss": 2.712,
      "theoretical_loss": 3.4437972431369746,
      "tokens_seen": 1908998144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042876767676767677,
      "loss": 2.772,
      "theoretical_loss": 3.4437872954508175,
      "tokens_seen": 1909063680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287474747474748,
      "loss": 2.712,
      "theoretical_loss": 3.443777348201762,
      "tokens_seen": 1909129216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042872727272727274,
      "loss": 2.7066,
      "theoretical_loss": 3.443767401389773,
      "tokens_seen": 1909194752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287070707070707,
      "loss": 2.7274,
      "theoretical_loss": 3.4437574550148176,
      "tokens_seen": 1909260288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286868686868687,
      "loss": 2.6661,
      "theoretical_loss": 3.4437475090768603,
      "tokens_seen": 1909325824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042866666666666666,
      "loss": 2.663,
      "theoretical_loss": 3.4437375635758674,
      "tokens_seen": 1909391360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042864646464646467,
      "loss": 2.6362,
      "theoretical_loss": 3.4437276185118053,
      "tokens_seen": 1909456896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042862626262626263,
      "loss": 2.514,
      "theoretical_loss": 3.4437176738846382,
      "tokens_seen": 1909522432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286060606060606,
      "loss": 2.5079,
      "theoretical_loss": 3.4437077296943337,
      "tokens_seen": 1909587968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285858585858586,
      "loss": 2.801,
      "theoretical_loss": 3.4436977859408566,
      "tokens_seen": 1909653504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042856565656565655,
      "loss": 2.6743,
      "theoretical_loss": 3.443687842624173,
      "tokens_seen": 1909719040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042854545454545456,
      "loss": 2.6999,
      "theoretical_loss": 3.443677899744249,
      "tokens_seen": 1909784576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285252525252525,
      "loss": 2.603,
      "theoretical_loss": 3.4436679573010496,
      "tokens_seen": 1909850112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042850505050505053,
      "loss": 2.7413,
      "theoretical_loss": 3.4436580152945413,
      "tokens_seen": 1909915648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284848484848485,
      "loss": 2.3256,
      "theoretical_loss": 3.44364807372469,
      "tokens_seen": 1909981184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042846464646464645,
      "loss": 2.6853,
      "theoretical_loss": 3.4436381325914613,
      "tokens_seen": 1910046720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042844444444444446,
      "loss": 2.4406,
      "theoretical_loss": 3.443628191894821,
      "tokens_seen": 1910112256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284242424242424,
      "loss": 2.7554,
      "theoretical_loss": 3.443618251634735,
      "tokens_seen": 1910177792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284040404040404,
      "loss": 2.4338,
      "theoretical_loss": 3.443608311811169,
      "tokens_seen": 1910243328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283838383838384,
      "loss": 2.5513,
      "theoretical_loss": 3.443598372424089,
      "tokens_seen": 1910308864
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283636363636364,
      "loss": 2.9312,
      "theoretical_loss": 3.443588433473461,
      "tokens_seen": 1910374400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042834343434343435,
      "loss": 2.6278,
      "theoretical_loss": 3.443578494959251,
      "tokens_seen": 1910439936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042832323232323236,
      "loss": 2.8923,
      "theoretical_loss": 3.4435685568814245,
      "tokens_seen": 1910505472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283030303030303,
      "loss": 2.7032,
      "theoretical_loss": 3.4435586192399477,
      "tokens_seen": 1910571008
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1051849,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.255452871322632,
      "objective/train/theoretical_loss": 3.443548682034786,
      "objective/train/tokens_used": 269495776,
      "theoretical_loss": 3.443548682034786,
      "tokens_seen": 1910636544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282828282828283,
      "loss": 2.7182,
      "theoretical_loss": 3.443548682034786,
      "tokens_seen": 1910636544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282626262626263,
      "loss": 2.5879,
      "theoretical_loss": 3.443538745265905,
      "tokens_seen": 1910702080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042824242424242424,
      "loss": 2.5467,
      "theoretical_loss": 3.443528808933272,
      "tokens_seen": 1910767616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042822222222222225,
      "loss": 2.8346,
      "theoretical_loss": 3.443518873036852,
      "tokens_seen": 1910833152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282020202020202,
      "loss": 2.3565,
      "theoretical_loss": 3.44350893757661,
      "tokens_seen": 1910898688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042818181818181817,
      "loss": 2.7003,
      "theoretical_loss": 3.443499002552514,
      "tokens_seen": 1910964224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281616161616162,
      "loss": 2.3673,
      "theoretical_loss": 3.443489067964528,
      "tokens_seen": 1911029760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281414141414142,
      "loss": 2.8019,
      "theoretical_loss": 3.4434791338126187,
      "tokens_seen": 1911095296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042812121212121215,
      "loss": 2.7316,
      "theoretical_loss": 3.443469200096752,
      "tokens_seen": 1911160832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281010101010101,
      "loss": 2.8181,
      "theoretical_loss": 3.4434592668168933,
      "tokens_seen": 1911226368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042808080808080806,
      "loss": 2.7831,
      "theoretical_loss": 3.4434493339730095,
      "tokens_seen": 1911291904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042806060606060607,
      "loss": 2.8366,
      "theoretical_loss": 3.4434394015650653,
      "tokens_seen": 1911357440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280404040404041,
      "loss": 2.5423,
      "theoretical_loss": 3.4434294695930276,
      "tokens_seen": 1911422976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042802020202020204,
      "loss": 2.5892,
      "theoretical_loss": 3.443419538056862,
      "tokens_seen": 1911488512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428,
      "loss": 2.5842,
      "theoretical_loss": 3.4434096069565348,
      "tokens_seen": 1911554048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042797979797979795,
      "loss": 2.7545,
      "theoretical_loss": 3.4433996762920107,
      "tokens_seen": 1911619584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427959595959596,
      "loss": 2.591,
      "theoretical_loss": 3.443389746063257,
      "tokens_seen": 1911685120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427939393939394,
      "loss": 2.8032,
      "theoretical_loss": 3.443379816270239,
      "tokens_seen": 1911750656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042791919191919193,
      "loss": 2.6334,
      "theoretical_loss": 3.4433698869129232,
      "tokens_seen": 1911816192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278989898989899,
      "loss": 2.7227,
      "theoretical_loss": 3.443359957991275,
      "tokens_seen": 1911881728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042787878787878784,
      "loss": 2.7052,
      "theoretical_loss": 3.4433500295052597,
      "tokens_seen": 1911947264
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278585858585859,
      "loss": 2.5298,
      "theoretical_loss": 3.4433401014548446,
      "tokens_seen": 1912012800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042783838383838387,
      "loss": 2.5337,
      "theoretical_loss": 3.4433301738399953,
      "tokens_seen": 1912078336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278181818181818,
      "loss": 2.5673,
      "theoretical_loss": 3.4433202466606767,
      "tokens_seen": 1912143872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277979797979798,
      "loss": 2.8075,
      "theoretical_loss": 3.4433103199168564,
      "tokens_seen": 1912209408
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1053170,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4347217082977295,
      "objective/train/theoretical_loss": 3.4433003936084994,
      "objective/train/tokens_used": 271134176,
      "theoretical_loss": 3.4433003936084994,
      "tokens_seen": 1912274944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277777777777778,
      "loss": 2.6555,
      "theoretical_loss": 3.4433003936084994,
      "tokens_seen": 1912274944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277575757575758,
      "loss": 2.5505,
      "theoretical_loss": 3.443290467735572,
      "tokens_seen": 1912340480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042773737373737376,
      "loss": 2.7639,
      "theoretical_loss": 3.44328054229804,
      "tokens_seen": 1912406016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277171717171717,
      "loss": 2.8434,
      "theoretical_loss": 3.4432706172958687,
      "tokens_seen": 1912471552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276969696969697,
      "loss": 2.7693,
      "theoretical_loss": 3.4432606927290257,
      "tokens_seen": 1912537088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276767676767677,
      "loss": 2.601,
      "theoretical_loss": 3.4432507685974754,
      "tokens_seen": 1912602624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276565656565657,
      "loss": 2.3856,
      "theoretical_loss": 3.443240844901185,
      "tokens_seen": 1912668160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042763636363636365,
      "loss": 2.6203,
      "theoretical_loss": 3.4432309216401196,
      "tokens_seen": 1912733696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276161616161616,
      "loss": 2.762,
      "theoretical_loss": 3.4432209988142457,
      "tokens_seen": 1912799232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275959595959596,
      "loss": 2.5514,
      "theoretical_loss": 3.4432110764235295,
      "tokens_seen": 1912864768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275757575757576,
      "loss": 2.6362,
      "theoretical_loss": 3.4432011544679364,
      "tokens_seen": 1912930304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275555555555556,
      "loss": 2.6411,
      "theoretical_loss": 3.443191232947433,
      "tokens_seen": 1912995840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042753535353535354,
      "loss": 2.7044,
      "theoretical_loss": 3.4431813118619847,
      "tokens_seen": 1913061376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275151515151515,
      "loss": 2.5536,
      "theoretical_loss": 3.443171391211558,
      "tokens_seen": 1913126912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274949494949495,
      "loss": 2.7948,
      "theoretical_loss": 3.4431614709961185,
      "tokens_seen": 1913192448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042747474747474747,
      "loss": 2.7015,
      "theoretical_loss": 3.4431515512156325,
      "tokens_seen": 1913257984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274545454545455,
      "loss": 2.7094,
      "theoretical_loss": 3.4431416318700663,
      "tokens_seen": 1913323520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042743434343434344,
      "loss": 2.7221,
      "theoretical_loss": 3.443131712959386,
      "tokens_seen": 1913389056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042741414141414145,
      "loss": 2.7083,
      "theoretical_loss": 3.443121794483557,
      "tokens_seen": 1913454592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273939393939394,
      "loss": 2.5024,
      "theoretical_loss": 3.4431118764425452,
      "tokens_seen": 1913520128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042737373737373736,
      "loss": 2.5784,
      "theoretical_loss": 3.4431019588363174,
      "tokens_seen": 1913585664
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273535353535354,
      "loss": 2.864,
      "theoretical_loss": 3.4430920416648396,
      "tokens_seen": 1913651200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042733333333333333,
      "loss": 2.8075,
      "theoretical_loss": 3.443082124928077,
      "tokens_seen": 1913716736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042731313131313134,
      "loss": 2.7673,
      "theoretical_loss": 3.443072208625997,
      "tokens_seen": 1913782272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272929292929293,
      "loss": 2.662,
      "theoretical_loss": 3.4430622927585643,
      "tokens_seen": 1913847808
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1053822,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.910274028778076,
      "objective/train/theoretical_loss": 3.443052377325746,
      "objective/train/tokens_used": 272772576,
      "theoretical_loss": 3.443052377325746,
      "tokens_seen": 1913913344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042727272727272726,
      "loss": 2.6898,
      "theoretical_loss": 3.443052377325746,
      "tokens_seen": 1913913344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042725252525252527,
      "loss": 2.5891,
      "theoretical_loss": 3.4430424623275075,
      "tokens_seen": 1913978880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272323232323233,
      "loss": 2.6262,
      "theoretical_loss": 3.4430325477638153,
      "tokens_seen": 1914044416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042721212121212123,
      "loss": 2.5299,
      "theoretical_loss": 3.443022633634635,
      "tokens_seen": 1914109952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271919191919192,
      "loss": 2.6919,
      "theoretical_loss": 3.4430127199399334,
      "tokens_seen": 1914175488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042717171717171715,
      "loss": 2.5389,
      "theoretical_loss": 3.443002806679676,
      "tokens_seen": 1914241024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042715151515151516,
      "loss": 2.4234,
      "theoretical_loss": 3.442992893853829,
      "tokens_seen": 1914306560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042713131313131317,
      "loss": 2.7233,
      "theoretical_loss": 3.4429829814623583,
      "tokens_seen": 1914372096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271111111111111,
      "loss": 2.4909,
      "theoretical_loss": 3.4429730695052303,
      "tokens_seen": 1914437632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270909090909091,
      "loss": 2.7478,
      "theoretical_loss": 3.442963157982412,
      "tokens_seen": 1914503168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042707070707070704,
      "loss": 2.5509,
      "theoretical_loss": 3.4429532468938673,
      "tokens_seen": 1914568704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042705050505050505,
      "loss": 2.525,
      "theoretical_loss": 3.4429433362395643,
      "tokens_seen": 1914634240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042703030303030306,
      "loss": 2.7176,
      "theoretical_loss": 3.442933426019468,
      "tokens_seen": 1914699776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427010101010101,
      "loss": 2.7536,
      "theoretical_loss": 3.442923516233545,
      "tokens_seen": 1914765312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000426989898989899,
      "loss": 2.6505,
      "theoretical_loss": 3.4429136068817616,
      "tokens_seen": 1914830848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042696969696969693,
      "loss": 2.54,
      "theoretical_loss": 3.4429036979640832,
      "tokens_seen": 1914896384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000426949494949495,
      "loss": 2.7148,
      "theoretical_loss": 3.4428937894804763,
      "tokens_seen": 1914961920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042692929292929296,
      "loss": 2.4606,
      "theoretical_loss": 3.442883881430908,
      "tokens_seen": 1915027456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269090909090909,
      "loss": 2.6427,
      "theoretical_loss": 3.4428739738153427,
      "tokens_seen": 1915092992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042688888888888887,
      "loss": 2.5241,
      "theoretical_loss": 3.442864066633747,
      "tokens_seen": 1915158528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268686868686868,
      "loss": 2.7025,
      "theoretical_loss": 3.442854159886088,
      "tokens_seen": 1915224064
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268484848484849,
      "loss": 2.8231,
      "theoretical_loss": 3.442844253572331,
      "tokens_seen": 1915289600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042682828282828285,
      "loss": 2.6401,
      "theoretical_loss": 3.4428343476924432,
      "tokens_seen": 1915355136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268080808080808,
      "loss": 2.7003,
      "theoretical_loss": 3.4428244422463887,
      "tokens_seen": 1915420672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042678787878787876,
      "loss": 2.4902,
      "theoretical_loss": 3.4428145372341357,
      "tokens_seen": 1915486208
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 1054940,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.356700897216797,
      "objective/train/theoretical_loss": 3.4428046326556494,
      "objective/train/tokens_used": 274410976,
      "theoretical_loss": 3.4428046326556494,
      "tokens_seen": 1915551744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042676767676767677,
      "loss": 2.4488,
      "theoretical_loss": 3.4428046326556494,
      "tokens_seen": 1915551744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267474747474748,
      "loss": 2.7953,
      "theoretical_loss": 3.442794728510896,
      "tokens_seen": 1915617280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042672727272727274,
      "loss": 2.6184,
      "theoretical_loss": 3.4427848247998423,
      "tokens_seen": 1915682816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267070707070707,
      "loss": 2.4235,
      "theoretical_loss": 3.4427749215224535,
      "tokens_seen": 1915748352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042668686868686865,
      "loss": 2.7641,
      "theoretical_loss": 3.4427650186786964,
      "tokens_seen": 1915813888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266666666666667,
      "loss": 2.7819,
      "theoretical_loss": 3.442755116268537,
      "tokens_seen": 1915879424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266464646464647,
      "loss": 2.6544,
      "theoretical_loss": 3.4427452142919415,
      "tokens_seen": 1915944960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042662626262626263,
      "loss": 2.6319,
      "theoretical_loss": 3.442735312748876,
      "tokens_seen": 1916010496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266060606060606,
      "loss": 2.7644,
      "theoretical_loss": 3.442725411639307,
      "tokens_seen": 1916076032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265858585858586,
      "loss": 2.6702,
      "theoretical_loss": 3.4427155109632004,
      "tokens_seen": 1916141568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265656565656566,
      "loss": 2.5994,
      "theoretical_loss": 3.4427056107205223,
      "tokens_seen": 1916207104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042654545454545457,
      "loss": 2.7542,
      "theoretical_loss": 3.4426957109112393,
      "tokens_seen": 1916272640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265252525252525,
      "loss": 2.4553,
      "theoretical_loss": 3.442685811535317,
      "tokens_seen": 1916338176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265050505050505,
      "loss": 2.6332,
      "theoretical_loss": 3.4426759125927227,
      "tokens_seen": 1916403712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264848484848485,
      "loss": 2.6213,
      "theoretical_loss": 3.4426660140834215,
      "tokens_seen": 1916469248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264646464646465,
      "loss": 2.6587,
      "theoretical_loss": 3.44265611600738,
      "tokens_seen": 1916534784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042644444444444446,
      "loss": 2.6661,
      "theoretical_loss": 3.442646218364565,
      "tokens_seen": 1916600320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264242424242424,
      "loss": 2.4961,
      "theoretical_loss": 3.442636321154942,
      "tokens_seen": 1916665856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042640404040404043,
      "loss": 2.3362,
      "theoretical_loss": 3.442626424378477,
      "tokens_seen": 1916731392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263838383838384,
      "loss": 2.6343,
      "theoretical_loss": 3.4426165280351366,
      "tokens_seen": 1916796928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263636363636364,
      "loss": 2.6596,
      "theoretical_loss": 3.4426066321248876,
      "tokens_seen": 1916862464
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042634343434343435,
      "loss": 2.8295,
      "theoretical_loss": 3.4425967366476957,
      "tokens_seen": 1916928000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263232323232323,
      "loss": 2.6982,
      "theoretical_loss": 3.442586841603527,
      "tokens_seen": 1916993536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263030303030303,
      "loss": 2.4981,
      "theoretical_loss": 3.4425769469923475,
      "tokens_seen": 1917059072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262828282828283,
      "loss": 2.7179,
      "theoretical_loss": 3.442567052814124,
      "tokens_seen": 1917124608
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1055508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.873598575592041,
      "objective/train/theoretical_loss": 3.4425571590688233,
      "objective/train/tokens_used": 276049376,
      "theoretical_loss": 3.4425571590688233,
      "tokens_seen": 1917190144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262626262626263,
      "loss": 2.8046,
      "theoretical_loss": 3.4425571590688233,
      "tokens_seen": 1917190144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042624242424242425,
      "loss": 2.6275,
      "theoretical_loss": 3.442547265756411,
      "tokens_seen": 1917255680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042622222222222226,
      "loss": 2.5438,
      "theoretical_loss": 3.4425373728768527,
      "tokens_seen": 1917321216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262020202020202,
      "loss": 2.5688,
      "theoretical_loss": 3.4425274804301154,
      "tokens_seen": 1917386752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042618181818181817,
      "loss": 2.5858,
      "theoretical_loss": 3.442517588416165,
      "tokens_seen": 1917452288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261616161616162,
      "loss": 2.3179,
      "theoretical_loss": 3.442507696834969,
      "tokens_seen": 1917517824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042614141414141414,
      "loss": 2.6822,
      "theoretical_loss": 3.442497805686492,
      "tokens_seen": 1917583360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042612121212121215,
      "loss": 2.8473,
      "theoretical_loss": 3.4424879149707017,
      "tokens_seen": 1917648896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261010101010101,
      "loss": 2.7661,
      "theoretical_loss": 3.442478024687563,
      "tokens_seen": 1917714432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042608080808080806,
      "loss": 2.6435,
      "theoretical_loss": 3.442468134837043,
      "tokens_seen": 1917779968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260606060606061,
      "loss": 2.5611,
      "theoretical_loss": 3.4424582454191084,
      "tokens_seen": 1917845504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260404040404041,
      "loss": 2.9118,
      "theoretical_loss": 3.4424483564337245,
      "tokens_seen": 1917911040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042602020202020204,
      "loss": 2.5781,
      "theoretical_loss": 3.4424384678808586,
      "tokens_seen": 1917976576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426,
      "loss": 2.6566,
      "theoretical_loss": 3.4424285797604757,
      "tokens_seen": 1918042112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042597979797979796,
      "loss": 2.5477,
      "theoretical_loss": 3.4424186920725437,
      "tokens_seen": 1918107648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042595959595959597,
      "loss": 2.6297,
      "theoretical_loss": 3.4424088048170276,
      "tokens_seen": 1918173184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425939393939394,
      "loss": 2.5819,
      "theoretical_loss": 3.4423989179938945,
      "tokens_seen": 1918238720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042591919191919194,
      "loss": 2.8178,
      "theoretical_loss": 3.4423890316031107,
      "tokens_seen": 1918304256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258989898989899,
      "loss": 2.5821,
      "theoretical_loss": 3.442379145644642,
      "tokens_seen": 1918369792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042587878787878785,
      "loss": 2.6713,
      "theoretical_loss": 3.442369260118455,
      "tokens_seen": 1918435328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258585858585859,
      "loss": 2.6253,
      "theoretical_loss": 3.442359375024516,
      "tokens_seen": 1918500864
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042583838383838387,
      "loss": 2.7187,
      "theoretical_loss": 3.4423494903627914,
      "tokens_seen": 1918566400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042581818181818183,
      "loss": 2.6976,
      "theoretical_loss": 3.442339606133247,
      "tokens_seen": 1918631936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257979797979798,
      "loss": 2.7475,
      "theoretical_loss": 3.4423297223358507,
      "tokens_seen": 1918697472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042577777777777774,
      "loss": 2.7081,
      "theoretical_loss": 3.442319838970567,
      "tokens_seen": 1918763008
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1056802,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.889559268951416,
      "objective/train/theoretical_loss": 3.4423099560373633,
      "objective/train/tokens_used": 277687776,
      "theoretical_loss": 3.4423099560373633,
      "tokens_seen": 1918828544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257575757575758,
      "loss": 2.8355,
      "theoretical_loss": 3.4423099560373633,
      "tokens_seen": 1918828544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042573737373737377,
      "loss": 2.7127,
      "theoretical_loss": 3.442300073536206,
      "tokens_seen": 1918894080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257171717171717,
      "loss": 2.5629,
      "theoretical_loss": 3.4422901914670607,
      "tokens_seen": 1918959616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256969696969697,
      "loss": 2.5977,
      "theoretical_loss": 3.4422803098298944,
      "tokens_seen": 1919025152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042567676767676764,
      "loss": 2.4683,
      "theoretical_loss": 3.4422704286246733,
      "tokens_seen": 1919090688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256565656565657,
      "loss": 2.502,
      "theoretical_loss": 3.442260547851364,
      "tokens_seen": 1919156224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042563636363636366,
      "loss": 2.6394,
      "theoretical_loss": 3.4422506675099322,
      "tokens_seen": 1919221760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256161616161616,
      "loss": 2.5128,
      "theoretical_loss": 3.442240787600345,
      "tokens_seen": 1919287296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042559595959595957,
      "loss": 2.5185,
      "theoretical_loss": 3.4422309081225686,
      "tokens_seen": 1919352832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255757575757576,
      "loss": 2.6894,
      "theoretical_loss": 3.4422210290765687,
      "tokens_seen": 1919418368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255555555555556,
      "loss": 2.5627,
      "theoretical_loss": 3.442211150462313,
      "tokens_seen": 1919483904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042553535353535355,
      "loss": 2.8682,
      "theoretical_loss": 3.4422012722797666,
      "tokens_seen": 1919549440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255151515151515,
      "loss": 2.5446,
      "theoretical_loss": 3.442191394528897,
      "tokens_seen": 1919614976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042549494949494946,
      "loss": 2.6554,
      "theoretical_loss": 3.4421815172096695,
      "tokens_seen": 1919680512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254747474747475,
      "loss": 2.808,
      "theoretical_loss": 3.4421716403220515,
      "tokens_seen": 1919746048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254545454545455,
      "loss": 2.5093,
      "theoretical_loss": 3.4421617638660083,
      "tokens_seen": 1919811584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042543434343434344,
      "loss": 2.6735,
      "theoretical_loss": 3.442151887841508,
      "tokens_seen": 1919877120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254141414141414,
      "loss": 2.6459,
      "theoretical_loss": 3.442142012248515,
      "tokens_seen": 1919942656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253939393939394,
      "loss": 2.4776,
      "theoretical_loss": 3.442132137086997,
      "tokens_seen": 1920008192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042537373737373737,
      "loss": 2.5164,
      "theoretical_loss": 3.4421222623569205,
      "tokens_seen": 1920073728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253535353535354,
      "loss": 2.7869,
      "theoretical_loss": 3.442112388058251,
      "tokens_seen": 1920139264
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042533333333333334,
      "loss": 2.5343,
      "theoretical_loss": 3.442102514190956,
      "tokens_seen": 1920204800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253131313131313,
      "loss": 2.6066,
      "theoretical_loss": 3.4420926407550008,
      "tokens_seen": 1920270336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252929292929293,
      "loss": 2.7356,
      "theoretical_loss": 3.442082767750353,
      "tokens_seen": 1920335872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042527272727272726,
      "loss": 2.6266,
      "theoretical_loss": 3.4420728951769783,
      "tokens_seen": 1920401408
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1058271,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8692092895507812,
      "objective/train/theoretical_loss": 3.4420630230348435,
      "objective/train/tokens_used": 279326176,
      "theoretical_loss": 3.4420630230348435,
      "tokens_seen": 1920466944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042525252525252527,
      "loss": 2.789,
      "theoretical_loss": 3.4420630230348435,
      "tokens_seen": 1920466944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042523232323232323,
      "loss": 2.5816,
      "theoretical_loss": 3.4420531513239148,
      "tokens_seen": 1920532480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042521212121212124,
      "loss": 2.4545,
      "theoretical_loss": 3.4420432800441585,
      "tokens_seen": 1920598016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251919191919192,
      "loss": 2.5497,
      "theoretical_loss": 3.4420334091955413,
      "tokens_seen": 1920663552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042517171717171715,
      "loss": 2.4662,
      "theoretical_loss": 3.44202353877803,
      "tokens_seen": 1920729088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042515151515151516,
      "loss": 2.581,
      "theoretical_loss": 3.4420136687915903,
      "tokens_seen": 1920794624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251313131313131,
      "loss": 2.5163,
      "theoretical_loss": 3.442003799236189,
      "tokens_seen": 1920860160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042511111111111113,
      "loss": 2.6854,
      "theoretical_loss": 3.441993930111793,
      "tokens_seen": 1920925696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250909090909091,
      "loss": 2.6518,
      "theoretical_loss": 3.4419840614183688,
      "tokens_seen": 1920991232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042507070707070705,
      "loss": 2.7454,
      "theoretical_loss": 3.4419741931558816,
      "tokens_seen": 1921056768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042505050505050506,
      "loss": 2.4956,
      "theoretical_loss": 3.4419643253242995,
      "tokens_seen": 1921122304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042503030303030307,
      "loss": 2.7089,
      "theoretical_loss": 3.441954457923588,
      "tokens_seen": 1921187840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425010101010101,
      "loss": 2.9059,
      "theoretical_loss": 3.441944590953714,
      "tokens_seen": 1921253376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424989898989899,
      "loss": 2.7142,
      "theoretical_loss": 3.4419347244146437,
      "tokens_seen": 1921318912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424969696969697,
      "loss": 2.66,
      "theoretical_loss": 3.4419248583063435,
      "tokens_seen": 1921384448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042494949494949495,
      "loss": 2.5657,
      "theoretical_loss": 3.4419149926287806,
      "tokens_seen": 1921449984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042492929292929296,
      "loss": 2.5889,
      "theoretical_loss": 3.441905127381921,
      "tokens_seen": 1921515520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249090909090909,
      "loss": 2.8425,
      "theoretical_loss": 3.4418952625657306,
      "tokens_seen": 1921581056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248888888888889,
      "loss": 2.642,
      "theoretical_loss": 3.4418853981801774,
      "tokens_seen": 1921646592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248686868686869,
      "loss": 2.6522,
      "theoretical_loss": 3.4418755342252267,
      "tokens_seen": 1921712128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248484848484849,
      "loss": 2.6508,
      "theoretical_loss": 3.441865670700846,
      "tokens_seen": 1921777664
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042482828282828285,
      "loss": 2.7685,
      "theoretical_loss": 3.4418558076070003,
      "tokens_seen": 1921843200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248080808080808,
      "loss": 2.5812,
      "theoretical_loss": 3.4418459449436574,
      "tokens_seen": 1921908736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042478787878787877,
      "loss": 2.419,
      "theoretical_loss": 3.4418360827107835,
      "tokens_seen": 1921974272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247676767676768,
      "loss": 2.5664,
      "theoretical_loss": 3.4418262209083457,
      "tokens_seen": 1922039808
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1058853,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.736966609954834,
      "objective/train/theoretical_loss": 3.4418163595363094,
      "objective/train/tokens_used": 280964576,
      "theoretical_loss": 3.4418163595363094,
      "tokens_seen": 1922105344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247474747474748,
      "loss": 2.7834,
      "theoretical_loss": 3.4418163595363094,
      "tokens_seen": 1922105344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042472727272727275,
      "loss": 2.6645,
      "theoretical_loss": 3.4418064985946417,
      "tokens_seen": 1922170880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247070707070707,
      "loss": 2.6125,
      "theoretical_loss": 3.441796638083309,
      "tokens_seen": 1922236416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042468686868686866,
      "loss": 2.6596,
      "theoretical_loss": 3.441786778002278,
      "tokens_seen": 1922301952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246666666666667,
      "loss": 2.631,
      "theoretical_loss": 3.4417769183515157,
      "tokens_seen": 1922367488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246464646464647,
      "loss": 2.5757,
      "theoretical_loss": 3.4417670591309877,
      "tokens_seen": 1922433024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042462626262626264,
      "loss": 2.5161,
      "theoretical_loss": 3.441757200340662,
      "tokens_seen": 1922498560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246060606060606,
      "loss": 2.6606,
      "theoretical_loss": 3.4417473419805034,
      "tokens_seen": 1922564096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042458585858585855,
      "loss": 2.7274,
      "theoretical_loss": 3.4417374840504795,
      "tokens_seen": 1922629632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245656565656566,
      "loss": 2.6871,
      "theoretical_loss": 3.4417276265505565,
      "tokens_seen": 1922695168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245454545454546,
      "loss": 2.8311,
      "theoretical_loss": 3.441717769480701,
      "tokens_seen": 1922760704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042452525252525253,
      "loss": 2.5027,
      "theoretical_loss": 3.4417079128408803,
      "tokens_seen": 1922826240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245050505050505,
      "loss": 2.7313,
      "theoretical_loss": 3.4416980566310604,
      "tokens_seen": 1922891776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244848484848485,
      "loss": 2.6135,
      "theoretical_loss": 3.4416882008512073,
      "tokens_seen": 1922957312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244646464646465,
      "loss": 2.6602,
      "theoretical_loss": 3.4416783455012885,
      "tokens_seen": 1923022848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042444444444444447,
      "loss": 2.6453,
      "theoretical_loss": 3.4416684905812707,
      "tokens_seen": 1923088384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244242424242424,
      "loss": 2.57,
      "theoretical_loss": 3.4416586360911197,
      "tokens_seen": 1923153920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244040404040404,
      "loss": 2.7243,
      "theoretical_loss": 3.441648782030802,
      "tokens_seen": 1923219456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243838383838384,
      "loss": 2.8261,
      "theoretical_loss": 3.4416389284002853,
      "tokens_seen": 1923284992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243636363636364,
      "loss": 2.4296,
      "theoretical_loss": 3.4416290751995353,
      "tokens_seen": 1923350528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042434343434343436,
      "loss": 2.5943,
      "theoretical_loss": 3.4416192224285194,
      "tokens_seen": 1923416064
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243232323232323,
      "loss": 2.6942,
      "theoretical_loss": 3.4416093700872032,
      "tokens_seen": 1923481600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042430303030303033,
      "loss": 2.6438,
      "theoretical_loss": 3.441599518175554,
      "tokens_seen": 1923547136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242828282828283,
      "loss": 2.5493,
      "theoretical_loss": 3.441589666693538,
      "tokens_seen": 1923612672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242626262626263,
      "loss": 2.6233,
      "theoretical_loss": 3.4415798156411226,
      "tokens_seen": 1923678208
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1059748,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.226703643798828,
      "objective/train/theoretical_loss": 3.4415699650182736,
      "objective/train/tokens_used": 282602976,
      "theoretical_loss": 3.4415699650182736,
      "tokens_seen": 1923743744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042424242424242425,
      "loss": 2.5937,
      "theoretical_loss": 3.4415699650182736,
      "tokens_seen": 1923743744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242222222222222,
      "loss": 2.6444,
      "theoretical_loss": 3.4415601148249584,
      "tokens_seen": 1923809280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242020202020202,
      "loss": 2.6106,
      "theoretical_loss": 3.4415502650611427,
      "tokens_seen": 1923874816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241818181818182,
      "loss": 2.6631,
      "theoretical_loss": 3.441540415726794,
      "tokens_seen": 1923940352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241616161616162,
      "loss": 2.6376,
      "theoretical_loss": 3.4415305668218785,
      "tokens_seen": 1924005888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042414141414141415,
      "loss": 2.7537,
      "theoretical_loss": 3.4415207183463625,
      "tokens_seen": 1924071424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042412121212121216,
      "loss": 2.858,
      "theoretical_loss": 3.4415108703002137,
      "tokens_seen": 1924136960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241010101010101,
      "loss": 2.7537,
      "theoretical_loss": 3.441501022683398,
      "tokens_seen": 1924202496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042408080808080807,
      "loss": 2.7596,
      "theoretical_loss": 3.441491175495882,
      "tokens_seen": 1924268032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240606060606061,
      "loss": 2.5014,
      "theoretical_loss": 3.4414813287376327,
      "tokens_seen": 1924333568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042404040404040404,
      "loss": 2.6273,
      "theoretical_loss": 3.4414714824086166,
      "tokens_seen": 1924399104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042402020202020205,
      "loss": 2.4565,
      "theoretical_loss": 3.4414616365088007,
      "tokens_seen": 1924464640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424,
      "loss": 2.5366,
      "theoretical_loss": 3.4414517910381512,
      "tokens_seen": 1924530176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042397979797979796,
      "loss": 2.5527,
      "theoretical_loss": 3.441441945996635,
      "tokens_seen": 1924595712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423959595959596,
      "loss": 2.6158,
      "theoretical_loss": 3.441432101384218,
      "tokens_seen": 1924661248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042393939393939393,
      "loss": 2.5519,
      "theoretical_loss": 3.441422257200869,
      "tokens_seen": 1924726784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042391919191919194,
      "loss": 2.6452,
      "theoretical_loss": 3.4414124134465522,
      "tokens_seen": 1924792320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238989898989899,
      "loss": 2.4951,
      "theoretical_loss": 3.4414025701212365,
      "tokens_seen": 1924857856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042387878787878786,
      "loss": 2.6911,
      "theoretical_loss": 3.4413927272248865,
      "tokens_seen": 1924923392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042385858585858587,
      "loss": 2.6374,
      "theoretical_loss": 3.441382884757471,
      "tokens_seen": 1924988928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238383838383839,
      "loss": 2.995,
      "theoretical_loss": 3.4413730427189546,
      "tokens_seen": 1925054464
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042381818181818183,
      "loss": 2.9582,
      "theoretical_loss": 3.441363201109305,
      "tokens_seen": 1925120000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237979797979798,
      "loss": 2.7288,
      "theoretical_loss": 3.44135335992849,
      "tokens_seen": 1925185536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042377777777777775,
      "loss": 2.5508,
      "theoretical_loss": 3.4413435191764745,
      "tokens_seen": 1925251072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042375757575757576,
      "loss": 2.5517,
      "theoretical_loss": 3.4413336788532263,
      "tokens_seen": 1925316608
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1060184,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.955732583999634,
      "objective/train/theoretical_loss": 3.441323838958712,
      "objective/train/tokens_used": 284241376,
      "theoretical_loss": 3.441323838958712,
      "tokens_seen": 1925382144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042373737373737377,
      "loss": 2.7418,
      "theoretical_loss": 3.441323838958712,
      "tokens_seen": 1925382144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042371717171717173,
      "loss": 2.5483,
      "theoretical_loss": 3.4413139994928974,
      "tokens_seen": 1925447680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236969696969697,
      "loss": 2.646,
      "theoretical_loss": 3.4413041604557506,
      "tokens_seen": 1925513216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042367676767676764,
      "loss": 2.7451,
      "theoretical_loss": 3.441294321847238,
      "tokens_seen": 1925578752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236565656565657,
      "loss": 2.6215,
      "theoretical_loss": 3.4412844836673253,
      "tokens_seen": 1925644288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042363636363636366,
      "loss": 2.6979,
      "theoretical_loss": 3.44127464591598,
      "tokens_seen": 1925709824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236161616161616,
      "loss": 2.5959,
      "theoretical_loss": 3.4412648085931696,
      "tokens_seen": 1925775360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235959595959596,
      "loss": 2.5962,
      "theoretical_loss": 3.44125497169886,
      "tokens_seen": 1925840896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042357575757575753,
      "loss": 2.7386,
      "theoretical_loss": 3.4412451352330176,
      "tokens_seen": 1925906432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235555555555556,
      "loss": 2.6526,
      "theoretical_loss": 3.4412352991956094,
      "tokens_seen": 1925971968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042353535353535356,
      "loss": 2.552,
      "theoretical_loss": 3.441225463586603,
      "tokens_seen": 1926037504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235151515151515,
      "loss": 2.6995,
      "theoretical_loss": 3.4412156284059643,
      "tokens_seen": 1926103040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042349494949494947,
      "loss": 2.7203,
      "theoretical_loss": 3.44120579365366,
      "tokens_seen": 1926168576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234747474747475,
      "loss": 2.6077,
      "theoretical_loss": 3.441195959329658,
      "tokens_seen": 1926234112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234545454545455,
      "loss": 2.7539,
      "theoretical_loss": 3.4411861254339238,
      "tokens_seen": 1926299648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042343434343434345,
      "loss": 2.5654,
      "theoretical_loss": 3.4411762919664244,
      "tokens_seen": 1926365184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234141414141414,
      "loss": 2.6057,
      "theoretical_loss": 3.441166458927127,
      "tokens_seen": 1926430720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042339393939393936,
      "loss": 2.713,
      "theoretical_loss": 3.4411566263159985,
      "tokens_seen": 1926496256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233737373737374,
      "loss": 2.6969,
      "theoretical_loss": 3.441146794133005,
      "tokens_seen": 1926561792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233535353535354,
      "loss": 2.7568,
      "theoretical_loss": 3.4411369623781134,
      "tokens_seen": 1926627328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042333333333333334,
      "loss": 2.5438,
      "theoretical_loss": 3.4411271310512914,
      "tokens_seen": 1926692864
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233131313131313,
      "loss": 2.6664,
      "theoretical_loss": 3.441117300152505,
      "tokens_seen": 1926758400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232929292929293,
      "loss": 2.6309,
      "theoretical_loss": 3.441107469681721,
      "tokens_seen": 1926823936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232727272727273,
      "loss": 2.4625,
      "theoretical_loss": 3.441097639638907,
      "tokens_seen": 1926889472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232525252525253,
      "loss": 2.6909,
      "theoretical_loss": 3.4410878100240287,
      "tokens_seen": 1926955008
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1061402,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8198060989379883,
      "objective/train/theoretical_loss": 3.4410779808370533,
      "objective/train/tokens_used": 285879776,
      "theoretical_loss": 3.4410779808370533,
      "tokens_seen": 1927020544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042323232323232323,
      "loss": 2.7919,
      "theoretical_loss": 3.4410779808370533,
      "tokens_seen": 1927020544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232121212121212,
      "loss": 2.6017,
      "theoretical_loss": 3.441068152077948,
      "tokens_seen": 1927086080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231919191919192,
      "loss": 2.7291,
      "theoretical_loss": 3.4410583237466796,
      "tokens_seen": 1927151616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231717171717172,
      "loss": 2.5611,
      "theoretical_loss": 3.441048495843215,
      "tokens_seen": 1927217152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042315151515151517,
      "loss": 2.5936,
      "theoretical_loss": 3.4410386683675203,
      "tokens_seen": 1927282688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231313131313131,
      "loss": 2.8128,
      "theoretical_loss": 3.4410288413195627,
      "tokens_seen": 1927348224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042311111111111114,
      "loss": 2.6127,
      "theoretical_loss": 3.441019014699309,
      "tokens_seen": 1927413760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230909090909091,
      "loss": 2.662,
      "theoretical_loss": 3.4410091885067264,
      "tokens_seen": 1927479296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230707070707071,
      "loss": 2.8147,
      "theoretical_loss": 3.440999362741781,
      "tokens_seen": 1927544832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042305050505050506,
      "loss": 2.7461,
      "theoretical_loss": 3.440989537404441,
      "tokens_seen": 1927610368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423030303030303,
      "loss": 2.6908,
      "theoretical_loss": 3.4409797124946726,
      "tokens_seen": 1927675904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042301010101010103,
      "loss": 2.7072,
      "theoretical_loss": 3.4409698880124413,
      "tokens_seen": 1927741440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422989898989899,
      "loss": 2.3188,
      "theoretical_loss": 3.440960063957716,
      "tokens_seen": 1927806976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422969696969697,
      "loss": 2.7596,
      "theoretical_loss": 3.4409502403304626,
      "tokens_seen": 1927872512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042294949494949496,
      "loss": 2.6218,
      "theoretical_loss": 3.4409404171306477,
      "tokens_seen": 1927938048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042292929292929297,
      "loss": 2.5852,
      "theoretical_loss": 3.440930594358239,
      "tokens_seen": 1928003584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229090909090909,
      "loss": 2.6241,
      "theoretical_loss": 3.440920772013203,
      "tokens_seen": 1928069120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228888888888889,
      "loss": 2.7613,
      "theoretical_loss": 3.440910950095506,
      "tokens_seen": 1928134656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228686868686869,
      "loss": 2.7854,
      "theoretical_loss": 3.440901128605116,
      "tokens_seen": 1928200192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042284848484848485,
      "loss": 2.5871,
      "theoretical_loss": 3.4408913075419987,
      "tokens_seen": 1928265728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042282828282828286,
      "loss": 2.6012,
      "theoretical_loss": 3.4408814869061217,
      "tokens_seen": 1928331264
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228080808080808,
      "loss": 2.663,
      "theoretical_loss": 3.440871666697452,
      "tokens_seen": 1928396800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042278787878787877,
      "loss": 2.6573,
      "theoretical_loss": 3.440861846915956,
      "tokens_seen": 1928462336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227676767676768,
      "loss": 2.5695,
      "theoretical_loss": 3.4408520275616006,
      "tokens_seen": 1928527872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227474747474748,
      "loss": 2.8241,
      "theoretical_loss": 3.4408422086343533,
      "tokens_seen": 1928593408
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1062041,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.024728298187256,
      "objective/train/theoretical_loss": 3.4408323901341813,
      "objective/train/tokens_used": 287518176,
      "theoretical_loss": 3.4408323901341813,
      "tokens_seen": 1928658944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042272727272727275,
      "loss": 2.6987,
      "theoretical_loss": 3.4408323901341813,
      "tokens_seen": 1928658944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227070707070707,
      "loss": 2.6473,
      "theoretical_loss": 3.44082257206105,
      "tokens_seen": 1928724480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042268686868686867,
      "loss": 2.7776,
      "theoretical_loss": 3.440812754414927,
      "tokens_seen": 1928790016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226666666666667,
      "loss": 2.7868,
      "theoretical_loss": 3.4408029371957802,
      "tokens_seen": 1928855552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226464646464647,
      "loss": 2.6503,
      "theoretical_loss": 3.4407931204035753,
      "tokens_seen": 1928921088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042262626262626264,
      "loss": 2.5474,
      "theoretical_loss": 3.44078330403828,
      "tokens_seen": 1928986624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226060606060606,
      "loss": 2.6479,
      "theoretical_loss": 3.440773488099861,
      "tokens_seen": 1929052160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042258585858585856,
      "loss": 2.6409,
      "theoretical_loss": 3.4407636725882846,
      "tokens_seen": 1929117696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225656565656566,
      "loss": 2.5776,
      "theoretical_loss": 3.440753857503519,
      "tokens_seen": 1929183232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225454545454546,
      "loss": 2.6831,
      "theoretical_loss": 3.44074404284553,
      "tokens_seen": 1929248768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042252525252525254,
      "loss": 2.7039,
      "theoretical_loss": 3.440734228614285,
      "tokens_seen": 1929314304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225050505050505,
      "loss": 2.5256,
      "theoretical_loss": 3.4407244148097504,
      "tokens_seen": 1929379840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042248484848484845,
      "loss": 2.8189,
      "theoretical_loss": 3.4407146014318943,
      "tokens_seen": 1929445376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224646464646465,
      "loss": 2.7982,
      "theoretical_loss": 3.4407047884806827,
      "tokens_seen": 1929510912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224444444444445,
      "loss": 2.5438,
      "theoretical_loss": 3.4406949759560836,
      "tokens_seen": 1929576448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042242424242424243,
      "loss": 2.5654,
      "theoretical_loss": 3.4406851638580624,
      "tokens_seen": 1929641984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224040404040404,
      "loss": 2.5339,
      "theoretical_loss": 3.440675352186587,
      "tokens_seen": 1929707520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042238383838383834,
      "loss": 2.3803,
      "theoretical_loss": 3.440665540941625,
      "tokens_seen": 1929773056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223636363636364,
      "loss": 2.5308,
      "theoretical_loss": 3.4406557301231424,
      "tokens_seen": 1929838592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042234343434343437,
      "loss": 2.5711,
      "theoretical_loss": 3.4406459197311063,
      "tokens_seen": 1929904128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223232323232323,
      "loss": 2.6102,
      "theoretical_loss": 3.440636109765484,
      "tokens_seen": 1929969664
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223030303030303,
      "loss": 2.5921,
      "theoretical_loss": 3.440626300226242,
      "tokens_seen": 1930035200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222828282828283,
      "loss": 2.5174,
      "theoretical_loss": 3.440616491113348,
      "tokens_seen": 1930100736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222626262626263,
      "loss": 2.7388,
      "theoretical_loss": 3.4406066824267683,
      "tokens_seen": 1930166272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042224242424242426,
      "loss": 2.7482,
      "theoretical_loss": 3.44059687416647,
      "tokens_seen": 1930231808
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1063329,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.848512649536133,
      "objective/train/theoretical_loss": 3.440587066332421,
      "objective/train/tokens_used": 289156576,
      "theoretical_loss": 3.440587066332421,
      "tokens_seen": 1930297344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222222222222222,
      "loss": 2.5179,
      "theoretical_loss": 3.440587066332421,
      "tokens_seen": 1930297344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042220202020202017,
      "loss": 2.451,
      "theoretical_loss": 3.4405772589245873,
      "tokens_seen": 1930362880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221818181818182,
      "loss": 2.9025,
      "theoretical_loss": 3.440567451942936,
      "tokens_seen": 1930428416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221616161616162,
      "loss": 2.5348,
      "theoretical_loss": 3.440557645387435,
      "tokens_seen": 1930493952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042214141414141415,
      "loss": 2.5781,
      "theoretical_loss": 3.44054783925805,
      "tokens_seen": 1930559488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221212121212121,
      "loss": 2.7627,
      "theoretical_loss": 3.4405380335547493,
      "tokens_seen": 1930625024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221010101010101,
      "loss": 2.5685,
      "theoretical_loss": 3.4405282282774987,
      "tokens_seen": 1930690560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220808080808081,
      "loss": 2.5545,
      "theoretical_loss": 3.4405184234262656,
      "tokens_seen": 1930756096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220606060606061,
      "loss": 2.519,
      "theoretical_loss": 3.440508619001018,
      "tokens_seen": 1930821632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042204040404040404,
      "loss": 2.6168,
      "theoretical_loss": 3.440498815001722,
      "tokens_seen": 1930887168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422020202020202,
      "loss": 2.513,
      "theoretical_loss": 3.440489011428345,
      "tokens_seen": 1930952704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422,
      "loss": 2.6129,
      "theoretical_loss": 3.440479208280853,
      "tokens_seen": 1931018240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042197979797979797,
      "loss": 2.8086,
      "theoretical_loss": 3.440469405559215,
      "tokens_seen": 1931083776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000421959595959596,
      "loss": 2.7246,
      "theoretical_loss": 3.4404596032633963,
      "tokens_seen": 1931149312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042193939393939394,
      "loss": 2.5076,
      "theoretical_loss": 3.440449801393364,
      "tokens_seen": 1931214848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042191919191919195,
      "loss": 2.6279,
      "theoretical_loss": 3.440439999949087,
      "tokens_seen": 1931280384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218989898989899,
      "loss": 2.7267,
      "theoretical_loss": 3.440430198930531,
      "tokens_seen": 1931345920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042187878787878786,
      "loss": 2.6061,
      "theoretical_loss": 3.440420398337663,
      "tokens_seen": 1931411456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042185858585858587,
      "loss": 2.6032,
      "theoretical_loss": 3.44041059817045,
      "tokens_seen": 1931476992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042183838383838383,
      "loss": 2.4926,
      "theoretical_loss": 3.4404007984288594,
      "tokens_seen": 1931542528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042181818181818184,
      "loss": 2.5209,
      "theoretical_loss": 3.440390999112858,
      "tokens_seen": 1931608064
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217979797979798,
      "loss": 2.4727,
      "theoretical_loss": 3.4403812002224137,
      "tokens_seen": 1931673600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042177777777777775,
      "loss": 2.8118,
      "theoretical_loss": 3.440371401757493,
      "tokens_seen": 1931739136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042175757575757577,
      "loss": 2.6462,
      "theoretical_loss": 3.4403616037180624,
      "tokens_seen": 1931804672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217373737373738,
      "loss": 2.7645,
      "theoretical_loss": 3.44035180610409,
      "tokens_seen": 1931870208
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 1064052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8039486408233643,
      "objective/train/theoretical_loss": 3.440342008915542,
      "objective/train/tokens_used": 290794976,
      "theoretical_loss": 3.440342008915542,
      "tokens_seen": 1931935744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042171717171717173,
      "loss": 2.663,
      "theoretical_loss": 3.440342008915542,
      "tokens_seen": 1931935744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216969696969697,
      "loss": 2.5882,
      "theoretical_loss": 3.440332212152386,
      "tokens_seen": 1932001280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042167676767676765,
      "loss": 2.931,
      "theoretical_loss": 3.440322415814589,
      "tokens_seen": 1932066816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042165656565656566,
      "loss": 2.5881,
      "theoretical_loss": 3.4403126199021186,
      "tokens_seen": 1932132352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042163636363636367,
      "loss": 2.6762,
      "theoretical_loss": 3.4403028244149416,
      "tokens_seen": 1932197888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216161616161616,
      "loss": 2.4638,
      "theoretical_loss": 3.4402930293530245,
      "tokens_seen": 1932263424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215959595959596,
      "loss": 2.5095,
      "theoretical_loss": 3.440283234716335,
      "tokens_seen": 1932328960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042157575757575754,
      "loss": 2.6447,
      "theoretical_loss": 3.44027344050484,
      "tokens_seen": 1932394496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215555555555556,
      "loss": 2.6329,
      "theoretical_loss": 3.440263646718507,
      "tokens_seen": 1932460032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042153535353535356,
      "loss": 2.497,
      "theoretical_loss": 3.4402538533573024,
      "tokens_seen": 1932525568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215151515151515,
      "loss": 2.6166,
      "theoretical_loss": 3.440244060421194,
      "tokens_seen": 1932591104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214949494949495,
      "loss": 2.5755,
      "theoretical_loss": 3.440234267910149,
      "tokens_seen": 1932656640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214747474747475,
      "loss": 2.7671,
      "theoretical_loss": 3.4402244758241336,
      "tokens_seen": 1932722176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214545454545455,
      "loss": 2.7073,
      "theoretical_loss": 3.4402146841631165,
      "tokens_seen": 1932787712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042143434343434345,
      "loss": 2.6975,
      "theoretical_loss": 3.4402048929270634,
      "tokens_seen": 1932853248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214141414141414,
      "loss": 2.7931,
      "theoretical_loss": 3.4401951021159416,
      "tokens_seen": 1932918784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042139393939393937,
      "loss": 2.6141,
      "theoretical_loss": 3.4401853117297194,
      "tokens_seen": 1932984320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042137373737373743,
      "loss": 2.6676,
      "theoretical_loss": 3.4401755217683627,
      "tokens_seen": 1933049856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213535353535354,
      "loss": 2.5838,
      "theoretical_loss": 3.440165732231839,
      "tokens_seen": 1933115392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042133333333333335,
      "loss": 2.7085,
      "theoretical_loss": 3.440155943120116,
      "tokens_seen": 1933180928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213131313131313,
      "loss": 2.6783,
      "theoretical_loss": 3.440146154433161,
      "tokens_seen": 1933246464
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042129292929292926,
      "loss": 2.6153,
      "theoretical_loss": 3.4401363661709397,
      "tokens_seen": 1933312000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212727272727273,
      "loss": 2.6217,
      "theoretical_loss": 3.4401265783334205,
      "tokens_seen": 1933377536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212525252525253,
      "loss": 2.6967,
      "theoretical_loss": 3.4401167909205705,
      "tokens_seen": 1933443072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042123232323232324,
      "loss": 2.7279,
      "theoretical_loss": 3.440107003932357,
      "tokens_seen": 1933508608
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1065201,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.028951406478882,
      "objective/train/theoretical_loss": 3.440097217368746,
      "objective/train/tokens_used": 292433376,
      "theoretical_loss": 3.440097217368746,
      "tokens_seen": 1933574144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212121212121212,
      "loss": 2.831,
      "theoretical_loss": 3.440097217368746,
      "tokens_seen": 1933574144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211919191919192,
      "loss": 2.6147,
      "theoretical_loss": 3.4400874312297063,
      "tokens_seen": 1933639680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211717171717172,
      "loss": 2.6807,
      "theoretical_loss": 3.4400776455152036,
      "tokens_seen": 1933705216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211515151515152,
      "loss": 2.5114,
      "theoretical_loss": 3.4400678602252066,
      "tokens_seen": 1933770752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042113131313131313,
      "loss": 2.3262,
      "theoretical_loss": 3.4400580753596817,
      "tokens_seen": 1933836288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211111111111111,
      "loss": 2.5457,
      "theoretical_loss": 3.440048290918596,
      "tokens_seen": 1933901824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210909090909091,
      "loss": 2.5826,
      "theoretical_loss": 3.4400385069019164,
      "tokens_seen": 1933967360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210707070707071,
      "loss": 2.4071,
      "theoretical_loss": 3.4400287233096107,
      "tokens_seen": 1934032896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042105050505050507,
      "loss": 2.6196,
      "theoretical_loss": 3.4400189401416466,
      "tokens_seen": 1934098432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000421030303030303,
      "loss": 2.5962,
      "theoretical_loss": 3.4400091573979896,
      "tokens_seen": 1934163968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042101010101010104,
      "loss": 2.669,
      "theoretical_loss": 3.439999375078609,
      "tokens_seen": 1934229504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420989898989899,
      "loss": 2.4876,
      "theoretical_loss": 3.439989593183471,
      "tokens_seen": 1934295040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420969696969697,
      "loss": 2.4545,
      "theoretical_loss": 3.4399798117125426,
      "tokens_seen": 1934360576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042094949494949496,
      "loss": 2.6715,
      "theoretical_loss": 3.439970030665791,
      "tokens_seen": 1934426112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209292929292929,
      "loss": 2.3414,
      "theoretical_loss": 3.439960250043184,
      "tokens_seen": 1934491648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042090909090909093,
      "loss": 2.5181,
      "theoretical_loss": 3.4399504698446886,
      "tokens_seen": 1934557184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208888888888889,
      "loss": 2.4947,
      "theoretical_loss": 3.439940690070272,
      "tokens_seen": 1934622720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208686868686869,
      "loss": 2.554,
      "theoretical_loss": 3.4399309107199016,
      "tokens_seen": 1934688256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042084848484848485,
      "loss": 2.7057,
      "theoretical_loss": 3.439921131793544,
      "tokens_seen": 1934753792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042082828282828286,
      "loss": 2.5841,
      "theoretical_loss": 3.439911353291168,
      "tokens_seen": 1934819328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208080808080808,
      "loss": 2.5806,
      "theoretical_loss": 3.4399015752127386,
      "tokens_seen": 1934884864
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207878787878788,
      "loss": 2.7166,
      "theoretical_loss": 3.439891797558225,
      "tokens_seen": 1934950400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207676767676768,
      "loss": 2.9193,
      "theoretical_loss": 3.4398820203275937,
      "tokens_seen": 1935015936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042074747474747475,
      "loss": 2.5482,
      "theoretical_loss": 3.439872243520812,
      "tokens_seen": 1935081472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042072727272727276,
      "loss": 2.641,
      "theoretical_loss": 3.4398624671378473,
      "tokens_seen": 1935147008
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1066155,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.908733367919922,
      "objective/train/theoretical_loss": 3.4398526911786664,
      "objective/train/tokens_used": 294071776,
      "theoretical_loss": 3.4398526911786664,
      "tokens_seen": 1935212544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207070707070707,
      "loss": 2.8247,
      "theoretical_loss": 3.4398526911786664,
      "tokens_seen": 1935212544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042068686868686867,
      "loss": 2.7459,
      "theoretical_loss": 3.4398429156432373,
      "tokens_seen": 1935278080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206666666666667,
      "loss": 2.5958,
      "theoretical_loss": 3.439833140531527,
      "tokens_seen": 1935343616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042064646464646464,
      "loss": 2.6365,
      "theoretical_loss": 3.4398233658435027,
      "tokens_seen": 1935409152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042062626262626265,
      "loss": 2.5277,
      "theoretical_loss": 3.4398135915791315,
      "tokens_seen": 1935474688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206060606060606,
      "loss": 2.5039,
      "theoretical_loss": 3.4398038177383805,
      "tokens_seen": 1935540224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042058585858585856,
      "loss": 2.6533,
      "theoretical_loss": 3.439794044321218,
      "tokens_seen": 1935605760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205656565656566,
      "loss": 2.7235,
      "theoretical_loss": 3.4397842713276106,
      "tokens_seen": 1935671296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205454545454546,
      "loss": 2.5335,
      "theoretical_loss": 3.439774498757526,
      "tokens_seen": 1935736832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042052525252525254,
      "loss": 2.4197,
      "theoretical_loss": 3.439764726610931,
      "tokens_seen": 1935802368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205050505050505,
      "loss": 2.6907,
      "theoretical_loss": 3.4397549548877926,
      "tokens_seen": 1935867904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042048484848484846,
      "loss": 2.481,
      "theoretical_loss": 3.4397451835880792,
      "tokens_seen": 1935933440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042046464646464647,
      "loss": 2.4428,
      "theoretical_loss": 3.439735412711757,
      "tokens_seen": 1935998976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204444444444445,
      "loss": 2.7433,
      "theoretical_loss": 3.4397256422587947,
      "tokens_seen": 1936064512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042042424242424244,
      "loss": 2.5442,
      "theoretical_loss": 3.4397158722291583,
      "tokens_seen": 1936130048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204040404040404,
      "loss": 2.4129,
      "theoretical_loss": 3.4397061026228157,
      "tokens_seen": 1936195584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042038383838383835,
      "loss": 2.5154,
      "theoretical_loss": 3.439696333439734,
      "tokens_seen": 1936261120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203636363636364,
      "loss": 2.4758,
      "theoretical_loss": 3.439686564679881,
      "tokens_seen": 1936326656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042034343434343437,
      "loss": 2.7579,
      "theoretical_loss": 3.4396767963432238,
      "tokens_seen": 1936392192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042032323232323233,
      "loss": 2.5475,
      "theoretical_loss": 3.4396670284297297,
      "tokens_seen": 1936457728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203030303030303,
      "loss": 2.5622,
      "theoretical_loss": 3.4396572609393656,
      "tokens_seen": 1936523264
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042028282828282824,
      "loss": 2.6701,
      "theoretical_loss": 3.4396474938720996,
      "tokens_seen": 1936588800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202626262626263,
      "loss": 2.8353,
      "theoretical_loss": 3.439637727227899,
      "tokens_seen": 1936654336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042024242424242426,
      "loss": 2.7227,
      "theoretical_loss": 3.439627961006731,
      "tokens_seen": 1936719872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202222222222222,
      "loss": 2.5612,
      "theoretical_loss": 3.4396181952085625,
      "tokens_seen": 1936785408
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1066685,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.676300048828125,
      "objective/train/theoretical_loss": 3.4396084298333616,
      "objective/train/tokens_used": 295710176,
      "theoretical_loss": 3.4396084298333616,
      "tokens_seen": 1936850944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202020202020202,
      "loss": 2.6837,
      "theoretical_loss": 3.4396084298333616,
      "tokens_seen": 1936850944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201818181818182,
      "loss": 2.4526,
      "theoretical_loss": 3.4395986648810943,
      "tokens_seen": 1936916480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201616161616162,
      "loss": 2.698,
      "theoretical_loss": 3.43958890035173,
      "tokens_seen": 1936982016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042014141414141416,
      "loss": 2.6901,
      "theoretical_loss": 3.439579136245235,
      "tokens_seen": 1937047552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201212121212121,
      "loss": 2.6405,
      "theoretical_loss": 3.439569372561577,
      "tokens_seen": 1937113088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042010101010101007,
      "loss": 2.7245,
      "theoretical_loss": 3.439559609300722,
      "tokens_seen": 1937178624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200808080808081,
      "loss": 2.7824,
      "theoretical_loss": 3.439549846462639,
      "tokens_seen": 1937244160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200606060606061,
      "loss": 2.3995,
      "theoretical_loss": 3.4395400840472954,
      "tokens_seen": 1937309696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042004040404040405,
      "loss": 2.558,
      "theoretical_loss": 3.439530322054658,
      "tokens_seen": 1937375232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420020202020202,
      "loss": 2.6701,
      "theoretical_loss": 3.439520560484694,
      "tokens_seen": 1937440768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042,
      "loss": 2.4783,
      "theoretical_loss": 3.4395107993373712,
      "tokens_seen": 1937506304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419979797979798,
      "loss": 2.6693,
      "theoretical_loss": 3.439501038612657,
      "tokens_seen": 1937571840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419959595959596,
      "loss": 2.5367,
      "theoretical_loss": 3.4394912783105185,
      "tokens_seen": 1937637376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041993939393939394,
      "loss": 2.6768,
      "theoretical_loss": 3.4394815184309238,
      "tokens_seen": 1937702912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199191919191919,
      "loss": 2.794,
      "theoretical_loss": 3.4394717589738395,
      "tokens_seen": 1937768448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198989898989899,
      "loss": 2.5629,
      "theoretical_loss": 3.4394619999392333,
      "tokens_seen": 1937833984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041987878787878787,
      "loss": 2.8272,
      "theoretical_loss": 3.4394522413270727,
      "tokens_seen": 1937899520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198585858585859,
      "loss": 2.8337,
      "theoretical_loss": 3.4394424831373254,
      "tokens_seen": 1937965056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041983838383838383,
      "loss": 2.5516,
      "theoretical_loss": 3.439432725369958,
      "tokens_seen": 1938030592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041981818181818185,
      "loss": 2.689,
      "theoretical_loss": 3.439422968024939,
      "tokens_seen": 1938096128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197979797979798,
      "loss": 2.6582,
      "theoretical_loss": 3.4394132111022353,
      "tokens_seen": 1938161664
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197777777777778,
      "loss": 2.449,
      "theoretical_loss": 3.4394034546018144,
      "tokens_seen": 1938227200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197777777777778,
      "loss": 2.9064,
      "theoretical_loss": 3.4393936985236433,
      "tokens_seen": 1938292736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041975757575757577,
      "loss": 2.8677,
      "theoretical_loss": 3.439383942867691,
      "tokens_seen": 1938358272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041973737373737373,
      "loss": 2.6937,
      "theoretical_loss": 3.4393741876339226,
      "tokens_seen": 1938423808
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1067842,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.121251344680786,
      "objective/train/theoretical_loss": 3.439364432822307,
      "objective/train/tokens_used": 297348576,
      "theoretical_loss": 3.439364432822307,
      "tokens_seen": 1938489344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041971717171717174,
      "loss": 2.7767,
      "theoretical_loss": 3.439364432822307,
      "tokens_seen": 1938489344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196969696969697,
      "loss": 2.7315,
      "theoretical_loss": 3.439354678432812,
      "tokens_seen": 1938554880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196767676767677,
      "loss": 2.7844,
      "theoretical_loss": 3.439344924465404,
      "tokens_seen": 1938620416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041965656565656566,
      "loss": 2.6795,
      "theoretical_loss": 3.4393351709200513,
      "tokens_seen": 1938685952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196363636363637,
      "loss": 2.6743,
      "theoretical_loss": 3.439325417796721,
      "tokens_seen": 1938751488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041961616161616163,
      "loss": 2.605,
      "theoretical_loss": 3.4393156650953802,
      "tokens_seen": 1938817024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195959595959596,
      "loss": 2.6641,
      "theoretical_loss": 3.439305912815997,
      "tokens_seen": 1938882560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195757575757576,
      "loss": 2.8231,
      "theoretical_loss": 3.4392961609585386,
      "tokens_seen": 1938948096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041955555555555556,
      "loss": 2.7356,
      "theoretical_loss": 3.439286409522973,
      "tokens_seen": 1939013632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041953535353535357,
      "loss": 2.6914,
      "theoretical_loss": 3.439276658509267,
      "tokens_seen": 1939079168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195151515151515,
      "loss": 2.6307,
      "theoretical_loss": 3.4392669079173883,
      "tokens_seen": 1939144704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194949494949495,
      "loss": 2.339,
      "theoretical_loss": 3.4392571577473046,
      "tokens_seen": 1939210240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194747474747475,
      "loss": 2.614,
      "theoretical_loss": 3.439247407998983,
      "tokens_seen": 1939275776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194545454545455,
      "loss": 2.5329,
      "theoretical_loss": 3.4392376586723916,
      "tokens_seen": 1939341312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041943434343434346,
      "loss": 2.789,
      "theoretical_loss": 3.439227909767497,
      "tokens_seen": 1939406848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194141414141414,
      "loss": 2.7102,
      "theoretical_loss": 3.439218161284268,
      "tokens_seen": 1939472384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193939393939394,
      "loss": 2.7777,
      "theoretical_loss": 3.439208413222671,
      "tokens_seen": 1939537920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193737373737374,
      "loss": 2.593,
      "theoretical_loss": 3.4391986655826736,
      "tokens_seen": 1939603456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193535353535354,
      "loss": 2.662,
      "theoretical_loss": 3.439188918364244,
      "tokens_seen": 1939668992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041933333333333335,
      "loss": 2.8318,
      "theoretical_loss": 3.439179171567349,
      "tokens_seen": 1939734528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193131313131313,
      "loss": 2.8346,
      "theoretical_loss": 3.439169425191957,
      "tokens_seen": 1939800064
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041929292929292927,
      "loss": 2.4422,
      "theoretical_loss": 3.439159679238035,
      "tokens_seen": 1939865600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041927272727272733,
      "loss": 2.5032,
      "theoretical_loss": 3.4391499337055498,
      "tokens_seen": 1939931136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192525252525253,
      "loss": 2.6996,
      "theoretical_loss": 3.4391401885944703,
      "tokens_seen": 1939996672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041923232323232325,
      "loss": 2.7708,
      "theoretical_loss": 3.439130443904763,
      "tokens_seen": 1940062208
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1068281,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4002110958099365,
      "objective/train/theoretical_loss": 3.4391206996363968,
      "objective/train/tokens_used": 298986976,
      "theoretical_loss": 3.4391206996363968,
      "tokens_seen": 1940127744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192121212121212,
      "loss": 2.7074,
      "theoretical_loss": 3.4391206996363968,
      "tokens_seen": 1940127744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041919191919191916,
      "loss": 2.7498,
      "theoretical_loss": 3.439110955789338,
      "tokens_seen": 1940193280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191717171717172,
      "loss": 2.583,
      "theoretical_loss": 3.4391012123635534,
      "tokens_seen": 1940258816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191515151515152,
      "loss": 2.6829,
      "theoretical_loss": 3.4390914693590124,
      "tokens_seen": 1940324352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041913131313131314,
      "loss": 2.6073,
      "theoretical_loss": 3.439081726775682,
      "tokens_seen": 1940389888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191111111111111,
      "loss": 2.8032,
      "theoretical_loss": 3.4390719846135296,
      "tokens_seen": 1940455424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041909090909090905,
      "loss": 2.6993,
      "theoretical_loss": 3.4390622428725224,
      "tokens_seen": 1940520960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190707070707071,
      "loss": 2.5315,
      "theoretical_loss": 3.439052501552628,
      "tokens_seen": 1940586496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190505050505051,
      "loss": 2.6999,
      "theoretical_loss": 3.4390427606538148,
      "tokens_seen": 1940652032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041903030303030303,
      "loss": 2.6889,
      "theoretical_loss": 3.4390330201760495,
      "tokens_seen": 1940717568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419010101010101,
      "loss": 2.6122,
      "theoretical_loss": 3.439023280119301,
      "tokens_seen": 1940783104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418989898989899,
      "loss": 2.7696,
      "theoretical_loss": 3.439013540483535,
      "tokens_seen": 1940848640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418969696969697,
      "loss": 2.644,
      "theoretical_loss": 3.4390038012687203,
      "tokens_seen": 1940914176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041894949494949497,
      "loss": 2.4996,
      "theoretical_loss": 3.4389940624748236,
      "tokens_seen": 1940979712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189292929292929,
      "loss": 2.6993,
      "theoretical_loss": 3.438984324101814,
      "tokens_seen": 1941045248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189090909090909,
      "loss": 2.4784,
      "theoretical_loss": 3.4389745861496577,
      "tokens_seen": 1941110784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188888888888889,
      "loss": 2.6716,
      "theoretical_loss": 3.438964848618323,
      "tokens_seen": 1941176320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188686868686869,
      "loss": 2.7038,
      "theoretical_loss": 3.438955111507777,
      "tokens_seen": 1941241856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041884848484848486,
      "loss": 2.4977,
      "theoretical_loss": 3.438945374817988,
      "tokens_seen": 1941307392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188282828282828,
      "loss": 2.8805,
      "theoretical_loss": 3.4389356385489234,
      "tokens_seen": 1941372928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041880808080808083,
      "loss": 2.6442,
      "theoretical_loss": 3.43892590270055,
      "tokens_seen": 1941438464
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187878787878788,
      "loss": 2.3895,
      "theoretical_loss": 3.4389161672728368,
      "tokens_seen": 1941504000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187676767676768,
      "loss": 2.9731,
      "theoretical_loss": 3.4389064322657505,
      "tokens_seen": 1941569536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041874747474747475,
      "loss": 2.5621,
      "theoretical_loss": 3.438896697679259,
      "tokens_seen": 1941635072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187272727272727,
      "loss": 2.6273,
      "theoretical_loss": 3.4388869635133297,
      "tokens_seen": 1941700608
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1069559,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2460758686065674,
      "objective/train/theoretical_loss": 3.4388772297679306,
      "objective/train/tokens_used": 300625376,
      "theoretical_loss": 3.4388772297679306,
      "tokens_seen": 1941766144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187070707070707,
      "loss": 2.6064,
      "theoretical_loss": 3.4388772297679306,
      "tokens_seen": 1941766144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186868686868687,
      "loss": 2.589,
      "theoretical_loss": 3.438867496443029,
      "tokens_seen": 1941831680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186666666666667,
      "loss": 2.6783,
      "theoretical_loss": 3.438857763538593,
      "tokens_seen": 1941897216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041864646464646464,
      "loss": 2.7044,
      "theoretical_loss": 3.4388480310545892,
      "tokens_seen": 1941962752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041862626262626266,
      "loss": 2.364,
      "theoretical_loss": 3.4388382989909863,
      "tokens_seen": 1942028288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186060606060606,
      "loss": 2.6038,
      "theoretical_loss": 3.438828567347752,
      "tokens_seen": 1942093824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041858585858585857,
      "loss": 2.5249,
      "theoretical_loss": 3.4388188361248533,
      "tokens_seen": 1942159360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185656565656566,
      "loss": 2.7292,
      "theoretical_loss": 3.4388091053222585,
      "tokens_seen": 1942224896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041854545454545454,
      "loss": 2.9845,
      "theoretical_loss": 3.4387993749399346,
      "tokens_seen": 1942290432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041852525252525255,
      "loss": 2.6296,
      "theoretical_loss": 3.4387896449778497,
      "tokens_seen": 1942355968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185050505050505,
      "loss": 2.507,
      "theoretical_loss": 3.4387799154359717,
      "tokens_seen": 1942421504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041848484848484846,
      "loss": 2.5466,
      "theoretical_loss": 3.4387701863142675,
      "tokens_seen": 1942487040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184646464646465,
      "loss": 2.6209,
      "theoretical_loss": 3.4387604576127053,
      "tokens_seen": 1942552576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184444444444445,
      "loss": 2.6257,
      "theoretical_loss": 3.438750729331253,
      "tokens_seen": 1942618112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041842424242424244,
      "loss": 2.5931,
      "theoretical_loss": 3.438741001469878,
      "tokens_seen": 1942683648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184040404040404,
      "loss": 2.6329,
      "theoretical_loss": 3.438731274028547,
      "tokens_seen": 1942749184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041838383838383835,
      "loss": 2.6169,
      "theoretical_loss": 3.43872154700723,
      "tokens_seen": 1942814720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041836363636363637,
      "loss": 2.7174,
      "theoretical_loss": 3.438711820405893,
      "tokens_seen": 1942880256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183434343434344,
      "loss": 2.7735,
      "theoretical_loss": 3.438702094224504,
      "tokens_seen": 1942945792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041832323232323233,
      "loss": 2.5786,
      "theoretical_loss": 3.4386923684630304,
      "tokens_seen": 1943011328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183030303030303,
      "loss": 2.4048,
      "theoretical_loss": 3.438682643121441,
      "tokens_seen": 1943076864
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041828282828282825,
      "loss": 2.4826,
      "theoretical_loss": 3.4386729181997024,
      "tokens_seen": 1943142400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182626262626263,
      "loss": 2.6934,
      "theoretical_loss": 3.4386631936977827,
      "tokens_seen": 1943207936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041824242424242427,
      "loss": 2.691,
      "theoretical_loss": 3.43865346961565,
      "tokens_seen": 1943273472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182222222222222,
      "loss": 2.6558,
      "theoretical_loss": 3.4386437459532715,
      "tokens_seen": 1943339008
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1070184,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5214946269989014,
      "objective/train/theoretical_loss": 3.4386340227106147,
      "objective/train/tokens_used": 302263776,
      "theoretical_loss": 3.4386340227106147,
      "tokens_seen": 1943404544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182020202020202,
      "loss": 2.5596,
      "theoretical_loss": 3.4386340227106147,
      "tokens_seen": 1943404544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041818181818181814,
      "loss": 2.5195,
      "theoretical_loss": 3.438624299887648,
      "tokens_seen": 1943470080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181616161616162,
      "loss": 2.7051,
      "theoretical_loss": 3.438614577484339,
      "tokens_seen": 1943535616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041814141414141416,
      "loss": 2.7595,
      "theoretical_loss": 3.438604855500655,
      "tokens_seen": 1943601152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181212121212121,
      "loss": 2.9422,
      "theoretical_loss": 3.4385951339365644,
      "tokens_seen": 1943666688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181010101010101,
      "loss": 2.6183,
      "theoretical_loss": 3.438585412792034,
      "tokens_seen": 1943732224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041808080808080814,
      "loss": 2.8163,
      "theoretical_loss": 3.4385756920670327,
      "tokens_seen": 1943797760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180606060606061,
      "loss": 2.7602,
      "theoretical_loss": 3.438565971761528,
      "tokens_seen": 1943863296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041804040404040406,
      "loss": 2.8253,
      "theoretical_loss": 3.4385562518754864,
      "tokens_seen": 1943928832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418020202020202,
      "loss": 2.3846,
      "theoretical_loss": 3.4385465324088766,
      "tokens_seen": 1943994368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041799999999999997,
      "loss": 2.5301,
      "theoretical_loss": 3.438536813361667,
      "tokens_seen": 1944059904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041797979797979803,
      "loss": 2.6371,
      "theoretical_loss": 3.4385270947338245,
      "tokens_seen": 1944125440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417959595959596,
      "loss": 2.5773,
      "theoretical_loss": 3.4385173765253167,
      "tokens_seen": 1944190976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041793939393939395,
      "loss": 2.4892,
      "theoretical_loss": 3.4385076587361123,
      "tokens_seen": 1944256512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179191919191919,
      "loss": 2.7892,
      "theoretical_loss": 3.4384979413661787,
      "tokens_seen": 1944322048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178989898989899,
      "loss": 2.5332,
      "theoretical_loss": 3.438488224415483,
      "tokens_seen": 1944387584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178787878787879,
      "loss": 2.636,
      "theoretical_loss": 3.4384785078839935,
      "tokens_seen": 1944453120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178585858585859,
      "loss": 2.6614,
      "theoretical_loss": 3.438468791771678,
      "tokens_seen": 1944518656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041783838383838384,
      "loss": 2.7253,
      "theoretical_loss": 3.4384590760785043,
      "tokens_seen": 1944584192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178181818181818,
      "loss": 2.7251,
      "theoretical_loss": 3.4384493608044404,
      "tokens_seen": 1944649728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177979797979798,
      "loss": 2.6503,
      "theoretical_loss": 3.4384396459494537,
      "tokens_seen": 1944715264
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177777777777778,
      "loss": 2.4993,
      "theoretical_loss": 3.4384299315135123,
      "tokens_seen": 1944780800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177575757575758,
      "loss": 2.7246,
      "theoretical_loss": 3.438420217496584,
      "tokens_seen": 1944846336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041773737373737373,
      "loss": 2.3471,
      "theoretical_loss": 3.438410503898636,
      "tokens_seen": 1944911872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041771717171717174,
      "loss": 2.6327,
      "theoretical_loss": 3.438400790719637,
      "tokens_seen": 1944977408
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1071164,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9253249168395996,
      "objective/train/theoretical_loss": 3.4383910779595537,
      "objective/train/tokens_used": 303902176,
      "theoretical_loss": 3.4383910779595537,
      "tokens_seen": 1945042944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176969696969697,
      "loss": 2.591,
      "theoretical_loss": 3.4383910779595537,
      "tokens_seen": 1945042944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176767676767677,
      "loss": 2.5417,
      "theoretical_loss": 3.4383813656183557,
      "tokens_seen": 1945108480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041765656565656567,
      "loss": 2.5524,
      "theoretical_loss": 3.438371653696009,
      "tokens_seen": 1945174016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176363636363636,
      "loss": 2.686,
      "theoretical_loss": 3.4383619421924823,
      "tokens_seen": 1945239552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041761616161616164,
      "loss": 2.652,
      "theoretical_loss": 3.4383522311077432,
      "tokens_seen": 1945305088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175959595959596,
      "loss": 2.5249,
      "theoretical_loss": 3.43834252044176,
      "tokens_seen": 1945370624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175757575757576,
      "loss": 2.7325,
      "theoretical_loss": 3.4383328101945,
      "tokens_seen": 1945436160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041755555555555556,
      "loss": 2.6777,
      "theoretical_loss": 3.438323100365931,
      "tokens_seen": 1945501696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041753535353535357,
      "loss": 2.8154,
      "theoretical_loss": 3.4383133909560213,
      "tokens_seen": 1945567232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041751515151515153,
      "loss": 2.6971,
      "theoretical_loss": 3.4383036819647383,
      "tokens_seen": 1945632768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174949494949495,
      "loss": 2.6609,
      "theoretical_loss": 3.4382939733920503,
      "tokens_seen": 1945698304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174747474747475,
      "loss": 2.6654,
      "theoretical_loss": 3.4382842652379244,
      "tokens_seen": 1945763840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041745454545454545,
      "loss": 2.6757,
      "theoretical_loss": 3.4382745575023295,
      "tokens_seen": 1945829376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041743434343434347,
      "loss": 2.8648,
      "theoretical_loss": 3.4382648501852326,
      "tokens_seen": 1945894912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174141414141414,
      "loss": 2.7462,
      "theoretical_loss": 3.4382551432866024,
      "tokens_seen": 1945960448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173939393939394,
      "loss": 2.5422,
      "theoretical_loss": 3.438245436806406,
      "tokens_seen": 1946025984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173737373737374,
      "loss": 2.5824,
      "theoretical_loss": 3.438235730744611,
      "tokens_seen": 1946091520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041735353535353535,
      "loss": 2.5466,
      "theoretical_loss": 3.4382260251011862,
      "tokens_seen": 1946157056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041733333333333336,
      "loss": 2.5127,
      "theoretical_loss": 3.4382163198760995,
      "tokens_seen": 1946222592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173131313131313,
      "loss": 2.6517,
      "theoretical_loss": 3.4382066150693182,
      "tokens_seen": 1946288128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041729292929292927,
      "loss": 2.6034,
      "theoretical_loss": 3.43819691068081,
      "tokens_seen": 1946353664
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172727272727273,
      "loss": 2.6726,
      "theoretical_loss": 3.438187206710543,
      "tokens_seen": 1946419200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172525252525253,
      "loss": 2.6317,
      "theoretical_loss": 3.4381775031584856,
      "tokens_seen": 1946484736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041723232323232325,
      "loss": 2.7176,
      "theoretical_loss": 3.4381678000246056,
      "tokens_seen": 1946550272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172121212121212,
      "loss": 2.5587,
      "theoretical_loss": 3.43815809730887,
      "tokens_seen": 1946615808
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1071829,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5584614276885986,
      "objective/train/theoretical_loss": 3.4381483950112477,
      "objective/train/tokens_used": 305540576,
      "theoretical_loss": 3.4381483950112477,
      "tokens_seen": 1946681344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041719191919191916,
      "loss": 2.5635,
      "theoretical_loss": 3.4381483950112477,
      "tokens_seen": 1946681344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171717171717172,
      "loss": 2.6877,
      "theoretical_loss": 3.438138693131706,
      "tokens_seen": 1946746880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171515151515152,
      "loss": 2.7448,
      "theoretical_loss": 3.4381289916702134,
      "tokens_seen": 1946812416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041713131313131314,
      "loss": 2.5441,
      "theoretical_loss": 3.4381192906267373,
      "tokens_seen": 1946877952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171111111111111,
      "loss": 2.8343,
      "theoretical_loss": 3.438109590001246,
      "tokens_seen": 1946943488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041709090909090906,
      "loss": 2.7814,
      "theoretical_loss": 3.4380998897937065,
      "tokens_seen": 1947009024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170707070707071,
      "loss": 2.5301,
      "theoretical_loss": 3.438090190004088,
      "tokens_seen": 1947074560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170505050505051,
      "loss": 2.6603,
      "theoretical_loss": 3.438080490632358,
      "tokens_seen": 1947140096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041703030303030304,
      "loss": 2.4052,
      "theoretical_loss": 3.438070791678484,
      "tokens_seen": 1947205632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417010101010101,
      "loss": 2.5384,
      "theoretical_loss": 3.4380610931424345,
      "tokens_seen": 1947271168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041698989898989895,
      "loss": 2.5251,
      "theoretical_loss": 3.4380513950241767,
      "tokens_seen": 1947336704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000416969696969697,
      "loss": 2.754,
      "theoretical_loss": 3.438041697323679,
      "tokens_seen": 1947402240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041694949494949497,
      "loss": 2.5165,
      "theoretical_loss": 3.43803200004091,
      "tokens_seen": 1947467776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041692929292929293,
      "loss": 2.6118,
      "theoretical_loss": 3.4380223031758366,
      "tokens_seen": 1947533312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169090909090909,
      "loss": 2.6706,
      "theoretical_loss": 3.438012606728427,
      "tokens_seen": 1947598848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168888888888889,
      "loss": 2.5127,
      "theoretical_loss": 3.43800291069865,
      "tokens_seen": 1947664384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168686868686869,
      "loss": 2.6097,
      "theoretical_loss": 3.4379932150864723,
      "tokens_seen": 1947729920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041684848484848486,
      "loss": 2.7995,
      "theoretical_loss": 3.4379835198918625,
      "tokens_seen": 1947795456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168282828282828,
      "loss": 2.5767,
      "theoretical_loss": 3.4379738251147884,
      "tokens_seen": 1947860992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168080808080808,
      "loss": 2.7866,
      "theoretical_loss": 3.4379641307552182,
      "tokens_seen": 1947926528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167878787878788,
      "loss": 2.3767,
      "theoretical_loss": 3.43795443681312,
      "tokens_seen": 1947992064
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167676767676768,
      "loss": 2.6098,
      "theoretical_loss": 3.4379447432884613,
      "tokens_seen": 1948057600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041674747474747476,
      "loss": 2.5764,
      "theoretical_loss": 3.4379350501812107,
      "tokens_seen": 1948123136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167272727272727,
      "loss": 2.7819,
      "theoretical_loss": 3.437925357491335,
      "tokens_seen": 1948188672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167070707070707,
      "loss": 2.6237,
      "theoretical_loss": 3.437915665218804,
      "tokens_seen": 1948254208
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 1073161,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.831909656524658,
      "objective/train/theoretical_loss": 3.437905973363584,
      "objective/train/tokens_used": 307178976,
      "theoretical_loss": 3.437905973363584,
      "tokens_seen": 1948319744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004166868686868687,
      "loss": 2.6736,
      "theoretical_loss": 3.437905973363584,
      "tokens_seen": 1948319744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166666666666667,
      "loss": 2.6268,
      "theoretical_loss": 3.437896281925644,
      "tokens_seen": 1948385280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041664646464646465,
      "loss": 2.8293,
      "theoretical_loss": 3.4378865909049514,
      "tokens_seen": 1948450816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166262626262626,
      "loss": 2.5125,
      "theoretical_loss": 3.4378769003014744,
      "tokens_seen": 1948516352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166060606060606,
      "loss": 2.7312,
      "theoretical_loss": 3.437867210115181,
      "tokens_seen": 1948581888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165858585858586,
      "loss": 2.5568,
      "theoretical_loss": 3.43785752034604,
      "tokens_seen": 1948647424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165656565656566,
      "loss": 2.9045,
      "theoretical_loss": 3.437847830994018,
      "tokens_seen": 1948712960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041654545454545454,
      "loss": 2.6776,
      "theoretical_loss": 3.4378381420590842,
      "tokens_seen": 1948778496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041652525252525255,
      "loss": 2.5809,
      "theoretical_loss": 3.437828453541206,
      "tokens_seen": 1948844032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165050505050505,
      "loss": 2.5214,
      "theoretical_loss": 3.4378187654403516,
      "tokens_seen": 1948909568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041648484848484847,
      "loss": 2.5708,
      "theoretical_loss": 3.437809077756489,
      "tokens_seen": 1948975104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164646464646465,
      "loss": 2.7088,
      "theoretical_loss": 3.437799390489586,
      "tokens_seen": 1949040640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041644444444444444,
      "loss": 2.5165,
      "theoretical_loss": 3.437789703639611,
      "tokens_seen": 1949106176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041642424242424245,
      "loss": 2.6073,
      "theoretical_loss": 3.4377800172065323,
      "tokens_seen": 1949171712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164040404040404,
      "loss": 2.5349,
      "theoretical_loss": 3.4377703311903174,
      "tokens_seen": 1949237248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041638383838383836,
      "loss": 2.5828,
      "theoretical_loss": 3.437760645590934,
      "tokens_seen": 1949302784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041636363636363637,
      "loss": 2.6568,
      "theoretical_loss": 3.437750960408351,
      "tokens_seen": 1949368320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163434343434344,
      "loss": 2.7027,
      "theoretical_loss": 3.4377412756425363,
      "tokens_seen": 1949433856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041632323232323234,
      "loss": 2.5977,
      "theoretical_loss": 3.437731591293457,
      "tokens_seen": 1949499392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163030303030303,
      "loss": 2.5681,
      "theoretical_loss": 3.4377219073610825,
      "tokens_seen": 1949564928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162828282828283,
      "loss": 2.4887,
      "theoretical_loss": 3.4377122238453803,
      "tokens_seen": 1949630464
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041626262626262626,
      "loss": 2.7247,
      "theoretical_loss": 3.437702540746318,
      "tokens_seen": 1949696000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162424242424243,
      "loss": 2.6263,
      "theoretical_loss": 3.4376928580638646,
      "tokens_seen": 1949761536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041622222222222223,
      "loss": 2.7817,
      "theoretical_loss": 3.4376831757979875,
      "tokens_seen": 1949827072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162020202020202,
      "loss": 2.6438,
      "theoretical_loss": 3.437673493948655,
      "tokens_seen": 1949892608
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1073760,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.301966428756714,
      "objective/train/theoretical_loss": 3.437663812515835,
      "objective/train/tokens_used": 308817376,
      "theoretical_loss": 3.437663812515835,
      "tokens_seen": 1949958144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161818181818182,
      "loss": 2.5007,
      "theoretical_loss": 3.437663812515835,
      "tokens_seen": 1949958144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161616161616162,
      "loss": 2.8909,
      "theoretical_loss": 3.437654131499496,
      "tokens_seen": 1950023680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041614141414141417,
      "loss": 2.6493,
      "theoretical_loss": 3.4376444508996054,
      "tokens_seen": 1950089216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161212121212121,
      "loss": 2.5106,
      "theoretical_loss": 3.437634770716132,
      "tokens_seen": 1950154752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161010101010101,
      "loss": 2.4883,
      "theoretical_loss": 3.4376250909490436,
      "tokens_seen": 1950220288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160808080808081,
      "loss": 2.6714,
      "theoretical_loss": 3.4376154115983084,
      "tokens_seen": 1950285824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160606060606061,
      "loss": 2.4813,
      "theoretical_loss": 3.4376057326638945,
      "tokens_seen": 1950351360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041604040404040406,
      "loss": 2.7367,
      "theoretical_loss": 3.4375960541457697,
      "tokens_seen": 1950416896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000416020202020202,
      "loss": 2.6954,
      "theoretical_loss": 3.437586376043902,
      "tokens_seen": 1950482432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000416,
      "loss": 2.605,
      "theoretical_loss": 3.43757669835826,
      "tokens_seen": 1950547968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041597979797979804,
      "loss": 2.552,
      "theoretical_loss": 3.437567021088812,
      "tokens_seen": 1950613504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415959595959596,
      "loss": 2.6863,
      "theoretical_loss": 3.4375573442355254,
      "tokens_seen": 1950679040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041593939393939395,
      "loss": 2.6037,
      "theoretical_loss": 3.4375476677983685,
      "tokens_seen": 1950744576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159191919191919,
      "loss": 2.7154,
      "theoretical_loss": 3.43753799177731,
      "tokens_seen": 1950810112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041589898989898987,
      "loss": 2.7183,
      "theoretical_loss": 3.437528316172317,
      "tokens_seen": 1950875648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041587878787878793,
      "loss": 2.7948,
      "theoretical_loss": 3.437518640983359,
      "tokens_seen": 1950941184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158585858585859,
      "loss": 2.5989,
      "theoretical_loss": 3.4375089662104035,
      "tokens_seen": 1951006720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041583838383838385,
      "loss": 2.6351,
      "theoretical_loss": 3.4374992918534177,
      "tokens_seen": 1951072256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158181818181818,
      "loss": 2.4972,
      "theoretical_loss": 3.4374896179123713,
      "tokens_seen": 1951137792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041579797979797976,
      "loss": 2.5836,
      "theoretical_loss": 3.4374799443872317,
      "tokens_seen": 1951203328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157777777777778,
      "loss": 2.4592,
      "theoretical_loss": 3.4374702712779666,
      "tokens_seen": 1951268864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157575757575758,
      "loss": 2.6278,
      "theoretical_loss": 3.4374605985845452,
      "tokens_seen": 1951334400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041573737373737374,
      "loss": 2.7606,
      "theoretical_loss": 3.4374509263069353,
      "tokens_seen": 1951399936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157171717171717,
      "loss": 2.5732,
      "theoretical_loss": 3.437441254445104,
      "tokens_seen": 1951465472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156969696969697,
      "loss": 2.6677,
      "theoretical_loss": 3.4374315829990207,
      "tokens_seen": 1951531008
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1074960,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.260453701019287,
      "objective/train/theoretical_loss": 3.437421911968653,
      "objective/train/tokens_used": 310455776,
      "theoretical_loss": 3.437421911968653,
      "tokens_seen": 1951596544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156767676767677,
      "loss": 2.4721,
      "theoretical_loss": 3.437421911968653,
      "tokens_seen": 1951596544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156565656565657,
      "loss": 2.5506,
      "theoretical_loss": 3.43741224135397,
      "tokens_seen": 1951662080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041563636363636363,
      "loss": 2.6751,
      "theoretical_loss": 3.4374025711549385,
      "tokens_seen": 1951727616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156161616161616,
      "loss": 2.5199,
      "theoretical_loss": 3.4373929013715276,
      "tokens_seen": 1951793152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155959595959596,
      "loss": 2.6086,
      "theoretical_loss": 3.437383232003705,
      "tokens_seen": 1951858688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155757575757576,
      "loss": 2.5168,
      "theoretical_loss": 3.4373735630514393,
      "tokens_seen": 1951924224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041555555555555557,
      "loss": 2.6447,
      "theoretical_loss": 3.4373638945146983,
      "tokens_seen": 1951989760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155353535353535,
      "loss": 2.4685,
      "theoretical_loss": 3.4373542263934507,
      "tokens_seen": 1952055296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041551515151515154,
      "loss": 2.5438,
      "theoretical_loss": 3.4373445586876636,
      "tokens_seen": 1952120832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154949494949495,
      "loss": 2.5168,
      "theoretical_loss": 3.4373348913973065,
      "tokens_seen": 1952186368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154747474747475,
      "loss": 2.5002,
      "theoretical_loss": 3.4373252245223473,
      "tokens_seen": 1952251904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041545454545454546,
      "loss": 2.6449,
      "theoretical_loss": 3.4373155580627532,
      "tokens_seen": 1952317440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154343434343434,
      "loss": 2.5862,
      "theoretical_loss": 3.437305892018494,
      "tokens_seen": 1952382976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041541414141414143,
      "loss": 2.6386,
      "theoretical_loss": 3.437296226389537,
      "tokens_seen": 1952448512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153939393939394,
      "loss": 2.448,
      "theoretical_loss": 3.43728656117585,
      "tokens_seen": 1952514048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153737373737374,
      "loss": 2.7955,
      "theoretical_loss": 3.4372768963774023,
      "tokens_seen": 1952579584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041535353535353535,
      "loss": 2.5514,
      "theoretical_loss": 3.437267231994161,
      "tokens_seen": 1952645120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041533333333333336,
      "loss": 2.7479,
      "theoretical_loss": 3.4372575680260953,
      "tokens_seen": 1952710656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153131313131313,
      "loss": 2.6988,
      "theoretical_loss": 3.437247904473173,
      "tokens_seen": 1952776192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152929292929293,
      "loss": 2.716,
      "theoretical_loss": 3.437238241335362,
      "tokens_seen": 1952841728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152727272727273,
      "loss": 2.7452,
      "theoretical_loss": 3.4372285786126318,
      "tokens_seen": 1952907264
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041525252525252525,
      "loss": 2.5785,
      "theoretical_loss": 3.437218916304949,
      "tokens_seen": 1952972800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041523232323232326,
      "loss": 2.5327,
      "theoretical_loss": 3.4372092544122825,
      "tokens_seen": 1953038336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152121212121212,
      "loss": 2.6263,
      "theoretical_loss": 3.4371995929346006,
      "tokens_seen": 1953103872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041519191919191917,
      "loss": 2.5292,
      "theoretical_loss": 3.437189931871872,
      "tokens_seen": 1953169408
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1075533,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5056207180023193,
      "objective/train/theoretical_loss": 3.437180271224064,
      "objective/train/tokens_used": 312094176,
      "theoretical_loss": 3.437180271224064,
      "tokens_seen": 1953234944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151717171717172,
      "loss": 2.6481,
      "theoretical_loss": 3.437180271224064,
      "tokens_seen": 1953234944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151515151515152,
      "loss": 2.801,
      "theoretical_loss": 3.437170610991146,
      "tokens_seen": 1953300480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041513131313131315,
      "loss": 2.3028,
      "theoretical_loss": 3.4371609511730856,
      "tokens_seen": 1953366016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151111111111111,
      "loss": 2.6984,
      "theoretical_loss": 3.437151291769851,
      "tokens_seen": 1953431552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041509090909090906,
      "loss": 2.6107,
      "theoretical_loss": 3.437141632781411,
      "tokens_seen": 1953497088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150707070707071,
      "loss": 2.8986,
      "theoretical_loss": 3.4371319742077326,
      "tokens_seen": 1953562624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150505050505051,
      "loss": 2.4985,
      "theoretical_loss": 3.4371223160487854,
      "tokens_seen": 1953628160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041503030303030304,
      "loss": 2.4593,
      "theoretical_loss": 3.437112658304537,
      "tokens_seen": 1953693696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415010101010101,
      "loss": 2.7836,
      "theoretical_loss": 3.4371030009749566,
      "tokens_seen": 1953759232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041498989898989896,
      "loss": 2.5408,
      "theoretical_loss": 3.437093344060011,
      "tokens_seen": 1953824768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414969696969697,
      "loss": 2.7091,
      "theoretical_loss": 3.4370836875596695,
      "tokens_seen": 1953890304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414949494949495,
      "loss": 2.745,
      "theoretical_loss": 3.437074031473901,
      "tokens_seen": 1953955840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041492929292929293,
      "loss": 2.6009,
      "theoretical_loss": 3.437064375802672,
      "tokens_seen": 1954021376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149090909090909,
      "loss": 2.3948,
      "theoretical_loss": 3.4370547205459516,
      "tokens_seen": 1954086912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041488888888888885,
      "loss": 2.5832,
      "theoretical_loss": 3.4370450657037086,
      "tokens_seen": 1954152448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148686868686869,
      "loss": 2.6297,
      "theoretical_loss": 3.4370354112759114,
      "tokens_seen": 1954217984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041484848484848487,
      "loss": 2.5004,
      "theoretical_loss": 3.4370257572625276,
      "tokens_seen": 1954283520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041482828282828283,
      "loss": 2.7232,
      "theoretical_loss": 3.4370161036635256,
      "tokens_seen": 1954349056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148080808080808,
      "loss": 2.7519,
      "theoretical_loss": 3.437006450478875,
      "tokens_seen": 1954414592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147878787878788,
      "loss": 2.6824,
      "theoretical_loss": 3.4369967977085416,
      "tokens_seen": 1954480128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147676767676768,
      "loss": 2.5209,
      "theoretical_loss": 3.436987145352496,
      "tokens_seen": 1954545664
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041474747474747476,
      "loss": 2.6794,
      "theoretical_loss": 3.436977493410706,
      "tokens_seen": 1954611200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147272727272727,
      "loss": 2.577,
      "theoretical_loss": 3.436967841883139,
      "tokens_seen": 1954676736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147070707070707,
      "loss": 2.6933,
      "theoretical_loss": 3.436958190769764,
      "tokens_seen": 1954742272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146868686868687,
      "loss": 2.7894,
      "theoretical_loss": 3.4369485400705497,
      "tokens_seen": 1954807808
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1076714,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.991243839263916,
      "objective/train/theoretical_loss": 3.4369388897854636,
      "objective/train/tokens_used": 313732576,
      "theoretical_loss": 3.4369388897854636,
      "tokens_seen": 1954873344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146666666666667,
      "loss": 2.5591,
      "theoretical_loss": 3.4369388897854636,
      "tokens_seen": 1954873344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041464646464646466,
      "loss": 2.6222,
      "theoretical_loss": 3.436929239914475,
      "tokens_seen": 1954938880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146262626262626,
      "loss": 2.4555,
      "theoretical_loss": 3.436919590457552,
      "tokens_seen": 1955004416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146060606060606,
      "loss": 2.699,
      "theoretical_loss": 3.436909941414662,
      "tokens_seen": 1955069952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041458585858585863,
      "loss": 2.587,
      "theoretical_loss": 3.4369002927857744,
      "tokens_seen": 1955135488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145656565656566,
      "loss": 2.4276,
      "theoretical_loss": 3.436890644570857,
      "tokens_seen": 1955201024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041454545454545455,
      "loss": 2.6416,
      "theoretical_loss": 3.436880996769879,
      "tokens_seen": 1955266560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145252525252525,
      "loss": 2.5871,
      "theoretical_loss": 3.436871349382808,
      "tokens_seen": 1955332096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145050505050505,
      "loss": 2.6925,
      "theoretical_loss": 3.436861702409612,
      "tokens_seen": 1955397632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041448484848484853,
      "loss": 2.7467,
      "theoretical_loss": 3.4368520558502604,
      "tokens_seen": 1955463168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144646464646465,
      "loss": 2.5199,
      "theoretical_loss": 3.4368424097047208,
      "tokens_seen": 1955528704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041444444444444444,
      "loss": 2.6255,
      "theoretical_loss": 3.436832763972962,
      "tokens_seen": 1955594240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041442424242424245,
      "loss": 2.7099,
      "theoretical_loss": 3.4368231186549525,
      "tokens_seen": 1955659776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144040404040404,
      "loss": 2.6867,
      "theoretical_loss": 3.43681347375066,
      "tokens_seen": 1955725312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143838383838384,
      "loss": 2.7098,
      "theoretical_loss": 3.436803829260054,
      "tokens_seen": 1955790848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143636363636364,
      "loss": 2.3875,
      "theoretical_loss": 3.4367941851831016,
      "tokens_seen": 1955856384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041434343434343433,
      "loss": 2.6654,
      "theoretical_loss": 3.4367845415197715,
      "tokens_seen": 1955921920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041432323232323235,
      "loss": 2.6225,
      "theoretical_loss": 3.4367748982700332,
      "tokens_seen": 1955987456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143030303030303,
      "loss": 2.6428,
      "theoretical_loss": 3.436765255433854,
      "tokens_seen": 1956052992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142828282828283,
      "loss": 2.6854,
      "theoretical_loss": 3.436755613011203,
      "tokens_seen": 1956118528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041426262626262627,
      "loss": 2.4692,
      "theoretical_loss": 3.436745971002048,
      "tokens_seen": 1956184064
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142424242424243,
      "loss": 2.7986,
      "theoretical_loss": 3.4367363294063575,
      "tokens_seen": 1956249600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041422222222222224,
      "loss": 2.6346,
      "theoretical_loss": 3.4367266882241,
      "tokens_seen": 1956315136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142020202020202,
      "loss": 2.8425,
      "theoretical_loss": 3.436717047455245,
      "tokens_seen": 1956380672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141818181818182,
      "loss": 2.5537,
      "theoretical_loss": 3.436707407099759,
      "tokens_seen": 1956446208
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1077107,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.886265754699707,
      "objective/train/theoretical_loss": 3.4366977671576118,
      "objective/train/tokens_used": 315370976,
      "theoretical_loss": 3.4366977671576118,
      "tokens_seen": 1956511744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041416161616161616,
      "loss": 2.6059,
      "theoretical_loss": 3.4366977671576118,
      "tokens_seen": 1956511744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141414141414142,
      "loss": 2.7731,
      "theoretical_loss": 3.436688127628771,
      "tokens_seen": 1956577280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041412121212121213,
      "loss": 2.5947,
      "theoretical_loss": 3.4366784885132056,
      "tokens_seen": 1956642816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141010101010101,
      "loss": 2.571,
      "theoretical_loss": 3.436668849810884,
      "tokens_seen": 1956708352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140808080808081,
      "loss": 2.4155,
      "theoretical_loss": 3.436659211521774,
      "tokens_seen": 1956773888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041406060606060606,
      "loss": 2.5718,
      "theoretical_loss": 3.4366495736458456,
      "tokens_seen": 1956839424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041404040404040407,
      "loss": 2.4914,
      "theoretical_loss": 3.4366399361830657,
      "tokens_seen": 1956904960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414020202020202,
      "loss": 2.7574,
      "theoretical_loss": 3.436630299133403,
      "tokens_seen": 1956970496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414,
      "loss": 2.5857,
      "theoretical_loss": 3.4366206624968267,
      "tokens_seen": 1957036032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413979797979798,
      "loss": 2.6283,
      "theoretical_loss": 3.4366110262733045,
      "tokens_seen": 1957101568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413959595959596,
      "loss": 2.6388,
      "theoretical_loss": 3.4366013904628048,
      "tokens_seen": 1957167104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041393939393939396,
      "loss": 2.4807,
      "theoretical_loss": 3.4365917550652973,
      "tokens_seen": 1957232640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139191919191919,
      "loss": 2.5695,
      "theoretical_loss": 3.4365821200807494,
      "tokens_seen": 1957298176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041389898989898987,
      "loss": 2.7659,
      "theoretical_loss": 3.4365724855091293,
      "tokens_seen": 1957363712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138787878787879,
      "loss": 2.6308,
      "theoretical_loss": 3.4365628513504065,
      "tokens_seen": 1957429248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138585858585859,
      "loss": 2.6777,
      "theoretical_loss": 3.4365532176045486,
      "tokens_seen": 1957494784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041383838383838385,
      "loss": 2.4554,
      "theoretical_loss": 3.436543584271524,
      "tokens_seen": 1957560320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138181818181818,
      "loss": 2.8151,
      "theoretical_loss": 3.4365339513513025,
      "tokens_seen": 1957625856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041379797979797977,
      "loss": 2.473,
      "theoretical_loss": 3.4365243188438512,
      "tokens_seen": 1957691392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041377777777777783,
      "loss": 2.5284,
      "theoretical_loss": 3.436514686749139,
      "tokens_seen": 1957756928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137575757575758,
      "loss": 2.5062,
      "theoretical_loss": 3.4365050550671348,
      "tokens_seen": 1957822464
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041373737373737374,
      "loss": 2.6967,
      "theoretical_loss": 3.4364954237978065,
      "tokens_seen": 1957888000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137171717171717,
      "loss": 2.6102,
      "theoretical_loss": 3.436485792941123,
      "tokens_seen": 1957953536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041369696969696966,
      "loss": 2.6571,
      "theoretical_loss": 3.436476162497053,
      "tokens_seen": 1958019072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136767676767677,
      "loss": 2.9256,
      "theoretical_loss": 3.4364665324655643,
      "tokens_seen": 1958084608
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1078288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1963400840759277,
      "objective/train/theoretical_loss": 3.4364569028466265,
      "objective/train/tokens_used": 317009376,
      "theoretical_loss": 3.4364569028466265,
      "tokens_seen": 1958150144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136565656565657,
      "loss": 2.6203,
      "theoretical_loss": 3.4364569028466265,
      "tokens_seen": 1958150144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041363636363636364,
      "loss": 2.7172,
      "theoretical_loss": 3.4364472736402067,
      "tokens_seen": 1958215680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136161616161616,
      "loss": 2.5027,
      "theoretical_loss": 3.436437644846275,
      "tokens_seen": 1958281216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135959595959596,
      "loss": 2.6738,
      "theoretical_loss": 3.4364280164647987,
      "tokens_seen": 1958346752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135757575757576,
      "loss": 2.5267,
      "theoretical_loss": 3.4364183884957464,
      "tokens_seen": 1958412288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135555555555556,
      "loss": 2.4939,
      "theoretical_loss": 3.4364087609390874,
      "tokens_seen": 1958477824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041353535353535353,
      "loss": 2.5811,
      "theoretical_loss": 3.43639913379479,
      "tokens_seen": 1958543360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135151515151515,
      "loss": 2.3717,
      "theoretical_loss": 3.436389507062822,
      "tokens_seen": 1958608896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134949494949495,
      "loss": 2.5317,
      "theoretical_loss": 3.4363798807431527,
      "tokens_seen": 1958674432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134747474747475,
      "loss": 2.5381,
      "theoretical_loss": 3.436370254835751,
      "tokens_seen": 1958739968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041345454545454547,
      "loss": 2.5417,
      "theoretical_loss": 3.4363606293405837,
      "tokens_seen": 1958805504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134343434343434,
      "loss": 2.6023,
      "theoretical_loss": 3.4363510042576215,
      "tokens_seen": 1958871040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041341414141414143,
      "loss": 2.8136,
      "theoretical_loss": 3.436341379586832,
      "tokens_seen": 1958936576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133939393939394,
      "loss": 2.6125,
      "theoretical_loss": 3.4363317553281836,
      "tokens_seen": 1959002112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133737373737374,
      "loss": 2.5957,
      "theoretical_loss": 3.436322131481645,
      "tokens_seen": 1959067648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041335353535353536,
      "loss": 2.6129,
      "theoretical_loss": 3.436312508047185,
      "tokens_seen": 1959133184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133333333333333,
      "loss": 2.5397,
      "theoretical_loss": 3.4363028850247717,
      "tokens_seen": 1959198720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133131313131313,
      "loss": 2.409,
      "theoretical_loss": 3.436293262414374,
      "tokens_seen": 1959264256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132929292929293,
      "loss": 2.4528,
      "theoretical_loss": 3.4362836402159607,
      "tokens_seen": 1959329792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132727272727273,
      "loss": 2.6568,
      "theoretical_loss": 3.4362740184294998,
      "tokens_seen": 1959395328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041325252525252525,
      "loss": 2.6499,
      "theoretical_loss": 3.43626439705496,
      "tokens_seen": 1959460864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041323232323232326,
      "loss": 2.3524,
      "theoretical_loss": 3.4362547760923103,
      "tokens_seen": 1959526400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132121212121212,
      "loss": 2.5487,
      "theoretical_loss": 3.4362451555415197,
      "tokens_seen": 1959591936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131919191919192,
      "loss": 2.5426,
      "theoretical_loss": 3.436235535402555,
      "tokens_seen": 1959657472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131717171717172,
      "loss": 2.3804,
      "theoretical_loss": 3.4362259156753874,
      "tokens_seen": 1959723008
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1079604,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.561680793762207,
      "objective/train/theoretical_loss": 3.4362162963599827,
      "objective/train/tokens_used": 318647776,
      "theoretical_loss": 3.4362162963599827,
      "tokens_seen": 1959788544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041315151515151514,
      "loss": 2.7004,
      "theoretical_loss": 3.4362162963599827,
      "tokens_seen": 1959788544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041313131313131315,
      "loss": 2.4122,
      "theoretical_loss": 3.436206677456312,
      "tokens_seen": 1959854080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131111111111111,
      "loss": 2.7408,
      "theoretical_loss": 3.436197058964342,
      "tokens_seen": 1959919616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041309090909090907,
      "loss": 2.7353,
      "theoretical_loss": 3.4361874408840425,
      "tokens_seen": 1959985152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130707070707071,
      "loss": 2.7172,
      "theoretical_loss": 3.4361778232153815,
      "tokens_seen": 1960050688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130505050505051,
      "loss": 2.5049,
      "theoretical_loss": 3.436168205958328,
      "tokens_seen": 1960116224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041303030303030305,
      "loss": 2.602,
      "theoretical_loss": 3.4361585891128508,
      "tokens_seen": 1960181760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413010101010101,
      "loss": 2.6146,
      "theoretical_loss": 3.436148972678917,
      "tokens_seen": 1960247296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041298989898989896,
      "loss": 2.7808,
      "theoretical_loss": 3.436139356656498,
      "tokens_seen": 1960312832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041296969696969697,
      "loss": 2.6078,
      "theoretical_loss": 3.43612974104556,
      "tokens_seen": 1960378368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412949494949495,
      "loss": 2.7435,
      "theoretical_loss": 3.436120125846072,
      "tokens_seen": 1960443904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041292929292929294,
      "loss": 2.5543,
      "theoretical_loss": 3.436110511058004,
      "tokens_seen": 1960509440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129090909090909,
      "loss": 2.8017,
      "theoretical_loss": 3.4361008966813236,
      "tokens_seen": 1960574976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128888888888889,
      "loss": 2.7139,
      "theoretical_loss": 3.4360912827159997,
      "tokens_seen": 1960640512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128686868686869,
      "loss": 2.7691,
      "theoretical_loss": 3.4360816691620006,
      "tokens_seen": 1960706048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128484848484849,
      "loss": 2.558,
      "theoretical_loss": 3.436072056019295,
      "tokens_seen": 1960771584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041282828282828283,
      "loss": 2.5107,
      "theoretical_loss": 3.4360624432878524,
      "tokens_seen": 1960837120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128080808080808,
      "loss": 2.4678,
      "theoretical_loss": 3.4360528309676406,
      "tokens_seen": 1960902656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127878787878788,
      "loss": 2.6761,
      "theoretical_loss": 3.4360432190586288,
      "tokens_seen": 1960968192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127676767676768,
      "loss": 2.6746,
      "theoretical_loss": 3.436033607560785,
      "tokens_seen": 1961033728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041274747474747477,
      "loss": 2.8905,
      "theoretical_loss": 3.4360239964740784,
      "tokens_seen": 1961099264
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127272727272727,
      "loss": 2.3687,
      "theoretical_loss": 3.4360143857984777,
      "tokens_seen": 1961164800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127070707070707,
      "loss": 2.7308,
      "theoretical_loss": 3.4360047755339513,
      "tokens_seen": 1961230336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041268686868686875,
      "loss": 2.6371,
      "theoretical_loss": 3.4359951656804677,
      "tokens_seen": 1961295872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126666666666667,
      "loss": 2.6793,
      "theoretical_loss": 3.4359855562379957,
      "tokens_seen": 1961361408
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1080156,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.839082717895508,
      "objective/train/theoretical_loss": 3.435975947206505,
      "objective/train/tokens_used": 320286176,
      "theoretical_loss": 3.435975947206505,
      "tokens_seen": 1961426944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041264646464646466,
      "loss": 2.7623,
      "theoretical_loss": 3.435975947206505,
      "tokens_seen": 1961426944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126262626262626,
      "loss": 2.7329,
      "theoretical_loss": 3.435966338585963,
      "tokens_seen": 1961492480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126060606060606,
      "loss": 2.5897,
      "theoretical_loss": 3.4359567303763385,
      "tokens_seen": 1961558016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041258585858585864,
      "loss": 2.4654,
      "theoretical_loss": 3.435947122577601,
      "tokens_seen": 1961623552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125656565656566,
      "loss": 2.7386,
      "theoretical_loss": 3.4359375151897185,
      "tokens_seen": 1961689088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041254545454545455,
      "loss": 2.6012,
      "theoretical_loss": 3.43592790821266,
      "tokens_seen": 1961754624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125252525252525,
      "loss": 2.6477,
      "theoretical_loss": 3.4359183016463946,
      "tokens_seen": 1961820160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041250505050505047,
      "loss": 2.8662,
      "theoretical_loss": 3.43590869549089,
      "tokens_seen": 1961885696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041248484848484853,
      "loss": 2.455,
      "theoretical_loss": 3.435899089746116,
      "tokens_seen": 1961951232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124646464646465,
      "loss": 2.8418,
      "theoretical_loss": 3.4358894844120407,
      "tokens_seen": 1962016768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041244444444444445,
      "loss": 2.666,
      "theoretical_loss": 3.435879879488633,
      "tokens_seen": 1962082304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124242424242424,
      "loss": 2.6833,
      "theoretical_loss": 3.435870274975861,
      "tokens_seen": 1962147840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124040404040404,
      "loss": 2.5037,
      "theoretical_loss": 3.4358606708736943,
      "tokens_seen": 1962213376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123838383838384,
      "loss": 2.5875,
      "theoretical_loss": 3.435851067182101,
      "tokens_seen": 1962278912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123636363636364,
      "loss": 2.5752,
      "theoretical_loss": 3.435841463901051,
      "tokens_seen": 1962344448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041234343434343434,
      "loss": 2.7171,
      "theoretical_loss": 3.435831861030512,
      "tokens_seen": 1962409984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123232323232323,
      "loss": 2.5322,
      "theoretical_loss": 3.4358222585704525,
      "tokens_seen": 1962475520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123030303030303,
      "loss": 2.621,
      "theoretical_loss": 3.4358126565208416,
      "tokens_seen": 1962541056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122828282828283,
      "loss": 2.6025,
      "theoretical_loss": 3.4358030548816485,
      "tokens_seen": 1962606592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122626262626263,
      "loss": 2.6889,
      "theoretical_loss": 3.435793453652842,
      "tokens_seen": 1962672128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041224242424242423,
      "loss": 2.5616,
      "theoretical_loss": 3.4357838528343896,
      "tokens_seen": 1962737664
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041222222222222224,
      "loss": 2.445,
      "theoretical_loss": 3.4357742524262616,
      "tokens_seen": 1962803200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122020202020202,
      "loss": 2.5863,
      "theoretical_loss": 3.4357646524284258,
      "tokens_seen": 1962868736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121818181818182,
      "loss": 2.5758,
      "theoretical_loss": 3.435755052840851,
      "tokens_seen": 1962934272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041216161616161617,
      "loss": 2.7752,
      "theoretical_loss": 3.4357454536635066,
      "tokens_seen": 1962999808
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1081069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6743087768554688,
      "objective/train/theoretical_loss": 3.435735854896361,
      "objective/train/tokens_used": 321924576,
      "theoretical_loss": 3.435735854896361,
      "tokens_seen": 1963065344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121414141414141,
      "loss": 2.397,
      "theoretical_loss": 3.435735854896361,
      "tokens_seen": 1963065344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041212121212121214,
      "loss": 2.7075,
      "theoretical_loss": 3.4357262565393825,
      "tokens_seen": 1963130880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121010101010101,
      "loss": 2.6683,
      "theoretical_loss": 3.4357166585925407,
      "tokens_seen": 1963196416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004120808080808081,
      "loss": 2.4466,
      "theoretical_loss": 3.4357070610558043,
      "tokens_seen": 1963261952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041206060606060606,
      "loss": 2.7529,
      "theoretical_loss": 3.4356974639291415,
      "tokens_seen": 1963327488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041204040404040407,
      "loss": 2.6604,
      "theoretical_loss": 3.4356878672125215,
      "tokens_seen": 1963393024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041202020202020203,
      "loss": 2.4729,
      "theoretical_loss": 3.435678270905913,
      "tokens_seen": 1963458560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412,
      "loss": 2.9114,
      "theoretical_loss": 3.4356686750092846,
      "tokens_seen": 1963524096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000411979797979798,
      "loss": 2.6102,
      "theoretical_loss": 3.4356590795226056,
      "tokens_seen": 1963589632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041195959595959595,
      "loss": 2.5086,
      "theoretical_loss": 3.4356494844458445,
      "tokens_seen": 1963655168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041193939393939396,
      "loss": 2.7648,
      "theoretical_loss": 3.43563988977897,
      "tokens_seen": 1963720704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119191919191919,
      "loss": 2.4557,
      "theoretical_loss": 3.435630295521951,
      "tokens_seen": 1963786240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118989898989899,
      "loss": 2.5652,
      "theoretical_loss": 3.4356207016747566,
      "tokens_seen": 1963851776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118787878787879,
      "loss": 2.5733,
      "theoretical_loss": 3.4356111082373553,
      "tokens_seen": 1963917312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118585858585859,
      "loss": 2.6456,
      "theoretical_loss": 3.435601515209716,
      "tokens_seen": 1963982848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041183838383838386,
      "loss": 2.5374,
      "theoretical_loss": 3.435591922591807,
      "tokens_seen": 1964048384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118181818181818,
      "loss": 2.4767,
      "theoretical_loss": 3.435582330383598,
      "tokens_seen": 1964113920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041179797979797977,
      "loss": 2.737,
      "theoretical_loss": 3.4355727385850576,
      "tokens_seen": 1964179456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117777777777778,
      "loss": 2.4674,
      "theoretical_loss": 3.435563147196155,
      "tokens_seen": 1964244992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117575757575758,
      "loss": 2.4722,
      "theoretical_loss": 3.435553556216858,
      "tokens_seen": 1964310528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041173737373737375,
      "loss": 2.5974,
      "theoretical_loss": 3.4355439656471356,
      "tokens_seen": 1964376064
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117171717171717,
      "loss": 2.8329,
      "theoretical_loss": 3.4355343754869576,
      "tokens_seen": 1964441600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041169696969696966,
      "loss": 2.654,
      "theoretical_loss": 3.435524785736292,
      "tokens_seen": 1964507136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041167676767676773,
      "loss": 2.7161,
      "theoretical_loss": 3.4355151963951083,
      "tokens_seen": 1964572672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116565656565657,
      "loss": 2.5793,
      "theoretical_loss": 3.4355056074633747,
      "tokens_seen": 1964638208
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 1081685,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.539858341217041,
      "objective/train/theoretical_loss": 3.4354960189410604,
      "objective/train/tokens_used": 323562976,
      "theoretical_loss": 3.4354960189410604,
      "tokens_seen": 1964703744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041163636363636364,
      "loss": 2.7547,
      "theoretical_loss": 3.4354960189410604,
      "tokens_seen": 1964703744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004116161616161616,
      "loss": 2.8963,
      "theoretical_loss": 3.435486430828134,
      "tokens_seen": 1964769280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041159595959595956,
      "loss": 2.6418,
      "theoretical_loss": 3.4354768431245652,
      "tokens_seen": 1964834816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115757575757576,
      "loss": 2.8092,
      "theoretical_loss": 3.435467255830322,
      "tokens_seen": 1964900352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115555555555556,
      "loss": 2.5482,
      "theoretical_loss": 3.4354576689453733,
      "tokens_seen": 1964965888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041153535353535354,
      "loss": 2.6656,
      "theoretical_loss": 3.4354480824696885,
      "tokens_seen": 1965031424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115151515151515,
      "loss": 2.71,
      "theoretical_loss": 3.4354384964032363,
      "tokens_seen": 1965096960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114949494949495,
      "loss": 2.7688,
      "theoretical_loss": 3.4354289107459848,
      "tokens_seen": 1965162496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114747474747475,
      "loss": 2.3955,
      "theoretical_loss": 3.435419325497904,
      "tokens_seen": 1965228032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041145454545454547,
      "loss": 2.6943,
      "theoretical_loss": 3.4354097406589625,
      "tokens_seen": 1965293568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041143434343434343,
      "loss": 2.6809,
      "theoretical_loss": 3.4354001562291288,
      "tokens_seen": 1965359104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114141414141414,
      "loss": 2.6625,
      "theoretical_loss": 3.435390572208372,
      "tokens_seen": 1965424640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113939393939394,
      "loss": 2.6743,
      "theoretical_loss": 3.4353809885966613,
      "tokens_seen": 1965490176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113737373737374,
      "loss": 2.4488,
      "theoretical_loss": 3.4353714053939655,
      "tokens_seen": 1965555712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041135353535353536,
      "loss": 2.6777,
      "theoretical_loss": 3.4353618226002527,
      "tokens_seen": 1965621248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113333333333333,
      "loss": 2.6546,
      "theoretical_loss": 3.435352240215493,
      "tokens_seen": 1965686784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041131313131313133,
      "loss": 2.5162,
      "theoretical_loss": 3.4353426582396542,
      "tokens_seen": 1965752320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112929292929293,
      "loss": 2.5366,
      "theoretical_loss": 3.4353330766727064,
      "tokens_seen": 1965817856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112727272727273,
      "loss": 2.6484,
      "theoretical_loss": 3.4353234955146177,
      "tokens_seen": 1965883392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041125252525252526,
      "loss": 2.574,
      "theoretical_loss": 3.4353139147653575,
      "tokens_seen": 1965948928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112323232323232,
      "loss": 2.3125,
      "theoretical_loss": 3.4353043344248944,
      "tokens_seen": 1966014464
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112121212121212,
      "loss": 2.5873,
      "theoretical_loss": 3.4352947544931967,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041119191919191924,
      "loss": 2.6623,
      "theoretical_loss": 3.435285174970235,
      "tokens_seen": 1966145536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111717171717172,
      "loss": 2.7218,
      "theoretical_loss": 3.435275595855977,
      "tokens_seen": 1966211072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041115151515151515,
      "loss": 2.4976,
      "theoretical_loss": 3.4352660171503913,
      "tokens_seen": 1966276608
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1082856,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7312498092651367,
      "objective/train/theoretical_loss": 3.4352564388534477,
      "objective/train/tokens_used": 325201376,
      "theoretical_loss": 3.4352564388534477,
      "tokens_seen": 1966342144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041113131313131316,
      "loss": 2.6919,
      "theoretical_loss": 3.4352564388534477,
      "tokens_seen": 1966342144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111111111111111,
      "loss": 2.5451,
      "theoretical_loss": 3.4352468609651154,
      "tokens_seen": 1966407680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041109090909090913,
      "loss": 2.4027,
      "theoretical_loss": 3.435237283485362,
      "tokens_seen": 1966473216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110707070707071,
      "loss": 2.6824,
      "theoretical_loss": 3.4352277064141585,
      "tokens_seen": 1966538752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041105050505050504,
      "loss": 2.4505,
      "theoretical_loss": 3.4352181297514717,
      "tokens_seen": 1966604288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041103030303030305,
      "loss": 2.4184,
      "theoretical_loss": 3.4352085534972723,
      "tokens_seen": 1966669824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000411010101010101,
      "loss": 2.5168,
      "theoretical_loss": 3.4351989776515275,
      "tokens_seen": 1966735360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410989898989899,
      "loss": 2.5689,
      "theoretical_loss": 3.435189402214208,
      "tokens_seen": 1966800896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410969696969697,
      "loss": 3.0605,
      "theoretical_loss": 3.435179827185282,
      "tokens_seen": 1966866432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410949494949495,
      "loss": 2.6067,
      "theoretical_loss": 3.4351702525647183,
      "tokens_seen": 1966931968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041092929292929295,
      "loss": 2.7054,
      "theoretical_loss": 3.435160678352486,
      "tokens_seen": 1966997504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109090909090909,
      "loss": 2.3814,
      "theoretical_loss": 3.4351511045485545,
      "tokens_seen": 1967063040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108888888888889,
      "loss": 2.8035,
      "theoretical_loss": 3.435141531152892,
      "tokens_seen": 1967128576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041086868686868687,
      "loss": 2.5896,
      "theoretical_loss": 3.4351319581654685,
      "tokens_seen": 1967194112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108484848484849,
      "loss": 2.9124,
      "theoretical_loss": 3.4351223855862525,
      "tokens_seen": 1967259648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041082828282828284,
      "loss": 2.7392,
      "theoretical_loss": 3.4351128134152122,
      "tokens_seen": 1967325184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108080808080808,
      "loss": 2.5456,
      "theoretical_loss": 3.435103241652318,
      "tokens_seen": 1967390720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107878787878788,
      "loss": 2.6759,
      "theoretical_loss": 3.4350936702975376,
      "tokens_seen": 1967456256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041076767676767676,
      "loss": 2.6334,
      "theoretical_loss": 3.435084099350841,
      "tokens_seen": 1967521792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107474747474748,
      "loss": 2.409,
      "theoretical_loss": 3.435074528812197,
      "tokens_seen": 1967587328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041072727272727273,
      "loss": 2.4961,
      "theoretical_loss": 3.4350649586815742,
      "tokens_seen": 1967652864
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107070707070707,
      "loss": 2.6618,
      "theoretical_loss": 3.435055388958942,
      "tokens_seen": 1967718400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106868686868687,
      "loss": 2.6335,
      "theoretical_loss": 3.4350458196442695,
      "tokens_seen": 1967783936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106666666666667,
      "loss": 2.6031,
      "theoretical_loss": 3.435036250737525,
      "tokens_seen": 1967849472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041064646464646467,
      "loss": 2.2975,
      "theoretical_loss": 3.4350266822386786,
      "tokens_seen": 1967915008
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1083372,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8083508014678955,
      "objective/train/theoretical_loss": 3.4350171141476986,
      "objective/train/tokens_used": 326839776,
      "theoretical_loss": 3.4350171141476986,
      "tokens_seen": 1967980544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106262626262626,
      "loss": 2.7665,
      "theoretical_loss": 3.4350171141476986,
      "tokens_seen": 1967980544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106060606060606,
      "loss": 2.6247,
      "theoretical_loss": 3.435007546464554,
      "tokens_seen": 1968046080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105858585858586,
      "loss": 2.6699,
      "theoretical_loss": 3.4349979791892142,
      "tokens_seen": 1968111616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105656565656566,
      "loss": 2.6583,
      "theoretical_loss": 3.434988412321648,
      "tokens_seen": 1968177152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041054545454545456,
      "loss": 2.6518,
      "theoretical_loss": 3.4349788458618247,
      "tokens_seen": 1968242688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105252525252525,
      "loss": 2.5485,
      "theoretical_loss": 3.434969279809713,
      "tokens_seen": 1968308224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105050505050505,
      "loss": 2.3527,
      "theoretical_loss": 3.434959714165282,
      "tokens_seen": 1968373760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041048484848484854,
      "loss": 2.6313,
      "theoretical_loss": 3.4349501489285013,
      "tokens_seen": 1968439296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104646464646465,
      "loss": 2.6005,
      "theoretical_loss": 3.4349405840993392,
      "tokens_seen": 1968504832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041044444444444445,
      "loss": 2.6,
      "theoretical_loss": 3.434931019677765,
      "tokens_seen": 1968570368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104242424242424,
      "loss": 2.7116,
      "theoretical_loss": 3.434921455663748,
      "tokens_seen": 1968635904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041040404040404037,
      "loss": 2.48,
      "theoretical_loss": 3.4349118920572574,
      "tokens_seen": 1968701440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041038383838383843,
      "loss": 2.661,
      "theoretical_loss": 3.4349023288582616,
      "tokens_seen": 1968766976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103636363636364,
      "loss": 2.5923,
      "theoretical_loss": 3.4348927660667306,
      "tokens_seen": 1968832512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041034343434343435,
      "loss": 2.3861,
      "theoretical_loss": 3.4348832036826322,
      "tokens_seen": 1968898048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103232323232323,
      "loss": 2.492,
      "theoretical_loss": 3.434873641705937,
      "tokens_seen": 1968963584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103030303030303,
      "loss": 2.7685,
      "theoretical_loss": 3.434864080136612,
      "tokens_seen": 1969029120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102828282828283,
      "loss": 2.8386,
      "theoretical_loss": 3.434854518974629,
      "tokens_seen": 1969094656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102626262626263,
      "loss": 2.8557,
      "theoretical_loss": 3.434844958219955,
      "tokens_seen": 1969160192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041024242424242424,
      "loss": 2.5892,
      "theoretical_loss": 3.4348353978725603,
      "tokens_seen": 1969225728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102222222222222,
      "loss": 2.5535,
      "theoretical_loss": 3.434825837932413,
      "tokens_seen": 1969291264
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102020202020202,
      "loss": 2.66,
      "theoretical_loss": 3.4348162783994827,
      "tokens_seen": 1969356800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101818181818182,
      "loss": 2.5879,
      "theoretical_loss": 3.4348067192737384,
      "tokens_seen": 1969422336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101616161616162,
      "loss": 2.7818,
      "theoretical_loss": 3.4347971605551493,
      "tokens_seen": 1969487872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041014141414141413,
      "loss": 2.5091,
      "theoretical_loss": 3.4347876022436843,
      "tokens_seen": 1969553408
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1084308,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9995901584625244,
      "objective/train/theoretical_loss": 3.434778044339313,
      "objective/train/tokens_used": 328478176,
      "theoretical_loss": 3.434778044339313,
      "tokens_seen": 1969618944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041012121212121214,
      "loss": 2.6201,
      "theoretical_loss": 3.434778044339313,
      "tokens_seen": 1969618944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101010101010101,
      "loss": 2.7089,
      "theoretical_loss": 3.434768486842005,
      "tokens_seen": 1969684480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100808080808081,
      "loss": 2.6555,
      "theoretical_loss": 3.4347589297517276,
      "tokens_seen": 1969750016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041006060606060607,
      "loss": 2.3727,
      "theoretical_loss": 3.434749373068451,
      "tokens_seen": 1969815552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410040404040404,
      "loss": 2.6942,
      "theoretical_loss": 3.4347398167921446,
      "tokens_seen": 1969881088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041002020202020203,
      "loss": 2.5807,
      "theoretical_loss": 3.4347302609227772,
      "tokens_seen": 1969946624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041,
      "loss": 2.6507,
      "theoretical_loss": 3.434720705460318,
      "tokens_seen": 1970012160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409979797979798,
      "loss": 2.5104,
      "theoretical_loss": 3.434711150404736,
      "tokens_seen": 1970077696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040995959595959596,
      "loss": 2.57,
      "theoretical_loss": 3.4347015957560005,
      "tokens_seen": 1970143232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040993939393939397,
      "loss": 2.3079,
      "theoretical_loss": 3.4346920415140803,
      "tokens_seen": 1970208768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099191919191919,
      "loss": 2.6171,
      "theoretical_loss": 3.4346824876789452,
      "tokens_seen": 1970274304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098989898989899,
      "loss": 2.5322,
      "theoretical_loss": 3.434672934250564,
      "tokens_seen": 1970339840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098787878787879,
      "loss": 2.6325,
      "theoretical_loss": 3.434663381228906,
      "tokens_seen": 1970405376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040985858585858585,
      "loss": 2.6111,
      "theoretical_loss": 3.434653828613939,
      "tokens_seen": 1970470912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040983838383838386,
      "loss": 2.6792,
      "theoretical_loss": 3.434644276405635,
      "tokens_seen": 1970536448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098181818181818,
      "loss": 2.6454,
      "theoretical_loss": 3.434634724603961,
      "tokens_seen": 1970601984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097979797979798,
      "loss": 2.408,
      "theoretical_loss": 3.434625173208886,
      "tokens_seen": 1970667520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097777777777778,
      "loss": 2.5299,
      "theoretical_loss": 3.43461562222038,
      "tokens_seen": 1970733056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097575757575758,
      "loss": 2.701,
      "theoretical_loss": 3.4346060716384126,
      "tokens_seen": 1970798592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040973737373737376,
      "loss": 2.6637,
      "theoretical_loss": 3.434596521462952,
      "tokens_seen": 1970864128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097171717171717,
      "loss": 2.4817,
      "theoretical_loss": 3.434586971693968,
      "tokens_seen": 1970929664
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040969696969696967,
      "loss": 2.6051,
      "theoretical_loss": 3.4345774223314294,
      "tokens_seen": 1970995200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096767676767677,
      "loss": 2.4224,
      "theoretical_loss": 3.434567873375306,
      "tokens_seen": 1971060736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096565656565657,
      "loss": 2.8169,
      "theoretical_loss": 3.4345583248255656,
      "tokens_seen": 1971126272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040963636363636365,
      "loss": 2.439,
      "theoretical_loss": 3.4345487766821794,
      "tokens_seen": 1971191808
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1085454,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7234983444213867,
      "objective/train/theoretical_loss": 3.434539228945115,
      "objective/train/tokens_used": 330116576,
      "theoretical_loss": 3.434539228945115,
      "tokens_seen": 1971257344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096161616161616,
      "loss": 2.4934,
      "theoretical_loss": 3.434539228945115,
      "tokens_seen": 1971257344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040959595959595956,
      "loss": 2.5968,
      "theoretical_loss": 3.4345296816143422,
      "tokens_seen": 1971322880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040957575757575763,
      "loss": 2.8397,
      "theoretical_loss": 3.43452013468983,
      "tokens_seen": 1971388416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095555555555556,
      "loss": 2.6118,
      "theoretical_loss": 3.4345105881715483,
      "tokens_seen": 1971453952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040953535353535354,
      "loss": 2.785,
      "theoretical_loss": 3.4345010420594653,
      "tokens_seen": 1971519488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095151515151515,
      "loss": 2.6511,
      "theoretical_loss": 3.4344914963535507,
      "tokens_seen": 1971585024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040949494949494945,
      "loss": 2.5009,
      "theoretical_loss": 3.434481951053774,
      "tokens_seen": 1971650560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094747474747475,
      "loss": 2.7502,
      "theoretical_loss": 3.4344724061601037,
      "tokens_seen": 1971716096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094545454545455,
      "loss": 2.4277,
      "theoretical_loss": 3.4344628616725097,
      "tokens_seen": 1971781632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040943434343434343,
      "loss": 2.6531,
      "theoretical_loss": 3.434453317590961,
      "tokens_seen": 1971847168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094141414141414,
      "loss": 2.5949,
      "theoretical_loss": 3.4344437739154268,
      "tokens_seen": 1971912704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040939393939393946,
      "loss": 2.7611,
      "theoretical_loss": 3.434434230645876,
      "tokens_seen": 1971978240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093737373737374,
      "loss": 2.6381,
      "theoretical_loss": 3.434424687782279,
      "tokens_seen": 1972043776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040935353535353537,
      "loss": 2.4472,
      "theoretical_loss": 3.4344151453246035,
      "tokens_seen": 1972109312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093333333333333,
      "loss": 2.448,
      "theoretical_loss": 3.4344056032728196,
      "tokens_seen": 1972174848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093131313131313,
      "loss": 2.6905,
      "theoretical_loss": 3.4343960616268965,
      "tokens_seen": 1972240384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040929292929292935,
      "loss": 2.8451,
      "theoretical_loss": 3.4343865203868034,
      "tokens_seen": 1972305920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092727272727273,
      "loss": 2.4575,
      "theoretical_loss": 3.4343769795525096,
      "tokens_seen": 1972371456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040925252525252526,
      "loss": 2.8152,
      "theoretical_loss": 3.4343674391239842,
      "tokens_seen": 1972436992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092323232323232,
      "loss": 2.4402,
      "theoretical_loss": 3.4343578991011965,
      "tokens_seen": 1972502528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092121212121212,
      "loss": 2.4436,
      "theoretical_loss": 3.4343483594841158,
      "tokens_seen": 1972568064
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040919191919191924,
      "loss": 2.845,
      "theoretical_loss": 3.4343388202727114,
      "tokens_seen": 1972633600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091717171717172,
      "loss": 2.7507,
      "theoretical_loss": 3.4343292814669524,
      "tokens_seen": 1972699136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040915151515151515,
      "loss": 2.6888,
      "theoretical_loss": 3.4343197430668084,
      "tokens_seen": 1972764672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091313131313131,
      "loss": 2.6904,
      "theoretical_loss": 3.4343102050722485,
      "tokens_seen": 1972830208
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1086100,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.58170747756958,
      "objective/train/theoretical_loss": 3.434300667483242,
      "objective/train/tokens_used": 331754976,
      "theoretical_loss": 3.434300667483242,
      "tokens_seen": 1972895744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091111111111111,
      "loss": 2.5805,
      "theoretical_loss": 3.434300667483242,
      "tokens_seen": 1972895744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040909090909090913,
      "loss": 2.5536,
      "theoretical_loss": 3.434291130299758,
      "tokens_seen": 1972961280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090707070707071,
      "loss": 2.4578,
      "theoretical_loss": 3.4342815935217663,
      "tokens_seen": 1973026816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040905050505050505,
      "loss": 2.7724,
      "theoretical_loss": 3.4342720571492356,
      "tokens_seen": 1973092352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409030303030303,
      "loss": 2.7268,
      "theoretical_loss": 3.434262521182135,
      "tokens_seen": 1973157888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409010101010101,
      "loss": 2.3653,
      "theoretical_loss": 3.434252985620435,
      "tokens_seen": 1973223424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408989898989899,
      "loss": 2.8847,
      "theoretical_loss": 3.4342434504641037,
      "tokens_seen": 1973288960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408969696969697,
      "loss": 2.5669,
      "theoretical_loss": 3.4342339157131114,
      "tokens_seen": 1973354496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040894949494949494,
      "loss": 2.6521,
      "theoretical_loss": 3.434224381367426,
      "tokens_seen": 1973420032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040892929292929295,
      "loss": 2.7036,
      "theoretical_loss": 3.434214847427018,
      "tokens_seen": 1973485568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089090909090909,
      "loss": 2.5247,
      "theoretical_loss": 3.434205313891857,
      "tokens_seen": 1973551104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088888888888889,
      "loss": 2.5833,
      "theoretical_loss": 3.434195780761911,
      "tokens_seen": 1973616640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088686868686869,
      "loss": 2.6482,
      "theoretical_loss": 3.4341862480371503,
      "tokens_seen": 1973682176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040884848484848483,
      "loss": 2.8507,
      "theoretical_loss": 3.434176715717544,
      "tokens_seen": 1973747712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040882828282828284,
      "loss": 2.4306,
      "theoretical_loss": 3.434167183803061,
      "tokens_seen": 1973813248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088080808080808,
      "loss": 2.6021,
      "theoretical_loss": 3.4341576522936714,
      "tokens_seen": 1973878784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087878787878788,
      "loss": 2.7193,
      "theoretical_loss": 3.4341481211893443,
      "tokens_seen": 1973944320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040876767676767677,
      "loss": 2.6304,
      "theoretical_loss": 3.4341385904900483,
      "tokens_seen": 1974009856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087474747474748,
      "loss": 2.5909,
      "theoretical_loss": 3.4341290601957537,
      "tokens_seen": 1974075392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040872727272727274,
      "loss": 2.6612,
      "theoretical_loss": 3.4341195303064294,
      "tokens_seen": 1974140928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087070707070707,
      "loss": 2.5381,
      "theoretical_loss": 3.4341100008220447,
      "tokens_seen": 1974206464
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086868686868687,
      "loss": 2.8058,
      "theoretical_loss": 3.4341004717425694,
      "tokens_seen": 1974272000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040866666666666666,
      "loss": 2.653,
      "theoretical_loss": 3.4340909430679725,
      "tokens_seen": 1974337536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040864646464646467,
      "loss": 2.6647,
      "theoretical_loss": 3.4340814147982233,
      "tokens_seen": 1974403072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040862626262626263,
      "loss": 2.6635,
      "theoretical_loss": 3.4340718869332907,
      "tokens_seen": 1974468608
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1087285,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9132044315338135,
      "objective/train/theoretical_loss": 3.4340623594731454,
      "objective/train/tokens_used": 333393376,
      "theoretical_loss": 3.4340623594731454,
      "tokens_seen": 1974534144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086060606060606,
      "loss": 2.8076,
      "theoretical_loss": 3.4340623594731454,
      "tokens_seen": 1974534144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085858585858586,
      "loss": 2.4489,
      "theoretical_loss": 3.4340528324177555,
      "tokens_seen": 1974599680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085656565656566,
      "loss": 2.5364,
      "theoretical_loss": 3.4340433057670907,
      "tokens_seen": 1974665216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040854545454545457,
      "loss": 2.5742,
      "theoretical_loss": 3.434033779521121,
      "tokens_seen": 1974730752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085252525252525,
      "loss": 2.5656,
      "theoretical_loss": 3.434024253679815,
      "tokens_seen": 1974796288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085050505050505,
      "loss": 2.7908,
      "theoretical_loss": 3.4340147282431426,
      "tokens_seen": 1974861824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084848484848485,
      "loss": 2.8222,
      "theoretical_loss": 3.434005203211073,
      "tokens_seen": 1974927360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084646464646465,
      "loss": 2.6662,
      "theoretical_loss": 3.4339956785835755,
      "tokens_seen": 1974992896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040844444444444446,
      "loss": 2.673,
      "theoretical_loss": 3.4339861543606194,
      "tokens_seen": 1975058432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084242424242424,
      "loss": 2.5198,
      "theoretical_loss": 3.4339766305421744,
      "tokens_seen": 1975123968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040840404040404037,
      "loss": 2.5674,
      "theoretical_loss": 3.4339671071282094,
      "tokens_seen": 1975189504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040838383838383844,
      "loss": 2.503,
      "theoretical_loss": 3.433957584118695,
      "tokens_seen": 1975255040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083636363636364,
      "loss": 2.6536,
      "theoretical_loss": 3.4339480615135987,
      "tokens_seen": 1975320576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040834343434343435,
      "loss": 2.5693,
      "theoretical_loss": 3.4339385393128916,
      "tokens_seen": 1975386112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083232323232323,
      "loss": 2.4123,
      "theoretical_loss": 3.4339290175165424,
      "tokens_seen": 1975451648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040830303030303026,
      "loss": 2.7539,
      "theoretical_loss": 3.4339194961245205,
      "tokens_seen": 1975517184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040828282828282833,
      "loss": 2.8219,
      "theoretical_loss": 3.4339099751367956,
      "tokens_seen": 1975582720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082626262626263,
      "loss": 2.4702,
      "theoretical_loss": 3.4339004545533367,
      "tokens_seen": 1975648256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040824242424242424,
      "loss": 2.5782,
      "theoretical_loss": 3.433890934374114,
      "tokens_seen": 1975713792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082222222222222,
      "loss": 2.6308,
      "theoretical_loss": 3.433881414599096,
      "tokens_seen": 1975779328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082020202020202,
      "loss": 2.5198,
      "theoretical_loss": 3.4338718952282523,
      "tokens_seen": 1975844864
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081818181818182,
      "loss": 2.5395,
      "theoretical_loss": 3.4338623762615526,
      "tokens_seen": 1975910400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081616161616162,
      "loss": 2.5631,
      "theoretical_loss": 3.4338528576989664,
      "tokens_seen": 1975975936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040814141414141414,
      "loss": 2.5006,
      "theoretical_loss": 3.433843339540463,
      "tokens_seen": 1976041472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081212121212121,
      "loss": 2.6091,
      "theoretical_loss": 3.4338338217860116,
      "tokens_seen": 1976107008
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1088636,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3727774620056152,
      "objective/train/theoretical_loss": 3.4338243044355825,
      "objective/train/tokens_used": 335031776,
      "theoretical_loss": 3.4338243044355825,
      "tokens_seen": 1976172544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081010101010101,
      "loss": 2.5955,
      "theoretical_loss": 3.4338243044355825,
      "tokens_seen": 1976172544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080808080808081,
      "loss": 2.664,
      "theoretical_loss": 3.433814787489144,
      "tokens_seen": 1976238080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040806060606060607,
      "loss": 2.7374,
      "theoretical_loss": 3.4338052709466664,
      "tokens_seen": 1976303616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040804040404040403,
      "loss": 2.6428,
      "theoretical_loss": 3.433795754808119,
      "tokens_seen": 1976369152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040802020202020204,
      "loss": 2.6984,
      "theoretical_loss": 3.4337862390734717,
      "tokens_seen": 1976434688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408,
      "loss": 2.7755,
      "theoretical_loss": 3.4337767237426924,
      "tokens_seen": 1976500224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407979797979798,
      "loss": 2.5779,
      "theoretical_loss": 3.433767208815752,
      "tokens_seen": 1976565760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040795959595959596,
      "loss": 2.5811,
      "theoretical_loss": 3.43375769429262,
      "tokens_seen": 1976631296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079393939393939,
      "loss": 2.634,
      "theoretical_loss": 3.4337481801732648,
      "tokens_seen": 1976696832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040791919191919193,
      "loss": 2.7384,
      "theoretical_loss": 3.4337386664576566,
      "tokens_seen": 1976762368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078989898989899,
      "loss": 2.517,
      "theoretical_loss": 3.433729153145765,
      "tokens_seen": 1976827904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078787878787879,
      "loss": 2.4324,
      "theoretical_loss": 3.433719640237559,
      "tokens_seen": 1976893440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040785858585858586,
      "loss": 2.6609,
      "theoretical_loss": 3.4337101277330087,
      "tokens_seen": 1976958976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040783838383838387,
      "loss": 2.5041,
      "theoretical_loss": 3.433700615632083,
      "tokens_seen": 1977024512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078181818181818,
      "loss": 2.4608,
      "theoretical_loss": 3.433691103934752,
      "tokens_seen": 1977090048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077979797979798,
      "loss": 2.6694,
      "theoretical_loss": 3.4336815926409843,
      "tokens_seen": 1977155584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077777777777778,
      "loss": 2.6661,
      "theoretical_loss": 3.4336720817507507,
      "tokens_seen": 1977221120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040775757575757575,
      "loss": 2.6757,
      "theoretical_loss": 3.433662571264019,
      "tokens_seen": 1977286656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040773737373737376,
      "loss": 2.6244,
      "theoretical_loss": 3.4336530611807605,
      "tokens_seen": 1977352192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077171717171717,
      "loss": 2.7615,
      "theoretical_loss": 3.4336435515009436,
      "tokens_seen": 1977417728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040769696969696973,
      "loss": 2.8321,
      "theoretical_loss": 3.433634042224538,
      "tokens_seen": 1977483264
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076767676767677,
      "loss": 2.6079,
      "theoretical_loss": 3.433624533351513,
      "tokens_seen": 1977548800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076565656565657,
      "loss": 2.5141,
      "theoretical_loss": 3.4336150248818393,
      "tokens_seen": 1977614336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040763636363636365,
      "loss": 2.5957,
      "theoretical_loss": 3.4336055168154846,
      "tokens_seen": 1977679872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076161616161616,
      "loss": 2.5912,
      "theoretical_loss": 3.43359600915242,
      "tokens_seen": 1977745408
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1089194,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.324873208999634,
      "objective/train/theoretical_loss": 3.4335865018926146,
      "objective/train/tokens_used": 336670176,
      "theoretical_loss": 3.4335865018926146,
      "tokens_seen": 1977810944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075959595959596,
      "loss": 2.7609,
      "theoretical_loss": 3.4335865018926146,
      "tokens_seen": 1977810944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075757575757576,
      "loss": 2.6968,
      "theoretical_loss": 3.433576995036037,
      "tokens_seen": 1977876480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075555555555556,
      "loss": 2.7448,
      "theoretical_loss": 3.433567488582658,
      "tokens_seen": 1977942016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040753535353535355,
      "loss": 2.5191,
      "theoretical_loss": 3.433557982532447,
      "tokens_seen": 1978007552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075151515151515,
      "loss": 2.62,
      "theoretical_loss": 3.433548476885372,
      "tokens_seen": 1978073088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074949494949495,
      "loss": 2.632,
      "theoretical_loss": 3.433538971641405,
      "tokens_seen": 1978138624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040747474747474747,
      "loss": 2.6667,
      "theoretical_loss": 3.433529466800514,
      "tokens_seen": 1978204160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074545454545455,
      "loss": 2.6104,
      "theoretical_loss": 3.433519962362668,
      "tokens_seen": 1978269696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040743434343434344,
      "loss": 2.6996,
      "theoretical_loss": 3.433510458327838,
      "tokens_seen": 1978335232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074141414141414,
      "loss": 2.6062,
      "theoretical_loss": 3.433500954695993,
      "tokens_seen": 1978400768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073939393939394,
      "loss": 2.5305,
      "theoretical_loss": 3.4334914514671024,
      "tokens_seen": 1978466304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073737373737374,
      "loss": 2.5134,
      "theoretical_loss": 3.433481948641136,
      "tokens_seen": 1978531840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073535353535354,
      "loss": 2.7429,
      "theoretical_loss": 3.433472446218063,
      "tokens_seen": 1978597376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040733333333333333,
      "loss": 2.6326,
      "theoretical_loss": 3.4334629441978537,
      "tokens_seen": 1978662912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073131313131313,
      "loss": 2.7723,
      "theoretical_loss": 3.4334534425804764,
      "tokens_seen": 1978728448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072929292929293,
      "loss": 2.8548,
      "theoretical_loss": 3.433443941365902,
      "tokens_seen": 1978793984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072727272727273,
      "loss": 2.289,
      "theoretical_loss": 3.4334344405541,
      "tokens_seen": 1978859520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040725252525252527,
      "loss": 2.4712,
      "theoretical_loss": 3.4334249401450387,
      "tokens_seen": 1978925056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072323232323232,
      "loss": 2.5277,
      "theoretical_loss": 3.433415440138689,
      "tokens_seen": 1978990592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072121212121212,
      "loss": 2.5607,
      "theoretical_loss": 3.43340594053502,
      "tokens_seen": 1979056128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040719191919191925,
      "loss": 2.6904,
      "theoretical_loss": 3.4333964413340015,
      "tokens_seen": 1979121664
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071717171717172,
      "loss": 2.6751,
      "theoretical_loss": 3.4333869425356025,
      "tokens_seen": 1979187200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040715151515151516,
      "loss": 2.2723,
      "theoretical_loss": 3.4333774441397935,
      "tokens_seen": 1979252736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071313131313131,
      "loss": 2.6722,
      "theoretical_loss": 3.4333679461465434,
      "tokens_seen": 1979318272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071111111111111,
      "loss": 2.723,
      "theoretical_loss": 3.433358448555822,
      "tokens_seen": 1979383808
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1090532,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0842976570129395,
      "objective/train/theoretical_loss": 3.4333489513675994,
      "objective/train/tokens_used": 338308576,
      "theoretical_loss": 3.4333489513675994,
      "tokens_seen": 1979449344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040709090909090914,
      "loss": 2.5824,
      "theoretical_loss": 3.4333489513675994,
      "tokens_seen": 1979449344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070707070707071,
      "loss": 2.7518,
      "theoretical_loss": 3.4333394545818443,
      "tokens_seen": 1979514880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040705050505050505,
      "loss": 2.7846,
      "theoretical_loss": 3.4333299581985273,
      "tokens_seen": 1979580416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407030303030303,
      "loss": 2.6402,
      "theoretical_loss": 3.433320462217617,
      "tokens_seen": 1979645952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407010101010101,
      "loss": 2.6241,
      "theoretical_loss": 3.433310966639084,
      "tokens_seen": 1979711488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040698989898989903,
      "loss": 2.7387,
      "theoretical_loss": 3.4333014714628973,
      "tokens_seen": 1979777024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000406969696969697,
      "loss": 2.5149,
      "theoretical_loss": 3.433291976689027,
      "tokens_seen": 1979842560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040694949494949495,
      "loss": 2.7728,
      "theoretical_loss": 3.433282482317442,
      "tokens_seen": 1979908096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069292929292929,
      "loss": 2.5955,
      "theoretical_loss": 3.4332729883481123,
      "tokens_seen": 1979973632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069090909090909,
      "loss": 2.6294,
      "theoretical_loss": 3.4332634947810083,
      "tokens_seen": 1980039168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068888888888889,
      "loss": 2.6673,
      "theoretical_loss": 3.4332540016160986,
      "tokens_seen": 1980104704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068686868686869,
      "loss": 2.6137,
      "theoretical_loss": 3.4332445088533534,
      "tokens_seen": 1980170240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040684848484848484,
      "loss": 2.6255,
      "theoretical_loss": 3.433235016492742,
      "tokens_seen": 1980235776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040682828282828285,
      "loss": 2.3272,
      "theoretical_loss": 3.433225524534235,
      "tokens_seen": 1980301312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068080808080808,
      "loss": 2.7693,
      "theoretical_loss": 3.4332160329778003,
      "tokens_seen": 1980366848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067878787878788,
      "loss": 2.7496,
      "theoretical_loss": 3.433206541823409,
      "tokens_seen": 1980432384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067676767676768,
      "loss": 2.5695,
      "theoretical_loss": 3.4331970510710303,
      "tokens_seen": 1980497920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040674747474747473,
      "loss": 2.5576,
      "theoretical_loss": 3.433187560720634,
      "tokens_seen": 1980563456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040672727272727274,
      "loss": 2.7761,
      "theoretical_loss": 3.4331780707721897,
      "tokens_seen": 1980628992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067070707070707,
      "loss": 2.7222,
      "theoretical_loss": 3.433168581225667,
      "tokens_seen": 1980694528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066868686868687,
      "loss": 2.6214,
      "theoretical_loss": 3.433159092081036,
      "tokens_seen": 1980760064
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040666666666666667,
      "loss": 2.6567,
      "theoretical_loss": 3.433149603338266,
      "tokens_seen": 1980825600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066464646464647,
      "loss": 2.5001,
      "theoretical_loss": 3.433140114997326,
      "tokens_seen": 1980891136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040662626262626264,
      "loss": 2.5997,
      "theoretical_loss": 3.433130627058187,
      "tokens_seen": 1980956672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066060606060606,
      "loss": 2.5609,
      "theoretical_loss": 3.433121139520818,
      "tokens_seen": 1981022208
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 1091139,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3586647510528564,
      "objective/train/theoretical_loss": 3.433111652385189,
      "objective/train/tokens_used": 339946976,
      "theoretical_loss": 3.433111652385189,
      "tokens_seen": 1981087744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004065858585858586,
      "loss": 2.4624,
      "theoretical_loss": 3.433111652385189,
      "tokens_seen": 1981087744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040656565656565656,
      "loss": 2.6518,
      "theoretical_loss": 3.4331021656512695,
      "tokens_seen": 1981153280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040654545454545457,
      "loss": 2.4553,
      "theoretical_loss": 3.433092679319029,
      "tokens_seen": 1981218816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040652525252525253,
      "loss": 2.6482,
      "theoretical_loss": 3.4330831933884367,
      "tokens_seen": 1981284352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004065050505050505,
      "loss": 2.6432,
      "theoretical_loss": 3.4330737078594638,
      "tokens_seen": 1981349888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064848484848485,
      "loss": 2.4884,
      "theoretical_loss": 3.4330642227320793,
      "tokens_seen": 1981415424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064646464646465,
      "loss": 2.748,
      "theoretical_loss": 3.433054738006253,
      "tokens_seen": 1981480960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040644444444444446,
      "loss": 2.5247,
      "theoretical_loss": 3.433045253681954,
      "tokens_seen": 1981546496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064242424242424,
      "loss": 2.5808,
      "theoretical_loss": 3.4330357697591527,
      "tokens_seen": 1981612032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064040404040404,
      "loss": 2.5666,
      "theoretical_loss": 3.4330262862378182,
      "tokens_seen": 1981677568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063838383838384,
      "loss": 2.5371,
      "theoretical_loss": 3.4330168031179213,
      "tokens_seen": 1981743104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063636363636364,
      "loss": 2.6204,
      "theoretical_loss": 3.433007320399431,
      "tokens_seen": 1981808640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040634343434343436,
      "loss": 2.6785,
      "theoretical_loss": 3.4329978380823167,
      "tokens_seen": 1981874176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063232323232323,
      "loss": 2.4334,
      "theoretical_loss": 3.4329883561665486,
      "tokens_seen": 1981939712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040630303030303027,
      "loss": 2.8279,
      "theoretical_loss": 3.432978874652097,
      "tokens_seen": 1982005248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040628282828282834,
      "loss": 2.5103,
      "theoretical_loss": 3.4329693935389303,
      "tokens_seen": 1982070784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062626262626263,
      "loss": 2.7507,
      "theoretical_loss": 3.4329599128270196,
      "tokens_seen": 1982136320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040624242424242425,
      "loss": 2.8613,
      "theoretical_loss": 3.432950432516334,
      "tokens_seen": 1982201856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062222222222222,
      "loss": 2.3773,
      "theoretical_loss": 3.4329409526068426,
      "tokens_seen": 1982267392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040620202020202016,
      "loss": 2.5538,
      "theoretical_loss": 3.4329314730985163,
      "tokens_seen": 1982332928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040618181818181823,
      "loss": 2.7964,
      "theoretical_loss": 3.4329219939913247,
      "tokens_seen": 1982398464
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061616161616162,
      "loss": 2.7069,
      "theoretical_loss": 3.4329125152852367,
      "tokens_seen": 1982464000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040614141414141414,
      "loss": 2.6791,
      "theoretical_loss": 3.432903036980223,
      "tokens_seen": 1982529536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061212121212121,
      "loss": 2.5445,
      "theoretical_loss": 3.4328935590762533,
      "tokens_seen": 1982595072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040610101010101006,
      "loss": 2.6502,
      "theoretical_loss": 3.4328840815732966,
      "tokens_seen": 1982660608
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1091959,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3474152088165283,
      "objective/train/theoretical_loss": 3.4328746044713236,
      "objective/train/tokens_used": 341585376,
      "theoretical_loss": 3.4328746044713236,
      "tokens_seen": 1982726144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060808080808081,
      "loss": 2.7832,
      "theoretical_loss": 3.4328746044713236,
      "tokens_seen": 1982726144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060606060606061,
      "loss": 2.7179,
      "theoretical_loss": 3.4328651277703033,
      "tokens_seen": 1982791680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040604040404040403,
      "loss": 2.835,
      "theoretical_loss": 3.4328556514702058,
      "tokens_seen": 1982857216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000406020202020202,
      "loss": 2.4766,
      "theoretical_loss": 3.432846175571001,
      "tokens_seen": 1982922752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040600000000000006,
      "loss": 2.4752,
      "theoretical_loss": 3.432836700072659,
      "tokens_seen": 1982988288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405979797979798,
      "loss": 2.816,
      "theoretical_loss": 3.4328272249751492,
      "tokens_seen": 1983053824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040595959595959597,
      "loss": 2.5177,
      "theoretical_loss": 3.432817750278441,
      "tokens_seen": 1983119360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040593939393939393,
      "loss": 2.6334,
      "theoretical_loss": 3.432808275982505,
      "tokens_seen": 1983184896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004059191919191919,
      "loss": 2.5028,
      "theoretical_loss": 3.4327988020873104,
      "tokens_seen": 1983250432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040589898989898995,
      "loss": 2.5369,
      "theoretical_loss": 3.4327893285928273,
      "tokens_seen": 1983315968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058787878787879,
      "loss": 2.7305,
      "theoretical_loss": 3.4327798554990254,
      "tokens_seen": 1983381504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040585858585858586,
      "loss": 2.7252,
      "theoretical_loss": 3.432770382805874,
      "tokens_seen": 1983447040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058383838383838,
      "loss": 2.6841,
      "theoretical_loss": 3.4327609105133443,
      "tokens_seen": 1983512576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040581818181818183,
      "loss": 2.7614,
      "theoretical_loss": 3.432751438621405,
      "tokens_seen": 1983578112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040579797979797984,
      "loss": 2.857,
      "theoretical_loss": 3.4327419671300263,
      "tokens_seen": 1983643648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057777777777778,
      "loss": 2.7226,
      "theoretical_loss": 3.4327324960391783,
      "tokens_seen": 1983709184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040575757575757576,
      "loss": 2.3924,
      "theoretical_loss": 3.43272302534883,
      "tokens_seen": 1983774720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057373737373737,
      "loss": 2.7689,
      "theoretical_loss": 3.4327135550589514,
      "tokens_seen": 1983840256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057171717171717,
      "loss": 2.5132,
      "theoretical_loss": 3.4327040851695134,
      "tokens_seen": 1983905792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040569696969696973,
      "loss": 2.5889,
      "theoretical_loss": 3.432694615680485,
      "tokens_seen": 1983971328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056767676767677,
      "loss": 2.7441,
      "theoretical_loss": 3.4326851465918358,
      "tokens_seen": 1984036864
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040565656565656565,
      "loss": 2.6829,
      "theoretical_loss": 3.4326756779035357,
      "tokens_seen": 1984102400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040563636363636366,
      "loss": 2.701,
      "theoretical_loss": 3.4326662096155554,
      "tokens_seen": 1984167936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056161616161616,
      "loss": 2.8375,
      "theoretical_loss": 3.4326567417278637,
      "tokens_seen": 1984233472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040559595959595963,
      "loss": 2.5529,
      "theoretical_loss": 3.4326472742404315,
      "tokens_seen": 1984299008
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1092471,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.014146327972412,
      "objective/train/theoretical_loss": 3.432637807153228,
      "objective/train/tokens_used": 343223776,
      "theoretical_loss": 3.432637807153228,
      "tokens_seen": 1984364544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055757575757576,
      "loss": 2.7457,
      "theoretical_loss": 3.432637807153228,
      "tokens_seen": 1984364544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040555555555555554,
      "loss": 2.7626,
      "theoretical_loss": 3.4326283404662226,
      "tokens_seen": 1984430080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040553535353535355,
      "loss": 2.7682,
      "theoretical_loss": 3.4326188741793864,
      "tokens_seen": 1984495616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055151515151515,
      "loss": 2.6067,
      "theoretical_loss": 3.432609408292689,
      "tokens_seen": 1984561152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054949494949495,
      "loss": 2.6628,
      "theoretical_loss": 3.4325999428060987,
      "tokens_seen": 1984626688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054747474747475,
      "loss": 2.8009,
      "theoretical_loss": 3.432590477719587,
      "tokens_seen": 1984692224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054545454545455,
      "loss": 2.7039,
      "theoretical_loss": 3.432581013033124,
      "tokens_seen": 1984757760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040543434343434344,
      "loss": 2.6553,
      "theoretical_loss": 3.4325715487466777,
      "tokens_seen": 1984823296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054141414141414,
      "loss": 2.5869,
      "theoretical_loss": 3.43256208486022,
      "tokens_seen": 1984888832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053939393939394,
      "loss": 2.6068,
      "theoretical_loss": 3.43255262137372,
      "tokens_seen": 1984954368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040537373737373737,
      "loss": 2.5967,
      "theoretical_loss": 3.4325431582871473,
      "tokens_seen": 1985019904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053535353535354,
      "loss": 2.4943,
      "theoretical_loss": 3.432533695600472,
      "tokens_seen": 1985085440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040533333333333334,
      "loss": 2.9349,
      "theoretical_loss": 3.4325242333136643,
      "tokens_seen": 1985150976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053131313131313,
      "loss": 2.7874,
      "theoretical_loss": 3.4325147714266944,
      "tokens_seen": 1985216512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052929292929293,
      "loss": 2.7921,
      "theoretical_loss": 3.432505309939531,
      "tokens_seen": 1985282048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052727272727273,
      "loss": 2.419,
      "theoretical_loss": 3.432495848852145,
      "tokens_seen": 1985347584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052525252525253,
      "loss": 2.7875,
      "theoretical_loss": 3.432486388164506,
      "tokens_seen": 1985413120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040523232323232323,
      "loss": 2.7653,
      "theoretical_loss": 3.4324769278765834,
      "tokens_seen": 1985478656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052121212121212,
      "loss": 2.9169,
      "theoretical_loss": 3.432467467988348,
      "tokens_seen": 1985544192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051919191919192,
      "loss": 2.6095,
      "theoretical_loss": 3.4324580084997693,
      "tokens_seen": 1985609728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051717171717172,
      "loss": 2.718,
      "theoretical_loss": 3.4324485494108177,
      "tokens_seen": 1985675264
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040515151515151517,
      "loss": 2.6026,
      "theoretical_loss": 3.4324390907214624,
      "tokens_seen": 1985740800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051313131313131,
      "loss": 2.6342,
      "theoretical_loss": 3.4324296324316736,
      "tokens_seen": 1985806336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051111111111111,
      "loss": 2.5302,
      "theoretical_loss": 3.432420174541422,
      "tokens_seen": 1985871872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040509090909090915,
      "loss": 2.8093,
      "theoretical_loss": 3.4324107170506757,
      "tokens_seen": 1985937408
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1093318,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.581864595413208,
      "objective/train/theoretical_loss": 3.4324012599594065,
      "objective/train/tokens_used": 344862176,
      "theoretical_loss": 3.4324012599594065,
      "tokens_seen": 1986002944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050707070707071,
      "loss": 2.4938,
      "theoretical_loss": 3.4324012599594065,
      "tokens_seen": 1986002944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040505050505050506,
      "loss": 2.6114,
      "theoretical_loss": 3.432391803267583,
      "tokens_seen": 1986068480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405030303030303,
      "loss": 2.6796,
      "theoretical_loss": 3.4323823469751766,
      "tokens_seen": 1986134016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040501010101010097,
      "loss": 2.6682,
      "theoretical_loss": 3.432372891082156,
      "tokens_seen": 1986199552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040498989898989904,
      "loss": 2.6571,
      "theoretical_loss": 3.4323634355884916,
      "tokens_seen": 1986265088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404969696969697,
      "loss": 2.5284,
      "theoretical_loss": 3.4323539804941534,
      "tokens_seen": 1986330624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040494949494949495,
      "loss": 2.685,
      "theoretical_loss": 3.432344525799111,
      "tokens_seen": 1986396160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049292929292929,
      "loss": 2.7383,
      "theoretical_loss": 3.432335071503335,
      "tokens_seen": 1986461696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049090909090909,
      "loss": 2.6443,
      "theoretical_loss": 3.4323256176067947,
      "tokens_seen": 1986527232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040488888888888893,
      "loss": 2.7833,
      "theoretical_loss": 3.4323161641094604,
      "tokens_seen": 1986592768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048686868686869,
      "loss": 2.7787,
      "theoretical_loss": 3.4323067110113024,
      "tokens_seen": 1986658304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040484848484848484,
      "loss": 2.7694,
      "theoretical_loss": 3.4322972583122904,
      "tokens_seen": 1986723840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048282828282828,
      "loss": 2.6806,
      "theoretical_loss": 3.4322878060123942,
      "tokens_seen": 1986789376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048080808080808,
      "loss": 2.8119,
      "theoretical_loss": 3.4322783541115838,
      "tokens_seen": 1986854912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047878787878788,
      "loss": 2.59,
      "theoretical_loss": 3.432268902609829,
      "tokens_seen": 1986920448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047676767676768,
      "loss": 2.721,
      "theoretical_loss": 3.4322594515071003,
      "tokens_seen": 1986985984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040474747474747474,
      "loss": 2.5832,
      "theoretical_loss": 3.432250000803368,
      "tokens_seen": 1987051520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040472727272727275,
      "loss": 2.7105,
      "theoretical_loss": 3.4322405504986007,
      "tokens_seen": 1987117056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047070707070707,
      "loss": 2.779,
      "theoretical_loss": 3.4322311005927704,
      "tokens_seen": 1987182592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046868686868687,
      "loss": 2.8089,
      "theoretical_loss": 3.432221651085845,
      "tokens_seen": 1987248128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040466666666666667,
      "loss": 2.6261,
      "theoretical_loss": 3.432212201977796,
      "tokens_seen": 1987313664
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040464646464646463,
      "loss": 2.5074,
      "theoretical_loss": 3.4322027532685926,
      "tokens_seen": 1987379200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040462626262626264,
      "loss": 2.4476,
      "theoretical_loss": 3.432193304958205,
      "tokens_seen": 1987444736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046060606060606,
      "loss": 2.7534,
      "theoretical_loss": 3.4321838570466037,
      "tokens_seen": 1987510272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045858585858586,
      "loss": 2.6793,
      "theoretical_loss": 3.432174409533758,
      "tokens_seen": 1987575808
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1094373,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.851362943649292,
      "objective/train/theoretical_loss": 3.4321649624196384,
      "objective/train/tokens_used": 346500576,
      "theoretical_loss": 3.4321649624196384,
      "tokens_seen": 1987641344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040456565656565657,
      "loss": 2.5657,
      "theoretical_loss": 3.4321649624196384,
      "tokens_seen": 1987641344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045454545454546,
      "loss": 2.3815,
      "theoretical_loss": 3.432155515704215,
      "tokens_seen": 1987706880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040452525252525253,
      "loss": 2.7743,
      "theoretical_loss": 3.432146069387458,
      "tokens_seen": 1987772416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045050505050505,
      "loss": 2.7407,
      "theoretical_loss": 3.4321366234693365,
      "tokens_seen": 1987837952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044848484848485,
      "loss": 2.5287,
      "theoretical_loss": 3.432127177949821,
      "tokens_seen": 1987903488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040446464646464646,
      "loss": 2.6323,
      "theoretical_loss": 3.432117732828882,
      "tokens_seen": 1987969024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040444444444444447,
      "loss": 2.6108,
      "theoretical_loss": 3.4321082881064884,
      "tokens_seen": 1988034560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044242424242424,
      "loss": 2.4885,
      "theoretical_loss": 3.4320988437826117,
      "tokens_seen": 1988100096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044040404040404,
      "loss": 2.5306,
      "theoretical_loss": 3.4320893998572215,
      "tokens_seen": 1988165632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004043838383838384,
      "loss": 2.5742,
      "theoretical_loss": 3.432079956330287,
      "tokens_seen": 1988231168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040436363636363635,
      "loss": 2.7162,
      "theoretical_loss": 3.4320705132017792,
      "tokens_seen": 1988296704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040434343434343436,
      "loss": 2.7681,
      "theoretical_loss": 3.4320610704716676,
      "tokens_seen": 1988362240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004043232323232323,
      "loss": 2.8476,
      "theoretical_loss": 3.4320516281399227,
      "tokens_seen": 1988427776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040430303030303033,
      "loss": 2.7233,
      "theoretical_loss": 3.4320421862065147,
      "tokens_seen": 1988493312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042828282828283,
      "loss": 2.949,
      "theoretical_loss": 3.432032744671413,
      "tokens_seen": 1988558848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042626262626263,
      "loss": 2.7295,
      "theoretical_loss": 3.4320233035345877,
      "tokens_seen": 1988624384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040424242424242425,
      "loss": 2.7111,
      "theoretical_loss": 3.4320138627960097,
      "tokens_seen": 1988689920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042222222222222,
      "loss": 2.7302,
      "theoretical_loss": 3.4320044224556483,
      "tokens_seen": 1988755456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042020202020202,
      "loss": 2.7109,
      "theoretical_loss": 3.4319949825134737,
      "tokens_seen": 1988820992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041818181818182,
      "loss": 2.9131,
      "theoretical_loss": 3.4319855429694566,
      "tokens_seen": 1988886528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041616161616162,
      "loss": 2.6445,
      "theoretical_loss": 3.4319761038235663,
      "tokens_seen": 1988952064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040414141414141415,
      "loss": 2.6379,
      "theoretical_loss": 3.431966665075773,
      "tokens_seen": 1989017600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041212121212121,
      "loss": 2.6443,
      "theoretical_loss": 3.4319572267260474,
      "tokens_seen": 1989083136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041010101010101,
      "loss": 2.6639,
      "theoretical_loss": 3.431947788774359,
      "tokens_seen": 1989148672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040808080808081,
      "loss": 2.9648,
      "theoretical_loss": 3.431938351220678,
      "tokens_seen": 1989214208
    },
    {
      "debugging/Self-BLEU-5": 0.4028481195939984,
      "debugging/distinct-1-grams": 0.8260221595150598,
      "debugging/distinct-2-grams": 0.9811327641385946,
      "debugging/entropy-1-grams": 5.420510909354855,
      "debugging/entropy-2-grams": 5.998582774683988,
      "debugging/length": 472.57142857142856,
      "debugging/num_segments": 7,
      "epoch": 0.2,
      "objective/train/docs_used": 1094923,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8087246417999268,
      "objective/train/theoretical_loss": 3.431928914064975,
      "objective/train/tokens_used": 348138976,
      "theoretical_loss": 3.431928914064975,
      "tokens_seen": 1989279744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040606060606061,
      "loss": 2.7047,
      "theoretical_loss": 3.431928914064975,
      "tokens_seen": 1989279744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040404040404040404,
      "loss": 2.6149,
      "theoretical_loss": 3.4319194773072192,
      "tokens_seen": 1989345280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404020202020202,
      "loss": 2.6278,
      "theoretical_loss": 3.4319100409473813,
      "tokens_seen": 1989410816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404,
      "loss": 2.6593,
      "theoretical_loss": 3.4319006049854313,
      "tokens_seen": 1989476352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403979797979798,
      "loss": 2.6405,
      "theoretical_loss": 3.4318911694213394,
      "tokens_seen": 1989541888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403959595959596,
      "loss": 2.745,
      "theoretical_loss": 3.431881734255076,
      "tokens_seen": 1989607424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040393939393939393,
      "loss": 2.841,
      "theoretical_loss": 3.431872299486611,
      "tokens_seen": 1989672960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039191919191919,
      "loss": 2.6359,
      "theoretical_loss": 3.4318628651159138,
      "tokens_seen": 1989738496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040389898989898995,
      "loss": 2.7627,
      "theoretical_loss": 3.4318534311429554,
      "tokens_seen": 1989804032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038787878787879,
      "loss": 2.4968,
      "theoretical_loss": 3.4318439975677055,
      "tokens_seen": 1989869568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040385858585858587,
      "loss": 2.5069,
      "theoretical_loss": 3.431834564390135,
      "tokens_seen": 1989935104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038383838383838,
      "loss": 2.7199,
      "theoretical_loss": 3.4318251316102124,
      "tokens_seen": 1990000640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038181818181818,
      "loss": 2.3488,
      "theoretical_loss": 3.4318156992279096,
      "tokens_seen": 1990066176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040379797979797985,
      "loss": 2.7606,
      "theoretical_loss": 3.4318062672431964,
      "tokens_seen": 1990131712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037777777777778,
      "loss": 2.6489,
      "theoretical_loss": 3.431796835656042,
      "tokens_seen": 1990197248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040375757575757576,
      "loss": 2.8979,
      "theoretical_loss": 3.431787404466417,
      "tokens_seen": 1990262784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037373737373737,
      "loss": 2.6212,
      "theoretical_loss": 3.431777973674292,
      "tokens_seen": 1990328320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040371717171717173,
      "loss": 2.8688,
      "theoretical_loss": 3.431768543279637,
      "tokens_seen": 1990393856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040369696969696974,
      "loss": 2.858,
      "theoretical_loss": 3.431759113282422,
      "tokens_seen": 1990459392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036767676767677,
      "loss": 2.5182,
      "theoretical_loss": 3.431749683682617,
      "tokens_seen": 1990524928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040365656565656565,
      "loss": 2.5937,
      "theoretical_loss": 3.4317402544801925,
      "tokens_seen": 1990590464
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036363636363636,
      "loss": 2.6897,
      "theoretical_loss": 3.4317308256751184,
      "tokens_seen": 1990656000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036161616161616,
      "loss": 2.7109,
      "theoretical_loss": 3.4317213972673652,
      "tokens_seen": 1990721536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040359595959595963,
      "loss": 2.7822,
      "theoretical_loss": 3.4317119692569023,
      "tokens_seen": 1990787072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035757575757576,
      "loss": 2.7954,
      "theoretical_loss": 3.431702541643701,
      "tokens_seen": 1990852608
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1096094,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.850614547729492,
      "objective/train/theoretical_loss": 3.431693114427731,
      "objective/train/tokens_used": 349777376,
      "theoretical_loss": 3.431693114427731,
      "tokens_seen": 1990918144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040355555555555555,
      "loss": 2.7385,
      "theoretical_loss": 3.431693114427731,
      "tokens_seen": 1990918144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040353535353535356,
      "loss": 2.9012,
      "theoretical_loss": 3.4316836876089623,
      "tokens_seen": 1990983680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035151515151515,
      "loss": 2.6501,
      "theoretical_loss": 3.431674261187365,
      "tokens_seen": 1991049216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034949494949495,
      "loss": 2.6113,
      "theoretical_loss": 3.4316648351629095,
      "tokens_seen": 1991114752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034747474747475,
      "loss": 2.6352,
      "theoretical_loss": 3.4316554095355665,
      "tokens_seen": 1991180288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040345454545454544,
      "loss": 2.5908,
      "theoretical_loss": 3.4316459843053053,
      "tokens_seen": 1991245824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040343434343434345,
      "loss": 2.466,
      "theoretical_loss": 3.4316365594720963,
      "tokens_seen": 1991311360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034141414141414,
      "loss": 2.9378,
      "theoretical_loss": 3.43162713503591,
      "tokens_seen": 1991376896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033939393939394,
      "loss": 2.8731,
      "theoretical_loss": 3.431617710996717,
      "tokens_seen": 1991442432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033737373737374,
      "loss": 2.5793,
      "theoretical_loss": 3.431608287354487,
      "tokens_seen": 1991507968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033535353535354,
      "loss": 2.8316,
      "theoretical_loss": 3.4315988641091906,
      "tokens_seen": 1991573504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040333333333333334,
      "loss": 2.7059,
      "theoretical_loss": 3.4315894412607966,
      "tokens_seen": 1991639040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033131313131313,
      "loss": 2.9928,
      "theoretical_loss": 3.431580018809277,
      "tokens_seen": 1991704576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032929292929293,
      "loss": 2.709,
      "theoretical_loss": 3.431570596754601,
      "tokens_seen": 1991770112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040327272727272727,
      "loss": 2.7009,
      "theoretical_loss": 3.4315611750967396,
      "tokens_seen": 1991835648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032525252525253,
      "loss": 2.4751,
      "theoretical_loss": 3.431551753835662,
      "tokens_seen": 1991901184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040323232323232324,
      "loss": 2.5752,
      "theoretical_loss": 3.4315423329713397,
      "tokens_seen": 1991966720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032121212121212,
      "loss": 2.6393,
      "theoretical_loss": 3.4315329125037417,
      "tokens_seen": 1992032256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031919191919192,
      "loss": 2.8421,
      "theoretical_loss": 3.4315234924328393,
      "tokens_seen": 1992097792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031717171717172,
      "loss": 2.65,
      "theoretical_loss": 3.431514072758602,
      "tokens_seen": 1992163328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040315151515151517,
      "loss": 2.6166,
      "theoretical_loss": 3.4315046534810003,
      "tokens_seen": 1992228864
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040313131313131313,
      "loss": 2.6158,
      "theoretical_loss": 3.4314952346000047,
      "tokens_seen": 1992294400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031111111111111,
      "loss": 2.7504,
      "theoretical_loss": 3.4314858161155852,
      "tokens_seen": 1992359936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030909090909091,
      "loss": 2.5385,
      "theoretical_loss": 3.431476398027712,
      "tokens_seen": 1992425472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030707070707071,
      "loss": 2.4957,
      "theoretical_loss": 3.431466980336355,
      "tokens_seen": 1992491008
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1096583,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7095284461975098,
      "objective/train/theoretical_loss": 3.4314575630414854,
      "objective/train/tokens_used": 351415776,
      "theoretical_loss": 3.4314575630414854,
      "tokens_seen": 1992556544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040305050505050506,
      "loss": 2.6214,
      "theoretical_loss": 3.4314575630414854,
      "tokens_seen": 1992556544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403030303030303,
      "loss": 2.807,
      "theoretical_loss": 3.4314481461430724,
      "tokens_seen": 1992622080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403010101010101,
      "loss": 2.634,
      "theoretical_loss": 3.4314387296410875,
      "tokens_seen": 1992687616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040298989898989904,
      "loss": 2.7925,
      "theoretical_loss": 3.4314293135355003,
      "tokens_seen": 1992753152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402969696969697,
      "loss": 2.4985,
      "theoretical_loss": 3.431419897826281,
      "tokens_seen": 1992818688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040294949494949496,
      "loss": 2.8183,
      "theoretical_loss": 3.4314104825133995,
      "tokens_seen": 1992884224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004029292929292929,
      "loss": 2.7924,
      "theoretical_loss": 3.431401067596827,
      "tokens_seen": 1992949760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040290909090909087,
      "loss": 2.6961,
      "theoretical_loss": 3.431391653076533,
      "tokens_seen": 1993015296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040288888888888894,
      "loss": 2.4038,
      "theoretical_loss": 3.431382238952488,
      "tokens_seen": 1993080832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028686868686869,
      "loss": 2.7434,
      "theoretical_loss": 3.4313728252246634,
      "tokens_seen": 1993146368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040284848484848485,
      "loss": 2.718,
      "theoretical_loss": 3.4313634118930274,
      "tokens_seen": 1993211904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028282828282828,
      "loss": 2.7766,
      "theoretical_loss": 3.431353998957552,
      "tokens_seen": 1993277440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040280808080808076,
      "loss": 2.7098,
      "theoretical_loss": 3.4313445864182066,
      "tokens_seen": 1993342976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040278787878787883,
      "loss": 2.7335,
      "theoretical_loss": 3.4313351742749623,
      "tokens_seen": 1993408512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027676767676768,
      "loss": 2.6668,
      "theoretical_loss": 3.4313257625277886,
      "tokens_seen": 1993474048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040274747474747474,
      "loss": 2.5202,
      "theoretical_loss": 3.4313163511766565,
      "tokens_seen": 1993539584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027272727272727,
      "loss": 2.7599,
      "theoretical_loss": 3.4313069402215355,
      "tokens_seen": 1993605120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027070707070707,
      "loss": 2.6586,
      "theoretical_loss": 3.431297529662397,
      "tokens_seen": 1993670656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026868686868687,
      "loss": 2.8594,
      "theoretical_loss": 3.43128811949921,
      "tokens_seen": 1993736192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026666666666667,
      "loss": 2.6204,
      "theoretical_loss": 3.431278709731946,
      "tokens_seen": 1993801728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040264646464646464,
      "loss": 2.6484,
      "theoretical_loss": 3.431269300360575,
      "tokens_seen": 1993867264
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026262626262626,
      "loss": 2.6423,
      "theoretical_loss": 3.431259891385067,
      "tokens_seen": 1993932800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026060606060606,
      "loss": 2.6703,
      "theoretical_loss": 3.431250482805392,
      "tokens_seen": 1993998336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025858585858586,
      "loss": 2.5785,
      "theoretical_loss": 3.4312410746215214,
      "tokens_seen": 1994063872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040256565656565657,
      "loss": 2.655,
      "theoretical_loss": 3.4312316668334253,
      "tokens_seen": 1994129408
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1097915,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4799556732177734,
      "objective/train/theoretical_loss": 3.431222259441073,
      "objective/train/tokens_used": 353054176,
      "theoretical_loss": 3.431222259441073,
      "tokens_seen": 1994194944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040254545454545453,
      "loss": 2.6861,
      "theoretical_loss": 3.431222259441073,
      "tokens_seen": 1994194944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040252525252525254,
      "loss": 2.6419,
      "theoretical_loss": 3.4312128524444363,
      "tokens_seen": 1994260480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040250505050505055,
      "loss": 2.548,
      "theoretical_loss": 3.431203445843485,
      "tokens_seen": 1994326016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024848484848485,
      "loss": 2.5519,
      "theoretical_loss": 3.4311940396381884,
      "tokens_seen": 1994391552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040246464646464646,
      "loss": 2.6708,
      "theoretical_loss": 3.4311846338285186,
      "tokens_seen": 1994457088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024444444444444,
      "loss": 2.6083,
      "theoretical_loss": 3.4311752284144443,
      "tokens_seen": 1994522624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040242424242424243,
      "loss": 2.5466,
      "theoretical_loss": 3.4311658233959372,
      "tokens_seen": 1994588160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040240404040404044,
      "loss": 2.5841,
      "theoretical_loss": 3.4311564187729675,
      "tokens_seen": 1994653696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023838383838384,
      "loss": 2.6176,
      "theoretical_loss": 3.431147014545505,
      "tokens_seen": 1994719232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040236363636363636,
      "loss": 2.6555,
      "theoretical_loss": 3.4311376107135203,
      "tokens_seen": 1994784768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040234343434343437,
      "loss": 2.681,
      "theoretical_loss": 3.4311282072769833,
      "tokens_seen": 1994850304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023232323232323,
      "loss": 2.5542,
      "theoretical_loss": 3.4311188042358656,
      "tokens_seen": 1994915840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040230303030303034,
      "loss": 2.8682,
      "theoretical_loss": 3.4311094015901364,
      "tokens_seen": 1994981376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022828282828283,
      "loss": 2.7279,
      "theoretical_loss": 3.431099999339767,
      "tokens_seen": 1995046912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040226262626262625,
      "loss": 2.8612,
      "theoretical_loss": 3.431090597484727,
      "tokens_seen": 1995112448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040224242424242426,
      "loss": 2.6292,
      "theoretical_loss": 3.4310811960249867,
      "tokens_seen": 1995177984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022222222222222,
      "loss": 2.5724,
      "theoretical_loss": 3.431071794960517,
      "tokens_seen": 1995243520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040220202020202023,
      "loss": 2.6812,
      "theoretical_loss": 3.4310623942912892,
      "tokens_seen": 1995309056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021818181818182,
      "loss": 2.9322,
      "theoretical_loss": 3.431052994017272,
      "tokens_seen": 1995374592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021616161616162,
      "loss": 2.7268,
      "theoretical_loss": 3.431043594138436,
      "tokens_seen": 1995440128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040214141414141415,
      "loss": 2.7061,
      "theoretical_loss": 3.431034194654753,
      "tokens_seen": 1995505664
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021212121212121,
      "loss": 2.7505,
      "theoretical_loss": 3.4310247955661923,
      "tokens_seen": 1995571200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021010101010101,
      "loss": 2.5533,
      "theoretical_loss": 3.4310153968727244,
      "tokens_seen": 1995636736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020808080808081,
      "loss": 2.5434,
      "theoretical_loss": 3.43100599857432,
      "tokens_seen": 1995702272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020606060606061,
      "loss": 2.7473,
      "theoretical_loss": 3.4309966006709494,
      "tokens_seen": 1995767808
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1098508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5174620151519775,
      "objective/train/theoretical_loss": 3.4309872031625828,
      "objective/train/tokens_used": 354692576,
      "theoretical_loss": 3.4309872031625828,
      "tokens_seen": 1995833344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040204040404040405,
      "loss": 2.7295,
      "theoretical_loss": 3.4309872031625828,
      "tokens_seen": 1995833344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402020202020202,
      "loss": 2.3933,
      "theoretical_loss": 3.4309778060491913,
      "tokens_seen": 1995898880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402,
      "loss": 2.7622,
      "theoretical_loss": 3.4309684093307444,
      "tokens_seen": 1995964416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000401979797979798,
      "loss": 2.8072,
      "theoretical_loss": 3.430959013007213,
      "tokens_seen": 1996029952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000401959595959596,
      "loss": 2.4624,
      "theoretical_loss": 3.430949617078568,
      "tokens_seen": 1996095488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040193939393939394,
      "loss": 2.6096,
      "theoretical_loss": 3.4309402215447795,
      "tokens_seen": 1996161024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019191919191919,
      "loss": 2.8545,
      "theoretical_loss": 3.430930826405817,
      "tokens_seen": 1996226560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018989898989899,
      "loss": 2.6149,
      "theoretical_loss": 3.4309214316616528,
      "tokens_seen": 1996292096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018787878787879,
      "loss": 2.621,
      "theoretical_loss": 3.4309120373122557,
      "tokens_seen": 1996357632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018585858585859,
      "loss": 2.8016,
      "theoretical_loss": 3.4309026433575966,
      "tokens_seen": 1996423168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040183838383838383,
      "loss": 2.7612,
      "theoretical_loss": 3.4308932497976468,
      "tokens_seen": 1996488704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018181818181818,
      "loss": 2.5173,
      "theoretical_loss": 3.4308838566323754,
      "tokens_seen": 1996554240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040179797979797985,
      "loss": 2.5817,
      "theoretical_loss": 3.4308744638617537,
      "tokens_seen": 1996619776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017777777777778,
      "loss": 2.5235,
      "theoretical_loss": 3.4308650714857523,
      "tokens_seen": 1996685312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040175757575757577,
      "loss": 2.5699,
      "theoretical_loss": 3.4308556795043414,
      "tokens_seen": 1996750848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017373737373737,
      "loss": 2.7443,
      "theoretical_loss": 3.4308462879174915,
      "tokens_seen": 1996816384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017171717171717,
      "loss": 2.5264,
      "theoretical_loss": 3.4308368967251726,
      "tokens_seen": 1996881920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040169696969696975,
      "loss": 2.647,
      "theoretical_loss": 3.4308275059273563,
      "tokens_seen": 1996947456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016767676767677,
      "loss": 2.6705,
      "theoretical_loss": 3.430818115524012,
      "tokens_seen": 1997012992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040165656565656566,
      "loss": 2.7531,
      "theoretical_loss": 3.430808725515111,
      "tokens_seen": 1997078528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016363636363636,
      "loss": 2.7024,
      "theoretical_loss": 3.430799335900623,
      "tokens_seen": 1997144064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040161616161616163,
      "loss": 2.7327,
      "theoretical_loss": 3.4307899466805187,
      "tokens_seen": 1997209600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040159595959595964,
      "loss": 2.746,
      "theoretical_loss": 3.430780557854769,
      "tokens_seen": 1997275136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015757575757576,
      "loss": 2.7495,
      "theoretical_loss": 3.4307711694233443,
      "tokens_seen": 1997340672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040155555555555555,
      "loss": 2.608,
      "theoretical_loss": 3.430761781386215,
      "tokens_seen": 1997406208
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 1099394,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8477187156677246,
      "objective/train/theoretical_loss": 3.430752393743351,
      "objective/train/tokens_used": 356330976,
      "theoretical_loss": 3.430752393743351,
      "tokens_seen": 1997471744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015353535353535,
      "loss": 2.63,
      "theoretical_loss": 3.430752393743351,
      "tokens_seen": 1997471744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004015151515151515,
      "loss": 2.8339,
      "theoretical_loss": 3.430743006494724,
      "tokens_seen": 1997537280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040149494949494953,
      "loss": 2.9765,
      "theoretical_loss": 3.4307336196403035,
      "tokens_seen": 1997602816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014747474747475,
      "loss": 2.9089,
      "theoretical_loss": 3.4307242331800607,
      "tokens_seen": 1997668352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040145454545454545,
      "loss": 2.7895,
      "theoretical_loss": 3.4307148471139657,
      "tokens_seen": 1997733888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040143434343434346,
      "loss": 2.7117,
      "theoretical_loss": 3.430705461441989,
      "tokens_seen": 1997799424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014141414141414,
      "loss": 2.6922,
      "theoretical_loss": 3.4306960761641014,
      "tokens_seen": 1997864960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013939393939394,
      "loss": 2.6537,
      "theoretical_loss": 3.4306866912802736,
      "tokens_seen": 1997930496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013737373737374,
      "loss": 2.6435,
      "theoretical_loss": 3.430677306790475,
      "tokens_seen": 1997996032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040135353535353534,
      "loss": 2.7962,
      "theoretical_loss": 3.4306679226946777,
      "tokens_seen": 1998061568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040133333333333335,
      "loss": 2.5895,
      "theoretical_loss": 3.430658538992851,
      "tokens_seen": 1998127104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013131313131313,
      "loss": 2.7134,
      "theoretical_loss": 3.430649155684966,
      "tokens_seen": 1998192640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012929292929293,
      "loss": 2.6309,
      "theoretical_loss": 3.4306397727709936,
      "tokens_seen": 1998258176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012727272727273,
      "loss": 2.573,
      "theoretical_loss": 3.4306303902509034,
      "tokens_seen": 1998323712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012525252525253,
      "loss": 2.8616,
      "theoretical_loss": 3.4306210081246666,
      "tokens_seen": 1998389248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040123232323232324,
      "loss": 2.587,
      "theoretical_loss": 3.430611626392254,
      "tokens_seen": 1998454784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012121212121212,
      "loss": 2.7992,
      "theoretical_loss": 3.4306022450536355,
      "tokens_seen": 1998520320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011919191919192,
      "loss": 2.5505,
      "theoretical_loss": 3.4305928641087817,
      "tokens_seen": 1998585856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040117171717171717,
      "loss": 2.7951,
      "theoretical_loss": 3.4305834835576636,
      "tokens_seen": 1998651392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011515151515152,
      "loss": 2.6681,
      "theoretical_loss": 3.430574103400252,
      "tokens_seen": 1998716928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040113131313131313,
      "loss": 2.5223,
      "theoretical_loss": 3.4305647236365164,
      "tokens_seen": 1998782464
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011111111111111,
      "loss": 2.7467,
      "theoretical_loss": 3.430555344266428,
      "tokens_seen": 1998848000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004010909090909091,
      "loss": 2.7558,
      "theoretical_loss": 3.4305459652899577,
      "tokens_seen": 1998913536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040107070707070706,
      "loss": 2.6164,
      "theoretical_loss": 3.4305365867070754,
      "tokens_seen": 1998979072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040105050505050507,
      "loss": 2.878,
      "theoretical_loss": 3.430527208517752,
      "tokens_seen": 1999044608
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1100643,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.998358726501465,
      "objective/train/theoretical_loss": 3.4305178307219584,
      "objective/train/tokens_used": 357969376,
      "theoretical_loss": 3.4305178307219584,
      "tokens_seen": 1999110144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401030303030303,
      "loss": 2.8925,
      "theoretical_loss": 3.4305178307219584,
      "tokens_seen": 1999110144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401010101010101,
      "loss": 2.8018,
      "theoretical_loss": 3.430508453319665,
      "tokens_seen": 1999175680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400989898989899,
      "loss": 2.7074,
      "theoretical_loss": 3.430499076310842,
      "tokens_seen": 1999241216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400969696969697,
      "loss": 2.6598,
      "theoretical_loss": 3.4304896996954604,
      "tokens_seen": 1999306752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040094949494949496,
      "loss": 2.555,
      "theoretical_loss": 3.4304803234734904,
      "tokens_seen": 1999372288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009292929292929,
      "loss": 2.7593,
      "theoretical_loss": 3.430470947644903,
      "tokens_seen": 1999437824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009090909090909,
      "loss": 3.0023,
      "theoretical_loss": 3.4304615722096687,
      "tokens_seen": 1999503360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008888888888889,
      "loss": 2.7062,
      "theoretical_loss": 3.4304521971677584,
      "tokens_seen": 1999568896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008686868686869,
      "loss": 2.711,
      "theoretical_loss": 3.430442822519142,
      "tokens_seen": 1999634432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040084848484848486,
      "loss": 2.6887,
      "theoretical_loss": 3.4304334482637904,
      "tokens_seen": 1999699968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008282828282828,
      "loss": 2.6289,
      "theoretical_loss": 3.4304240744016745,
      "tokens_seen": 1999765504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008080808080808,
      "loss": 2.6162,
      "theoretical_loss": 3.430414700932765,
      "tokens_seen": 1999831040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040078787878787883,
      "loss": 2.7507,
      "theoretical_loss": 3.4304053278570317,
      "tokens_seen": 1999896576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007676767676768,
      "loss": 2.6243,
      "theoretical_loss": 3.430395955174446,
      "tokens_seen": 1999962112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040074747474747475,
      "loss": 2.6822,
      "theoretical_loss": 3.4303865828849784,
      "tokens_seen": 2000027648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007272727272727,
      "loss": 2.5374,
      "theoretical_loss": 3.4303772109885995,
      "tokens_seen": 2000093184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007070707070707,
      "loss": 2.8277,
      "theoretical_loss": 3.43036783948528,
      "tokens_seen": 2000158720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040068686868686873,
      "loss": 2.9846,
      "theoretical_loss": 3.43035846837499,
      "tokens_seen": 2000224256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006666666666667,
      "loss": 2.718,
      "theoretical_loss": 3.4303490976577007,
      "tokens_seen": 2000289792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040064646464646464,
      "loss": 2.4615,
      "theoretical_loss": 3.430339727333383,
      "tokens_seen": 2000355328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006262626262626,
      "loss": 2.7228,
      "theoretical_loss": 3.4303303574020063,
      "tokens_seen": 2000420864
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040060606060606066,
      "loss": 2.5237,
      "theoretical_loss": 3.4303209878635426,
      "tokens_seen": 2000486400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005858585858586,
      "loss": 2.866,
      "theoretical_loss": 3.430311618717962,
      "tokens_seen": 2000551936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005656565656566,
      "loss": 2.6097,
      "theoretical_loss": 3.4303022499652354,
      "tokens_seen": 2000617472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040054545454545453,
      "loss": 2.7682,
      "theoretical_loss": 3.4302928816053324,
      "tokens_seen": 2000683008
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1101270,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.287954092025757,
      "objective/train/theoretical_loss": 3.4302835136382255,
      "objective/train/tokens_used": 359607776,
      "theoretical_loss": 3.4302835136382255,
      "tokens_seen": 2000748544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005252525252525,
      "loss": 2.6407,
      "theoretical_loss": 3.4302835136382255,
      "tokens_seen": 2000748544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040050505050505056,
      "loss": 2.6081,
      "theoretical_loss": 3.4302741460638844,
      "tokens_seen": 2000814080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004848484848485,
      "loss": 2.9468,
      "theoretical_loss": 3.430264778882279,
      "tokens_seen": 2000879616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040046464646464647,
      "loss": 2.6506,
      "theoretical_loss": 3.430255412093381,
      "tokens_seen": 2000945152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004444444444444,
      "loss": 2.7632,
      "theoretical_loss": 3.430246045697161,
      "tokens_seen": 2001010688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040042424242424244,
      "loss": 2.7678,
      "theoretical_loss": 3.4302366796935893,
      "tokens_seen": 2001076224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040040404040404045,
      "loss": 2.7699,
      "theoretical_loss": 3.4302273140826367,
      "tokens_seen": 2001141760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003838383838384,
      "loss": 2.8298,
      "theoretical_loss": 3.430217948864274,
      "tokens_seen": 2001207296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040036363636363636,
      "loss": 2.6612,
      "theoretical_loss": 3.430208584038472,
      "tokens_seen": 2001272832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003434343434343,
      "loss": 2.684,
      "theoretical_loss": 3.430199219605201,
      "tokens_seen": 2001338368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040032323232323233,
      "loss": 2.6532,
      "theoretical_loss": 3.430189855564432,
      "tokens_seen": 2001403904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040030303030303034,
      "loss": 2.8199,
      "theoretical_loss": 3.4301804919161354,
      "tokens_seen": 2001469440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004002828282828283,
      "loss": 2.7311,
      "theoretical_loss": 3.430171128660283,
      "tokens_seen": 2001534976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040026262626262625,
      "loss": 2.6164,
      "theoretical_loss": 3.4301617657968437,
      "tokens_seen": 2001600512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040024242424242427,
      "loss": 2.7514,
      "theoretical_loss": 3.4301524033257893,
      "tokens_seen": 2001666048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004002222222222222,
      "loss": 2.7445,
      "theoretical_loss": 3.4301430412470904,
      "tokens_seen": 2001731584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040020202020202023,
      "loss": 2.5826,
      "theoretical_loss": 3.4301336795607176,
      "tokens_seen": 2001797120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001818181818182,
      "loss": 2.6789,
      "theoretical_loss": 3.4301243182666417,
      "tokens_seen": 2001862656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040016161616161615,
      "loss": 2.6448,
      "theoretical_loss": 3.4301149573648333,
      "tokens_seen": 2001928192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040014141414141416,
      "loss": 2.6389,
      "theoretical_loss": 3.430105596855263,
      "tokens_seen": 2001993728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001212121212121,
      "loss": 2.4755,
      "theoretical_loss": 3.430096236737902,
      "tokens_seen": 2002059264
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001010101010101,
      "loss": 2.7416,
      "theoretical_loss": 3.4300868770127213,
      "tokens_seen": 2002124800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000808080808081,
      "loss": 2.7288,
      "theoretical_loss": 3.43007751767969,
      "tokens_seen": 2002190336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000606060606061,
      "loss": 2.7819,
      "theoretical_loss": 3.4300681587387807,
      "tokens_seen": 2002255872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040004040404040405,
      "loss": 2.7553,
      "theoretical_loss": 3.430058800189963,
      "tokens_seen": 2002321408
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1102668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.553368330001831,
      "objective/train/theoretical_loss": 3.430049442033208,
      "objective/train/tokens_used": 361246176,
      "theoretical_loss": 3.430049442033208,
      "tokens_seen": 2002386944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400020202020202,
      "loss": 2.6432,
      "theoretical_loss": 3.430049442033208,
      "tokens_seen": 2002386944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004,
      "loss": 2.7453,
      "theoretical_loss": 3.4300400842684864,
      "tokens_seen": 2002452480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399979797979798,
      "loss": 2.5433,
      "theoretical_loss": 3.4300307268957693,
      "tokens_seen": 2002518016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399959595959596,
      "loss": 2.645,
      "theoretical_loss": 3.430021369915027,
      "tokens_seen": 2002583552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039993939393939394,
      "loss": 2.7866,
      "theoretical_loss": 3.4300120133262304,
      "tokens_seen": 2002649088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999191919191919,
      "loss": 2.7484,
      "theoretical_loss": 3.43000265712935,
      "tokens_seen": 2002714624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998989898989899,
      "loss": 2.4926,
      "theoretical_loss": 3.429993301324357,
      "tokens_seen": 2002780160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998787878787879,
      "loss": 2.6038,
      "theoretical_loss": 3.4299839459112222,
      "tokens_seen": 2002845696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998585858585859,
      "loss": 2.7284,
      "theoretical_loss": 3.4299745908899157,
      "tokens_seen": 2002911232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039983838383838384,
      "loss": 2.8709,
      "theoretical_loss": 3.4299652362604087,
      "tokens_seen": 2002976768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998181818181818,
      "loss": 2.6466,
      "theoretical_loss": 3.4299558820226723,
      "tokens_seen": 2003042304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997979797979798,
      "loss": 2.5868,
      "theoretical_loss": 3.429946528176677,
      "tokens_seen": 2003107840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997777777777778,
      "loss": 2.6806,
      "theoretical_loss": 3.4299371747223937,
      "tokens_seen": 2003173376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039975757575757577,
      "loss": 2.7211,
      "theoretical_loss": 3.429927821659793,
      "tokens_seen": 2003238912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039973737373737373,
      "loss": 2.6575,
      "theoretical_loss": 3.429918468988845,
      "tokens_seen": 2003304448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997171717171717,
      "loss": 2.3557,
      "theoretical_loss": 3.4299091167095215,
      "tokens_seen": 2003369984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039969696969696975,
      "loss": 2.5517,
      "theoretical_loss": 3.4298997648217933,
      "tokens_seen": 2003435520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996767676767677,
      "loss": 2.876,
      "theoretical_loss": 3.42989041332563,
      "tokens_seen": 2003501056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039965656565656567,
      "loss": 2.6888,
      "theoretical_loss": 3.4298810622210043,
      "tokens_seen": 2003566592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996363636363636,
      "loss": 2.7384,
      "theoretical_loss": 3.4298717115078854,
      "tokens_seen": 2003632128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996161616161616,
      "loss": 2.4843,
      "theoretical_loss": 3.429862361186245,
      "tokens_seen": 2003697664
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039959595959595964,
      "loss": 2.6384,
      "theoretical_loss": 3.4298530112560535,
      "tokens_seen": 2003763200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995757575757576,
      "loss": 2.7286,
      "theoretical_loss": 3.429843661717282,
      "tokens_seen": 2003828736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039955555555555556,
      "loss": 2.8028,
      "theoretical_loss": 3.4298343125699002,
      "tokens_seen": 2003894272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995353535353535,
      "loss": 2.7432,
      "theoretical_loss": 3.4298249638138807,
      "tokens_seen": 2003959808
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1103172,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3909668922424316,
      "objective/train/theoretical_loss": 3.4298156154491934,
      "objective/train/tokens_used": 362884576,
      "theoretical_loss": 3.4298156154491934,
      "tokens_seen": 2004025344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039951515151515147,
      "loss": 2.5196,
      "theoretical_loss": 3.4298156154491934,
      "tokens_seen": 2004025344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039949494949494954,
      "loss": 2.7479,
      "theoretical_loss": 3.4298062674758087,
      "tokens_seen": 2004090880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994747474747475,
      "loss": 2.5021,
      "theoretical_loss": 3.429796919893698,
      "tokens_seen": 2004156416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039945454545454545,
      "loss": 2.791,
      "theoretical_loss": 3.4297875727028324,
      "tokens_seen": 2004221952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994343434343434,
      "loss": 2.8064,
      "theoretical_loss": 3.429778225903182,
      "tokens_seen": 2004287488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994141414141414,
      "loss": 2.7233,
      "theoretical_loss": 3.429768879494718,
      "tokens_seen": 2004353024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039939393939393943,
      "loss": 2.6588,
      "theoretical_loss": 3.4297595334774114,
      "tokens_seen": 2004418560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993737373737374,
      "loss": 2.6098,
      "theoretical_loss": 3.4297501878512326,
      "tokens_seen": 2004484096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039935353535353534,
      "loss": 2.7278,
      "theoretical_loss": 3.429740842616153,
      "tokens_seen": 2004549632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993333333333333,
      "loss": 2.5958,
      "theoretical_loss": 3.429731497772143,
      "tokens_seen": 2004615168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993131313131313,
      "loss": 2.8843,
      "theoretical_loss": 3.4297221533191737,
      "tokens_seen": 2004680704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992929292929293,
      "loss": 2.6212,
      "theoretical_loss": 3.429712809257216,
      "tokens_seen": 2004746240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992727272727273,
      "loss": 2.7094,
      "theoretical_loss": 3.4297034655862406,
      "tokens_seen": 2004811776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039925252525252524,
      "loss": 2.7256,
      "theoretical_loss": 3.429694122306218,
      "tokens_seen": 2004877312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039923232323232325,
      "loss": 2.8226,
      "theoretical_loss": 3.4296847794171197,
      "tokens_seen": 2004942848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992121212121212,
      "loss": 2.7651,
      "theoretical_loss": 3.4296754369189166,
      "tokens_seen": 2005008384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003991919191919192,
      "loss": 2.7283,
      "theoretical_loss": 3.4296660948115782,
      "tokens_seen": 2005073920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039917171717171717,
      "loss": 2.6783,
      "theoretical_loss": 3.4296567530950774,
      "tokens_seen": 2005139456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039915151515151513,
      "loss": 2.7694,
      "theoretical_loss": 3.429647411769384,
      "tokens_seen": 2005204992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039913131313131314,
      "loss": 2.855,
      "theoretical_loss": 3.4296380708344687,
      "tokens_seen": 2005270528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039911111111111115,
      "loss": 2.6749,
      "theoretical_loss": 3.429628730290303,
      "tokens_seen": 2005336064
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990909090909091,
      "loss": 2.4842,
      "theoretical_loss": 3.4296193901368577,
      "tokens_seen": 2005401600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039907070707070706,
      "loss": 2.9729,
      "theoretical_loss": 3.429610050374103,
      "tokens_seen": 2005467136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990505050505051,
      "loss": 2.7454,
      "theoretical_loss": 3.4296007110020104,
      "tokens_seen": 2005532672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039903030303030303,
      "loss": 2.6091,
      "theoretical_loss": 3.4295913720205506,
      "tokens_seen": 2005598208
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1104281,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4382286071777344,
      "objective/train/theoretical_loss": 3.4295820334296945,
      "objective/train/tokens_used": 364522976,
      "theoretical_loss": 3.4295820334296945,
      "tokens_seen": 2005663744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039901010101010104,
      "loss": 2.7794,
      "theoretical_loss": 3.4295820334296945,
      "tokens_seen": 2005663744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398989898989899,
      "loss": 2.5314,
      "theoretical_loss": 3.429572695229413,
      "tokens_seen": 2005729280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039896969696969696,
      "loss": 2.5465,
      "theoretical_loss": 3.429563357419677,
      "tokens_seen": 2005794816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039894949494949497,
      "loss": 2.6578,
      "theoretical_loss": 3.4295540200004577,
      "tokens_seen": 2005860352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003989292929292929,
      "loss": 2.7723,
      "theoretical_loss": 3.4295446829717258,
      "tokens_seen": 2005925888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039890909090909094,
      "loss": 2.5896,
      "theoretical_loss": 3.429535346333452,
      "tokens_seen": 2005991424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988888888888889,
      "loss": 2.4996,
      "theoretical_loss": 3.429526010085607,
      "tokens_seen": 2006056960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988686868686869,
      "loss": 2.8004,
      "theoretical_loss": 3.4295166742281626,
      "tokens_seen": 2006122496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039884848484848486,
      "loss": 2.5072,
      "theoretical_loss": 3.4295073387610895,
      "tokens_seen": 2006188032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988282828282828,
      "loss": 2.6983,
      "theoretical_loss": 3.4294980036843583,
      "tokens_seen": 2006253568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039880808080808083,
      "loss": 2.9369,
      "theoretical_loss": 3.4294886689979394,
      "tokens_seen": 2006319104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987878787878788,
      "loss": 2.5722,
      "theoretical_loss": 3.4294793347018047,
      "tokens_seen": 2006384640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987676767676768,
      "loss": 2.586,
      "theoretical_loss": 3.4294700007959245,
      "tokens_seen": 2006450176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039874747474747475,
      "loss": 2.8116,
      "theoretical_loss": 3.4294606672802703,
      "tokens_seen": 2006515712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987272727272727,
      "loss": 2.7575,
      "theoretical_loss": 3.4294513341548125,
      "tokens_seen": 2006581248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987070707070707,
      "loss": 2.8497,
      "theoretical_loss": 3.4294420014195226,
      "tokens_seen": 2006646784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039868686868686873,
      "loss": 2.8441,
      "theoretical_loss": 3.4294326690743713,
      "tokens_seen": 2006712320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986666666666667,
      "loss": 2.559,
      "theoretical_loss": 3.429423337119329,
      "tokens_seen": 2006777856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039864646464646465,
      "loss": 2.4556,
      "theoretical_loss": 3.429414005554367,
      "tokens_seen": 2006843392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986262626262626,
      "loss": 2.5888,
      "theoretical_loss": 3.429404674379457,
      "tokens_seen": 2006908928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986060606060606,
      "loss": 2.6556,
      "theoretical_loss": 3.429395343594569,
      "tokens_seen": 2006974464
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985858585858586,
      "loss": 2.6575,
      "theoretical_loss": 3.4293860131996743,
      "tokens_seen": 2007040000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985656565656566,
      "loss": 2.597,
      "theoretical_loss": 3.429376683194744,
      "tokens_seen": 2007105536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039854545454545454,
      "loss": 2.6176,
      "theoretical_loss": 3.4293673535797486,
      "tokens_seen": 2007171072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985252525252525,
      "loss": 2.7297,
      "theoretical_loss": 3.4293580243546598,
      "tokens_seen": 2007236608
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1105495,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7771756649017334,
      "objective/train/theoretical_loss": 3.4293486955194483,
      "objective/train/tokens_used": 366161376,
      "theoretical_loss": 3.4293486955194483,
      "tokens_seen": 2007302144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039850505050505056,
      "loss": 2.7023,
      "theoretical_loss": 3.4293486955194483,
      "tokens_seen": 2007302144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984848484848485,
      "loss": 2.7546,
      "theoretical_loss": 3.4293393670740846,
      "tokens_seen": 2007367680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984646464646465,
      "loss": 2.6322,
      "theoretical_loss": 3.42933003901854,
      "tokens_seen": 2007433216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039844444444444443,
      "loss": 2.806,
      "theoretical_loss": 3.429320711352786,
      "tokens_seen": 2007498752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984242424242424,
      "loss": 2.5904,
      "theoretical_loss": 3.429311384076793,
      "tokens_seen": 2007564288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039840404040404045,
      "loss": 2.8624,
      "theoretical_loss": 3.4293020571905317,
      "tokens_seen": 2007629824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983838383838384,
      "loss": 2.6415,
      "theoretical_loss": 3.429292730693974,
      "tokens_seen": 2007695360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039836363636363637,
      "loss": 2.5293,
      "theoretical_loss": 3.42928340458709,
      "tokens_seen": 2007760896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983434343434343,
      "loss": 2.5056,
      "theoretical_loss": 3.4292740788698515,
      "tokens_seen": 2007826432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039832323232323234,
      "loss": 2.6948,
      "theoretical_loss": 3.4292647535422285,
      "tokens_seen": 2007891968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039830303030303035,
      "loss": 2.602,
      "theoretical_loss": 3.4292554286041934,
      "tokens_seen": 2007957504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982828282828283,
      "loss": 2.6045,
      "theoretical_loss": 3.429246104055716,
      "tokens_seen": 2008023040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039826262626262626,
      "loss": 2.8573,
      "theoretical_loss": 3.4292367798967676,
      "tokens_seen": 2008088576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982424242424242,
      "loss": 2.5131,
      "theoretical_loss": 3.4292274561273195,
      "tokens_seen": 2008154112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039822222222222223,
      "loss": 2.6732,
      "theoretical_loss": 3.4292181327473426,
      "tokens_seen": 2008219648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039820202020202024,
      "loss": 2.591,
      "theoretical_loss": 3.429208809756808,
      "tokens_seen": 2008285184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981818181818182,
      "loss": 2.5137,
      "theoretical_loss": 3.429199487155686,
      "tokens_seen": 2008350720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039816161616161615,
      "loss": 2.5927,
      "theoretical_loss": 3.4291901649439493,
      "tokens_seen": 2008416256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039814141414141416,
      "loss": 2.8069,
      "theoretical_loss": 3.429180843121567,
      "tokens_seen": 2008481792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981212121212121,
      "loss": 2.6222,
      "theoretical_loss": 3.4291715216885112,
      "tokens_seen": 2008547328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039810101010101013,
      "loss": 2.7161,
      "theoretical_loss": 3.429162200644753,
      "tokens_seen": 2008612864
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980808080808081,
      "loss": 2.7619,
      "theoretical_loss": 3.429152879990263,
      "tokens_seen": 2008678400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039806060606060605,
      "loss": 2.697,
      "theoretical_loss": 3.429143559725013,
      "tokens_seen": 2008743936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039804040404040406,
      "loss": 2.5508,
      "theoretical_loss": 3.429134239848973,
      "tokens_seen": 2008809472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398020202020202,
      "loss": 2.6974,
      "theoretical_loss": 3.4291249203621144,
      "tokens_seen": 2008875008
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1105995,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0905230045318604,
      "objective/train/theoretical_loss": 3.4291156012644084,
      "objective/train/tokens_used": 367799776,
      "theoretical_loss": 3.4291156012644084,
      "tokens_seen": 2008940544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398,
      "loss": 2.7571,
      "theoretical_loss": 3.4291156012644084,
      "tokens_seen": 2008940544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397979797979798,
      "loss": 2.4411,
      "theoretical_loss": 3.429106282555826,
      "tokens_seen": 2009006080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397959595959596,
      "loss": 2.5121,
      "theoretical_loss": 3.4290969642363387,
      "tokens_seen": 2009071616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039793939393939395,
      "loss": 2.521,
      "theoretical_loss": 3.429087646305917,
      "tokens_seen": 2009137152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003979191919191919,
      "loss": 2.6003,
      "theoretical_loss": 3.4290783287645317,
      "tokens_seen": 2009202688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978989898989899,
      "loss": 2.8013,
      "theoretical_loss": 3.429069011612155,
      "tokens_seen": 2009268224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978787878787879,
      "loss": 2.8218,
      "theoretical_loss": 3.429059694848757,
      "tokens_seen": 2009333760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978585858585859,
      "loss": 2.5237,
      "theoretical_loss": 3.4290503784743085,
      "tokens_seen": 2009399296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039783838383838384,
      "loss": 2.6163,
      "theoretical_loss": 3.429041062488782,
      "tokens_seen": 2009464832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978181818181818,
      "loss": 2.7317,
      "theoretical_loss": 3.429031746892147,
      "tokens_seen": 2009530368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977979797979798,
      "loss": 2.5755,
      "theoretical_loss": 3.429022431684375,
      "tokens_seen": 2009595904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039777777777777777,
      "loss": 2.5838,
      "theoretical_loss": 3.4290131168654385,
      "tokens_seen": 2009661440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977575757575758,
      "loss": 2.6078,
      "theoretical_loss": 3.4290038024353064,
      "tokens_seen": 2009726976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039773737373737373,
      "loss": 2.7659,
      "theoretical_loss": 3.4289944883939514,
      "tokens_seen": 2009792512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977171717171717,
      "loss": 2.5651,
      "theoretical_loss": 3.428985174741344,
      "tokens_seen": 2009858048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976969696969697,
      "loss": 2.3688,
      "theoretical_loss": 3.428975861477455,
      "tokens_seen": 2009923584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976767676767677,
      "loss": 2.7842,
      "theoretical_loss": 3.4289665486022565,
      "tokens_seen": 2009989120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039765656565656567,
      "loss": 2.6497,
      "theoretical_loss": 3.4289572361157186,
      "tokens_seen": 2010054656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039763636363636363,
      "loss": 2.5598,
      "theoretical_loss": 3.428947924017813,
      "tokens_seen": 2010120192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976161616161616,
      "loss": 2.7141,
      "theoretical_loss": 3.42893861230851,
      "tokens_seen": 2010185728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975959595959596,
      "loss": 2.6488,
      "theoretical_loss": 3.428929300987782,
      "tokens_seen": 2010251264
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975757575757576,
      "loss": 2.6941,
      "theoretical_loss": 3.428919990055599,
      "tokens_seen": 2010316800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039755555555555556,
      "loss": 2.6565,
      "theoretical_loss": 3.428910679511932,
      "tokens_seen": 2010382336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975353535353535,
      "loss": 2.5045,
      "theoretical_loss": 3.428901369356754,
      "tokens_seen": 2010447872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975151515151515,
      "loss": 2.8396,
      "theoretical_loss": 3.4288920595900336,
      "tokens_seen": 2010513408
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1107334,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7171061038970947,
      "objective/train/theoretical_loss": 3.4288827502117436,
      "objective/train/tokens_used": 369438176,
      "theoretical_loss": 3.4288827502117436,
      "tokens_seen": 2010578944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039749494949494954,
      "loss": 2.6364,
      "theoretical_loss": 3.4288827502117436,
      "tokens_seen": 2010578944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003974747474747475,
      "loss": 2.616,
      "theoretical_loss": 3.4288734412218544,
      "tokens_seen": 2010644480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039745454545454546,
      "loss": 2.9022,
      "theoretical_loss": 3.4288641326203377,
      "tokens_seen": 2010710016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003974343434343434,
      "loss": 2.6037,
      "theoretical_loss": 3.428854824407164,
      "tokens_seen": 2010775552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039741414141414137,
      "loss": 2.5749,
      "theoretical_loss": 3.428845516582305,
      "tokens_seen": 2010841088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039739393939393944,
      "loss": 2.8528,
      "theoretical_loss": 3.4288362091457314,
      "tokens_seen": 2010906624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973737373737374,
      "loss": 2.7218,
      "theoretical_loss": 3.428826902097415,
      "tokens_seen": 2010972160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039735353535353535,
      "loss": 2.8239,
      "theoretical_loss": 3.428817595437326,
      "tokens_seen": 2011037696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973333333333333,
      "loss": 2.7298,
      "theoretical_loss": 3.4288082891654366,
      "tokens_seen": 2011103232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039731313131313137,
      "loss": 2.6982,
      "theoretical_loss": 3.428798983281717,
      "tokens_seen": 2011168768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039729292929292933,
      "loss": 2.8055,
      "theoretical_loss": 3.4287896777861384,
      "tokens_seen": 2011234304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972727272727273,
      "loss": 2.5813,
      "theoretical_loss": 3.4287803726786734,
      "tokens_seen": 2011299840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039725252525252524,
      "loss": 2.6005,
      "theoretical_loss": 3.4287710679592913,
      "tokens_seen": 2011365376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972323232323232,
      "loss": 2.8153,
      "theoretical_loss": 3.428761763627964,
      "tokens_seen": 2011430912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039721212121212126,
      "loss": 2.2978,
      "theoretical_loss": 3.4287524596846635,
      "tokens_seen": 2011496448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971919191919192,
      "loss": 2.5187,
      "theoretical_loss": 3.42874315612936,
      "tokens_seen": 2011561984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971717171717172,
      "loss": 2.6217,
      "theoretical_loss": 3.4287338529620244,
      "tokens_seen": 2011627520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039715151515151513,
      "loss": 2.6869,
      "theoretical_loss": 3.428724550182629,
      "tokens_seen": 2011693056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039713131313131315,
      "loss": 2.6536,
      "theoretical_loss": 3.4287152477911436,
      "tokens_seen": 2011758592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039711111111111116,
      "loss": 2.6725,
      "theoretical_loss": 3.4287059457875406,
      "tokens_seen": 2011824128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003970909090909091,
      "loss": 2.7913,
      "theoretical_loss": 3.428696644171791,
      "tokens_seen": 2011889664
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039707070707070707,
      "loss": 2.7083,
      "theoretical_loss": 3.428687342943866,
      "tokens_seen": 2011955200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397050505050505,
      "loss": 2.7214,
      "theoretical_loss": 3.428678042103736,
      "tokens_seen": 2012020736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039703030303030304,
      "loss": 2.839,
      "theoretical_loss": 3.428668741651373,
      "tokens_seen": 2012086272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039701010101010105,
      "loss": 2.6332,
      "theoretical_loss": 3.4286594415867477,
      "tokens_seen": 2012151808
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1107920,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5978739261627197,
      "objective/train/theoretical_loss": 3.4286501419098316,
      "objective/train/tokens_used": 371076576,
      "theoretical_loss": 3.4286501419098316,
      "tokens_seen": 2012217344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000396989898989899,
      "loss": 2.5902,
      "theoretical_loss": 3.4286501419098316,
      "tokens_seen": 2012217344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039696969696969696,
      "loss": 2.9242,
      "theoretical_loss": 3.428640842620596,
      "tokens_seen": 2012282880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000396949494949495,
      "loss": 2.8167,
      "theoretical_loss": 3.4286315437190122,
      "tokens_seen": 2012348416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039692929292929293,
      "loss": 2.5715,
      "theoretical_loss": 3.4286222452050508,
      "tokens_seen": 2012413952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039690909090909094,
      "loss": 2.6004,
      "theoretical_loss": 3.428612947078684,
      "tokens_seen": 2012479488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968888888888889,
      "loss": 2.6496,
      "theoretical_loss": 3.428603649339882,
      "tokens_seen": 2012545024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039686868686868686,
      "loss": 2.3663,
      "theoretical_loss": 3.4285943519886164,
      "tokens_seen": 2012610560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039684848484848487,
      "loss": 2.7802,
      "theoretical_loss": 3.4285850550248584,
      "tokens_seen": 2012676096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968282828282828,
      "loss": 2.6398,
      "theoretical_loss": 3.42857575844858,
      "tokens_seen": 2012741632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039680808080808083,
      "loss": 2.457,
      "theoretical_loss": 3.4285664622597514,
      "tokens_seen": 2012807168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967878787878788,
      "loss": 2.5914,
      "theoretical_loss": 3.428557166458344,
      "tokens_seen": 2012872704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967676767676768,
      "loss": 2.4352,
      "theoretical_loss": 3.4285478710443296,
      "tokens_seen": 2012938240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039674747474747476,
      "loss": 2.3116,
      "theoretical_loss": 3.428538576017679,
      "tokens_seen": 2013003776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967272727272727,
      "loss": 2.6761,
      "theoretical_loss": 3.4285292813783634,
      "tokens_seen": 2013069312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039670707070707073,
      "loss": 2.6778,
      "theoretical_loss": 3.4285199871263545,
      "tokens_seen": 2013134848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966868686868687,
      "loss": 2.6637,
      "theoretical_loss": 3.4285106932616225,
      "tokens_seen": 2013200384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966666666666667,
      "loss": 2.6304,
      "theoretical_loss": 3.4285013997841403,
      "tokens_seen": 2013265920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039664646464646465,
      "loss": 2.6416,
      "theoretical_loss": 3.428492106693878,
      "tokens_seen": 2013331456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966262626262626,
      "loss": 2.6832,
      "theoretical_loss": 3.428482813990807,
      "tokens_seen": 2013396992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966060606060606,
      "loss": 2.7265,
      "theoretical_loss": 3.4284735216748983,
      "tokens_seen": 2013462528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039658585858585863,
      "loss": 2.2403,
      "theoretical_loss": 3.4284642297461243,
      "tokens_seen": 2013528064
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965656565656566,
      "loss": 2.5304,
      "theoretical_loss": 3.428454938204455,
      "tokens_seen": 2013593600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039654545454545454,
      "loss": 2.6138,
      "theoretical_loss": 3.4284456470498625,
      "tokens_seen": 2013659136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965252525252525,
      "loss": 2.2908,
      "theoretical_loss": 3.4284363562823175,
      "tokens_seen": 2013724672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965050505050505,
      "loss": 2.5686,
      "theoretical_loss": 3.4284270659017917,
      "tokens_seen": 2013790208
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 1109221,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5685839653015137,
      "objective/train/theoretical_loss": 3.4284177759082564,
      "objective/train/tokens_used": 372714976,
      "theoretical_loss": 3.4284177759082564,
      "tokens_seen": 2013855744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003964848484848485,
      "loss": 2.6221,
      "theoretical_loss": 3.4284177759082564,
      "tokens_seen": 2013855744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964646464646465,
      "loss": 2.5984,
      "theoretical_loss": 3.4284084863016826,
      "tokens_seen": 2013921280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039644444444444444,
      "loss": 2.4075,
      "theoretical_loss": 3.428399197082042,
      "tokens_seen": 2013986816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964242424242424,
      "loss": 2.5087,
      "theoretical_loss": 3.4283899082493052,
      "tokens_seen": 2014052352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039640404040404046,
      "loss": 2.5658,
      "theoretical_loss": 3.428380619803444,
      "tokens_seen": 2014117888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963838383838384,
      "loss": 2.7391,
      "theoretical_loss": 3.4283713317444295,
      "tokens_seen": 2014183424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963636363636364,
      "loss": 2.7968,
      "theoretical_loss": 3.4283620440722333,
      "tokens_seen": 2014248960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039634343434343433,
      "loss": 2.6103,
      "theoretical_loss": 3.4283527567868264,
      "tokens_seen": 2014314496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963232323232323,
      "loss": 2.7116,
      "theoretical_loss": 3.4283434698881807,
      "tokens_seen": 2014380032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039630303030303035,
      "loss": 2.7277,
      "theoretical_loss": 3.4283341833762666,
      "tokens_seen": 2014445568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962828282828283,
      "loss": 2.7381,
      "theoretical_loss": 3.4283248972510556,
      "tokens_seen": 2014511104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039626262626262627,
      "loss": 2.6238,
      "theoretical_loss": 3.42831561151252,
      "tokens_seen": 2014576640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962424242424242,
      "loss": 2.3142,
      "theoretical_loss": 3.42830632616063,
      "tokens_seen": 2014642176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962222222222222,
      "loss": 2.741,
      "theoretical_loss": 3.428297041195357,
      "tokens_seen": 2014707712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039620202020202024,
      "loss": 2.7352,
      "theoretical_loss": 3.4282877566166734,
      "tokens_seen": 2014773248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961818181818182,
      "loss": 2.6948,
      "theoretical_loss": 3.428278472424549,
      "tokens_seen": 2014838784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039616161616161616,
      "loss": 2.7039,
      "theoretical_loss": 3.4282691886189562,
      "tokens_seen": 2014904320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961414141414141,
      "loss": 2.4815,
      "theoretical_loss": 3.4282599051998663,
      "tokens_seen": 2014969856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961212121212121,
      "loss": 2.6714,
      "theoretical_loss": 3.4282506221672504,
      "tokens_seen": 2015035392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039610101010101014,
      "loss": 2.7503,
      "theoretical_loss": 3.4282413395210796,
      "tokens_seen": 2015100928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960808080808081,
      "loss": 2.5656,
      "theoretical_loss": 3.4282320572613254,
      "tokens_seen": 2015166464
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039606060606060605,
      "loss": 2.3771,
      "theoretical_loss": 3.428222775387959,
      "tokens_seen": 2015232000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396040404040404,
      "loss": 2.5924,
      "theoretical_loss": 3.4282134939009525,
      "tokens_seen": 2015297536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396020202020202,
      "loss": 2.5664,
      "theoretical_loss": 3.4282042128002765,
      "tokens_seen": 2015363072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039600000000000003,
      "loss": 2.6235,
      "theoretical_loss": 3.4281949320859026,
      "tokens_seen": 2015428608
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1110352,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.888545274734497,
      "objective/train/theoretical_loss": 3.4281856517578024,
      "objective/train/tokens_used": 374353376,
      "theoretical_loss": 3.4281856517578024,
      "tokens_seen": 2015494144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395979797979798,
      "loss": 2.527,
      "theoretical_loss": 3.4281856517578024,
      "tokens_seen": 2015494144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039595959595959594,
      "loss": 2.8353,
      "theoretical_loss": 3.4281763718159466,
      "tokens_seen": 2015559680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039593939393939396,
      "loss": 2.3749,
      "theoretical_loss": 3.428167092260307,
      "tokens_seen": 2015625216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959191919191919,
      "loss": 2.5688,
      "theoretical_loss": 3.4281578130908548,
      "tokens_seen": 2015690752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958989898989899,
      "loss": 2.6914,
      "theoretical_loss": 3.428148534307562,
      "tokens_seen": 2015756288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958787878787879,
      "loss": 2.6239,
      "theoretical_loss": 3.4281392559103994,
      "tokens_seen": 2015821824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039585858585858584,
      "loss": 2.3188,
      "theoretical_loss": 3.428129977899338,
      "tokens_seen": 2015887360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039583838383838385,
      "loss": 2.5494,
      "theoretical_loss": 3.4281207002743503,
      "tokens_seen": 2015952896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958181818181818,
      "loss": 2.5859,
      "theoretical_loss": 3.4281114230354066,
      "tokens_seen": 2016018432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957979797979798,
      "loss": 2.6417,
      "theoretical_loss": 3.428102146182479,
      "tokens_seen": 2016083968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039577777777777777,
      "loss": 2.3921,
      "theoretical_loss": 3.428092869715538,
      "tokens_seen": 2016149504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957575757575758,
      "loss": 2.8298,
      "theoretical_loss": 3.4280835936345566,
      "tokens_seen": 2016215040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039573737373737374,
      "loss": 2.6659,
      "theoretical_loss": 3.4280743179395046,
      "tokens_seen": 2016280576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957171717171717,
      "loss": 2.5375,
      "theoretical_loss": 3.4280650426303536,
      "tokens_seen": 2016346112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956969696969697,
      "loss": 2.6967,
      "theoretical_loss": 3.428055767707076,
      "tokens_seen": 2016411648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039567676767676767,
      "loss": 2.5046,
      "theoretical_loss": 3.428046493169643,
      "tokens_seen": 2016477184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956565656565657,
      "loss": 2.6307,
      "theoretical_loss": 3.428037219018025,
      "tokens_seen": 2016542720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039563636363636363,
      "loss": 2.6712,
      "theoretical_loss": 3.428027945252194,
      "tokens_seen": 2016608256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039561616161616164,
      "loss": 2.6101,
      "theoretical_loss": 3.4280186718721217,
      "tokens_seen": 2016673792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955959595959596,
      "loss": 2.6178,
      "theoretical_loss": 3.428009398877779,
      "tokens_seen": 2016739328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955757575757576,
      "loss": 2.6002,
      "theoretical_loss": 3.428000126269138,
      "tokens_seen": 2016804864
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039555555555555557,
      "loss": 2.4496,
      "theoretical_loss": 3.427990854046169,
      "tokens_seen": 2016870400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955353535353535,
      "loss": 2.6548,
      "theoretical_loss": 3.4279815822088455,
      "tokens_seen": 2016935936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039551515151515154,
      "loss": 2.4536,
      "theoretical_loss": 3.427972310757136,
      "tokens_seen": 2017001472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954949494949495,
      "loss": 2.6407,
      "theoretical_loss": 3.4279630396910146,
      "tokens_seen": 2017067008
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1110721,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7697877883911133,
      "objective/train/theoretical_loss": 3.4279537690104513,
      "objective/train/tokens_used": 375991776,
      "theoretical_loss": 3.4279537690104513,
      "tokens_seen": 2017132544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954747474747475,
      "loss": 2.5957,
      "theoretical_loss": 3.4279537690104513,
      "tokens_seen": 2017132544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039545454545454546,
      "loss": 2.5364,
      "theoretical_loss": 3.427944498715418,
      "tokens_seen": 2017198080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954343434343434,
      "loss": 2.8179,
      "theoretical_loss": 3.427935228805886,
      "tokens_seen": 2017263616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039541414141414143,
      "loss": 2.6458,
      "theoretical_loss": 3.4279259592818265,
      "tokens_seen": 2017329152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039539393939393944,
      "loss": 2.58,
      "theoretical_loss": 3.4279166901432117,
      "tokens_seen": 2017394688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953737373737374,
      "loss": 2.2848,
      "theoretical_loss": 3.427907421390012,
      "tokens_seen": 2017460224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039535353535353535,
      "loss": 2.5158,
      "theoretical_loss": 3.4278981530221997,
      "tokens_seen": 2017525760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953333333333333,
      "loss": 2.5589,
      "theoretical_loss": 3.4278888850397458,
      "tokens_seen": 2017591296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953131313131313,
      "loss": 2.5954,
      "theoretical_loss": 3.4278796174426223,
      "tokens_seen": 2017656832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039529292929292933,
      "loss": 2.7181,
      "theoretical_loss": 3.4278703502308003,
      "tokens_seen": 2017722368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952727272727273,
      "loss": 2.5333,
      "theoretical_loss": 3.427861083404251,
      "tokens_seen": 2017787904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039525252525252525,
      "loss": 2.4276,
      "theoretical_loss": 3.4278518169629466,
      "tokens_seen": 2017853440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952323232323232,
      "loss": 2.5602,
      "theoretical_loss": 3.4278425509068575,
      "tokens_seen": 2017918976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039521212121212127,
      "loss": 2.7124,
      "theoretical_loss": 3.4278332852359563,
      "tokens_seen": 2017984512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951919191919192,
      "loss": 2.6338,
      "theoretical_loss": 3.4278240199502137,
      "tokens_seen": 2018050048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951717171717172,
      "loss": 2.5743,
      "theoretical_loss": 3.427814755049601,
      "tokens_seen": 2018115584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039515151515151514,
      "loss": 2.7207,
      "theoretical_loss": 3.4278054905340913,
      "tokens_seen": 2018181120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951313131313131,
      "loss": 2.7285,
      "theoretical_loss": 3.4277962264036543,
      "tokens_seen": 2018246656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039511111111111116,
      "loss": 2.3587,
      "theoretical_loss": 3.427786962658262,
      "tokens_seen": 2018312192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950909090909091,
      "loss": 2.4289,
      "theoretical_loss": 3.427777699297886,
      "tokens_seen": 2018377728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950707070707071,
      "loss": 2.6449,
      "theoretical_loss": 3.4277684363224985,
      "tokens_seen": 2018443264
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039505050505050503,
      "loss": 2.3899,
      "theoretical_loss": 3.4277591737320696,
      "tokens_seen": 2018508800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039503030303030304,
      "loss": 2.6567,
      "theoretical_loss": 3.427749911526572,
      "tokens_seen": 2018574336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039501010101010105,
      "loss": 2.7126,
      "theoretical_loss": 3.427740649705976,
      "tokens_seen": 2018639872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394989898989899,
      "loss": 2.6026,
      "theoretical_loss": 3.4277313882702547,
      "tokens_seen": 2018705408
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1111912,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.503018856048584,
      "objective/train/theoretical_loss": 3.4277221272193783,
      "objective/train/tokens_used": 377630176,
      "theoretical_loss": 3.4277221272193783,
      "tokens_seen": 2018770944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039496969696969697,
      "loss": 2.6544,
      "theoretical_loss": 3.4277221272193783,
      "tokens_seen": 2018770944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949494949494949,
      "loss": 2.7159,
      "theoretical_loss": 3.4277128665533194,
      "tokens_seen": 2018836480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039492929292929294,
      "loss": 2.6576,
      "theoretical_loss": 3.427703606272048,
      "tokens_seen": 2018902016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039490909090909095,
      "loss": 2.9404,
      "theoretical_loss": 3.427694346375537,
      "tokens_seen": 2018967552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948888888888889,
      "loss": 2.5293,
      "theoretical_loss": 3.4276850868637574,
      "tokens_seen": 2019033088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039486868686868686,
      "loss": 2.7324,
      "theoretical_loss": 3.4276758277366803,
      "tokens_seen": 2019098624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039484848484848487,
      "loss": 2.5299,
      "theoretical_loss": 3.4276665689942782,
      "tokens_seen": 2019164160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039482828282828283,
      "loss": 2.6959,
      "theoretical_loss": 3.427657310636522,
      "tokens_seen": 2019229696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039480808080808084,
      "loss": 2.5669,
      "theoretical_loss": 3.4276480526633835,
      "tokens_seen": 2019295232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947878787878788,
      "loss": 2.7492,
      "theoretical_loss": 3.4276387950748335,
      "tokens_seen": 2019360768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039476767676767675,
      "loss": 2.3663,
      "theoretical_loss": 3.4276295378708452,
      "tokens_seen": 2019426304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039474747474747476,
      "loss": 2.4772,
      "theoretical_loss": 3.427620281051388,
      "tokens_seen": 2019491840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947272727272727,
      "loss": 2.5573,
      "theoretical_loss": 3.4276110246164353,
      "tokens_seen": 2019557376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039470707070707073,
      "loss": 2.5873,
      "theoretical_loss": 3.4276017685659577,
      "tokens_seen": 2019622912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946868686868687,
      "loss": 2.6352,
      "theoretical_loss": 3.4275925128999267,
      "tokens_seen": 2019688448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946666666666667,
      "loss": 2.3906,
      "theoretical_loss": 3.4275832576183145,
      "tokens_seen": 2019753984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039464646464646466,
      "loss": 2.5083,
      "theoretical_loss": 3.4275740027210917,
      "tokens_seen": 2019819520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946262626262626,
      "loss": 2.5144,
      "theoretical_loss": 3.4275647482082308,
      "tokens_seen": 2019885056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946060606060606,
      "loss": 2.6141,
      "theoretical_loss": 3.4275554940797033,
      "tokens_seen": 2019950592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945858585858586,
      "loss": 2.5868,
      "theoretical_loss": 3.42754624033548,
      "tokens_seen": 2020016128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945656565656566,
      "loss": 2.3907,
      "theoretical_loss": 3.427536986975533,
      "tokens_seen": 2020081664
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039454545454545455,
      "loss": 2.5512,
      "theoretical_loss": 3.427527733999834,
      "tokens_seen": 2020147200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945252525252525,
      "loss": 2.4261,
      "theoretical_loss": 3.427518481408354,
      "tokens_seen": 2020212736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945050505050505,
      "loss": 2.4513,
      "theoretical_loss": 3.427509229201066,
      "tokens_seen": 2020278272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944848484848485,
      "loss": 2.4909,
      "theoretical_loss": 3.4274999773779395,
      "tokens_seen": 2020343808
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1112399,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.484231472015381,
      "objective/train/theoretical_loss": 3.4274907259389478,
      "objective/train/tokens_used": 379268576,
      "theoretical_loss": 3.4274907259389478,
      "tokens_seen": 2020409344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944646464646465,
      "loss": 2.6462,
      "theoretical_loss": 3.4274907259389478,
      "tokens_seen": 2020409344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039444444444444444,
      "loss": 2.5147,
      "theoretical_loss": 3.427481474884061,
      "tokens_seen": 2020474880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944242424242424,
      "loss": 2.5727,
      "theoretical_loss": 3.4274722242132523,
      "tokens_seen": 2020540416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944040404040404,
      "loss": 2.646,
      "theoretical_loss": 3.4274629739264926,
      "tokens_seen": 2020605952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943838383838384,
      "loss": 2.3592,
      "theoretical_loss": 3.427453724023753,
      "tokens_seen": 2020671488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943636363636364,
      "loss": 2.5641,
      "theoretical_loss": 3.427444474505006,
      "tokens_seen": 2020737024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039434343434343434,
      "loss": 2.6167,
      "theoretical_loss": 3.427435225370222,
      "tokens_seen": 2020802560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943232323232323,
      "loss": 2.5598,
      "theoretical_loss": 3.427425976619374,
      "tokens_seen": 2020868096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943030303030303,
      "loss": 2.7719,
      "theoretical_loss": 3.427416728252433,
      "tokens_seen": 2020933632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942828282828283,
      "loss": 2.5257,
      "theoretical_loss": 3.4274074802693706,
      "tokens_seen": 2020999168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039426262626262627,
      "loss": 2.7512,
      "theoretical_loss": 3.427398232670158,
      "tokens_seen": 2021064704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039424242424242423,
      "loss": 2.4711,
      "theoretical_loss": 3.4273889854547672,
      "tokens_seen": 2021130240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942222222222222,
      "loss": 2.6363,
      "theoretical_loss": 3.4273797386231704,
      "tokens_seen": 2021195776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039420202020202025,
      "loss": 2.66,
      "theoretical_loss": 3.4273704921753385,
      "tokens_seen": 2021261312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941818181818182,
      "loss": 2.5107,
      "theoretical_loss": 3.427361246111243,
      "tokens_seen": 2021326848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039416161616161616,
      "loss": 2.5211,
      "theoretical_loss": 3.427352000430856,
      "tokens_seen": 2021392384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941414141414141,
      "loss": 2.6448,
      "theoretical_loss": 3.4273427551341493,
      "tokens_seen": 2021457920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941212121212121,
      "loss": 2.4784,
      "theoretical_loss": 3.427333510221094,
      "tokens_seen": 2021523456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039410101010101014,
      "loss": 2.6683,
      "theoretical_loss": 3.4273242656916616,
      "tokens_seen": 2021588992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940808080808081,
      "loss": 2.4676,
      "theoretical_loss": 3.4273150215458243,
      "tokens_seen": 2021654528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039406060606060606,
      "loss": 2.4458,
      "theoretical_loss": 3.4273057777835536,
      "tokens_seen": 2021720064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394040404040404,
      "loss": 2.4926,
      "theoretical_loss": 3.4272965344048214,
      "tokens_seen": 2021785600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394020202020202,
      "loss": 2.4265,
      "theoretical_loss": 3.427287291409599,
      "tokens_seen": 2021851136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039400000000000004,
      "loss": 2.54,
      "theoretical_loss": 3.4272780487978576,
      "tokens_seen": 2021916672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000393979797979798,
      "loss": 2.4683,
      "theoretical_loss": 3.42726880656957,
      "tokens_seen": 2021982208
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1113663,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.153092622756958,
      "objective/train/theoretical_loss": 3.4272595647247073,
      "objective/train/tokens_used": 380906976,
      "theoretical_loss": 3.4272595647247073,
      "tokens_seen": 2022047744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039395959595959595,
      "loss": 2.5295,
      "theoretical_loss": 3.4272595647247073,
      "tokens_seen": 2022047744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939393939393939,
      "loss": 2.4037,
      "theoretical_loss": 3.42725032326324,
      "tokens_seen": 2022113280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039391919191919197,
      "loss": 2.6336,
      "theoretical_loss": 3.4272410821851422,
      "tokens_seen": 2022178816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039389898989898993,
      "loss": 2.4541,
      "theoretical_loss": 3.427231841490384,
      "tokens_seen": 2022244352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003938787878787879,
      "loss": 2.4697,
      "theoretical_loss": 3.427222601178937,
      "tokens_seen": 2022309888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039385858585858584,
      "loss": 2.6019,
      "theoretical_loss": 3.4272133612507734,
      "tokens_seen": 2022375424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039383838383838385,
      "loss": 2.6577,
      "theoretical_loss": 3.427204121705864,
      "tokens_seen": 2022440960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039381818181818186,
      "loss": 2.6264,
      "theoretical_loss": 3.427194882544182,
      "tokens_seen": 2022506496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937979797979798,
      "loss": 2.5512,
      "theoretical_loss": 3.4271856437656982,
      "tokens_seen": 2022572032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937777777777778,
      "loss": 2.5572,
      "theoretical_loss": 3.4271764053703837,
      "tokens_seen": 2022637568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039375757575757574,
      "loss": 2.7579,
      "theoretical_loss": 3.4271671673582116,
      "tokens_seen": 2022703104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039373737373737375,
      "loss": 2.4911,
      "theoretical_loss": 3.4271579297291526,
      "tokens_seen": 2022768640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039371717171717176,
      "loss": 2.6171,
      "theoretical_loss": 3.4271486924831787,
      "tokens_seen": 2022834176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936969696969697,
      "loss": 2.635,
      "theoretical_loss": 3.4271394556202615,
      "tokens_seen": 2022899712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039367676767676767,
      "loss": 2.4454,
      "theoretical_loss": 3.4271302191403725,
      "tokens_seen": 2022965248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936565656565657,
      "loss": 2.5407,
      "theoretical_loss": 3.4271209830434843,
      "tokens_seen": 2023030784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039363636363636364,
      "loss": 2.4783,
      "theoretical_loss": 3.427111747329567,
      "tokens_seen": 2023096320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039361616161616165,
      "loss": 2.815,
      "theoretical_loss": 3.4271025119985943,
      "tokens_seen": 2023161856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935959595959596,
      "loss": 2.2448,
      "theoretical_loss": 3.4270932770505365,
      "tokens_seen": 2023227392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039357575757575756,
      "loss": 2.5646,
      "theoretical_loss": 3.4270840424853657,
      "tokens_seen": 2023292928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935555555555556,
      "loss": 2.5724,
      "theoretical_loss": 3.427074808303054,
      "tokens_seen": 2023358464
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039353535353535353,
      "loss": 2.5373,
      "theoretical_loss": 3.427065574503572,
      "tokens_seen": 2023424000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039351515151515154,
      "loss": 2.6068,
      "theoretical_loss": 3.427056341086893,
      "tokens_seen": 2023489536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934949494949495,
      "loss": 2.4802,
      "theoretical_loss": 3.427047108052988,
      "tokens_seen": 2023555072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934747474747475,
      "loss": 2.5053,
      "theoretical_loss": 3.427037875401828,
      "tokens_seen": 2023620608
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1114958,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.473273754119873,
      "objective/train/theoretical_loss": 3.4270286431333856,
      "objective/train/tokens_used": 382545376,
      "theoretical_loss": 3.4270286431333856,
      "tokens_seen": 2023686144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039345454545454547,
      "loss": 2.3698,
      "theoretical_loss": 3.4270286431333856,
      "tokens_seen": 2023686144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934343434343434,
      "loss": 2.5781,
      "theoretical_loss": 3.4270194112476324,
      "tokens_seen": 2023751680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039341414141414144,
      "loss": 2.6656,
      "theoretical_loss": 3.4270101797445403,
      "tokens_seen": 2023817216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933939393939394,
      "loss": 2.383,
      "theoretical_loss": 3.4270009486240807,
      "tokens_seen": 2023882752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933737373737374,
      "loss": 2.6131,
      "theoretical_loss": 3.4269917178862253,
      "tokens_seen": 2023948288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039335353535353536,
      "loss": 2.5708,
      "theoretical_loss": 3.426982487530947,
      "tokens_seen": 2024013824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933333333333333,
      "loss": 2.3579,
      "theoretical_loss": 3.426973257558216,
      "tokens_seen": 2024079360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039331313131313133,
      "loss": 2.57,
      "theoretical_loss": 3.4269640279680043,
      "tokens_seen": 2024144896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039329292929292934,
      "loss": 2.662,
      "theoretical_loss": 3.4269547987602844,
      "tokens_seen": 2024210432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932727272727273,
      "loss": 2.6302,
      "theoretical_loss": 3.4269455699350275,
      "tokens_seen": 2024275968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039325252525252525,
      "loss": 2.3562,
      "theoretical_loss": 3.4269363414922056,
      "tokens_seen": 2024341504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932323232323232,
      "loss": 2.6728,
      "theoretical_loss": 3.426927113431791,
      "tokens_seen": 2024407040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932121212121212,
      "loss": 2.5875,
      "theoretical_loss": 3.4269178857537543,
      "tokens_seen": 2024472576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039319191919191923,
      "loss": 2.8369,
      "theoretical_loss": 3.426908658458068,
      "tokens_seen": 2024538112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931717171717172,
      "loss": 2.7961,
      "theoretical_loss": 3.426899431544704,
      "tokens_seen": 2024603648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039315151515151515,
      "loss": 2.401,
      "theoretical_loss": 3.426890205013634,
      "tokens_seen": 2024669184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931313131313131,
      "loss": 2.5557,
      "theoretical_loss": 3.4268809788648293,
      "tokens_seen": 2024734720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039311111111111117,
      "loss": 2.4,
      "theoretical_loss": 3.426871753098262,
      "tokens_seen": 2024800256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930909090909091,
      "loss": 2.7028,
      "theoretical_loss": 3.4268625277139044,
      "tokens_seen": 2024865792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930707070707071,
      "loss": 2.5791,
      "theoretical_loss": 3.426853302711727,
      "tokens_seen": 2024931328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039305050505050504,
      "loss": 2.6614,
      "theoretical_loss": 3.4268440780917038,
      "tokens_seen": 2024996864
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000393030303030303,
      "loss": 2.5811,
      "theoretical_loss": 3.426834853853804,
      "tokens_seen": 2025062400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039301010101010106,
      "loss": 2.6949,
      "theoretical_loss": 3.4268256299980013,
      "tokens_seen": 2025127936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392989898989899,
      "loss": 2.6975,
      "theoretical_loss": 3.4268164065242663,
      "tokens_seen": 2025193472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392969696969697,
      "loss": 2.5428,
      "theoretical_loss": 3.4268071834325715,
      "tokens_seen": 2025259008
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1115580,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.886110544204712,
      "objective/train/theoretical_loss": 3.4267979607228893,
      "objective/train/tokens_used": 384183776,
      "theoretical_loss": 3.4267979607228893,
      "tokens_seen": 2025324544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039294949494949493,
      "loss": 2.5216,
      "theoretical_loss": 3.4267979607228893,
      "tokens_seen": 2025324544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929292929292929,
      "loss": 2.5705,
      "theoretical_loss": 3.42678873839519,
      "tokens_seen": 2025390080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039290909090909095,
      "loss": 2.735,
      "theoretical_loss": 3.4267795164494466,
      "tokens_seen": 2025455616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928888888888889,
      "loss": 2.6675,
      "theoretical_loss": 3.42677029488563,
      "tokens_seen": 2025521152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039286868686868687,
      "loss": 2.453,
      "theoretical_loss": 3.4267610737037133,
      "tokens_seen": 2025586688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928484848484848,
      "loss": 2.6699,
      "theoretical_loss": 3.426751852903667,
      "tokens_seen": 2025652224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039282828282828283,
      "loss": 2.5889,
      "theoretical_loss": 3.426742632485464,
      "tokens_seen": 2025717760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039280808080808085,
      "loss": 2.5707,
      "theoretical_loss": 3.4267334124490754,
      "tokens_seen": 2025783296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003927878787878788,
      "loss": 2.5891,
      "theoretical_loss": 3.426724192794473,
      "tokens_seen": 2025848832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039276767676767676,
      "loss": 2.5181,
      "theoretical_loss": 3.426714973521629,
      "tokens_seen": 2025914368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003927474747474747,
      "loss": 2.4398,
      "theoretical_loss": 3.4267057546305155,
      "tokens_seen": 2025979904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039272727272727273,
      "loss": 2.5029,
      "theoretical_loss": 3.426696536121104,
      "tokens_seen": 2026045440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039270707070707074,
      "loss": 2.6772,
      "theoretical_loss": 3.426687317993366,
      "tokens_seen": 2026110976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926868686868687,
      "loss": 2.4527,
      "theoretical_loss": 3.4266781002472744,
      "tokens_seen": 2026176512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039266666666666665,
      "loss": 2.5978,
      "theoretical_loss": 3.4266688828828,
      "tokens_seen": 2026242048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039264646464646466,
      "loss": 2.4343,
      "theoretical_loss": 3.4266596658999147,
      "tokens_seen": 2026307584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926262626262626,
      "loss": 2.6968,
      "theoretical_loss": 3.426650449298591,
      "tokens_seen": 2026373120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039260606060606063,
      "loss": 2.589,
      "theoretical_loss": 3.4266412330788008,
      "tokens_seen": 2026438656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925858585858586,
      "loss": 2.604,
      "theoretical_loss": 3.4266320172405154,
      "tokens_seen": 2026504192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039256565656565654,
      "loss": 2.5602,
      "theoretical_loss": 3.426622801783707,
      "tokens_seen": 2026569728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039254545454545456,
      "loss": 2.5963,
      "theoretical_loss": 3.426613586708347,
      "tokens_seen": 2026635264
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925252525252525,
      "loss": 2.4481,
      "theoretical_loss": 3.426604372014408,
      "tokens_seen": 2026700800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925050505050505,
      "loss": 2.5384,
      "theoretical_loss": 3.426595157701861,
      "tokens_seen": 2026766336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924848484848485,
      "loss": 2.496,
      "theoretical_loss": 3.426585943770679,
      "tokens_seen": 2026831872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924646464646465,
      "loss": 2.6017,
      "theoretical_loss": 3.426576730220833,
      "tokens_seen": 2026897408
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1116847,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7708523273468018,
      "objective/train/theoretical_loss": 3.4265675170522956,
      "objective/train/tokens_used": 385822176,
      "theoretical_loss": 3.4265675170522956,
      "tokens_seen": 2026962944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039244444444444445,
      "loss": 2.6355,
      "theoretical_loss": 3.4265675170522956,
      "tokens_seen": 2026962944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924242424242424,
      "loss": 2.7949,
      "theoretical_loss": 3.426558304265038,
      "tokens_seen": 2027028480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924040404040404,
      "loss": 2.6705,
      "theoretical_loss": 3.426549091859032,
      "tokens_seen": 2027094016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923838383838384,
      "loss": 2.7159,
      "theoretical_loss": 3.426539879834251,
      "tokens_seen": 2027159552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923636363636364,
      "loss": 2.689,
      "theoretical_loss": 3.4265306681906647,
      "tokens_seen": 2027225088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039234343434343434,
      "loss": 2.5202,
      "theoretical_loss": 3.4265214569282465,
      "tokens_seen": 2027290624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923232323232323,
      "loss": 2.5135,
      "theoretical_loss": 3.4265122460469675,
      "tokens_seen": 2027356160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923030303030303,
      "loss": 2.5385,
      "theoretical_loss": 3.4265030355468005,
      "tokens_seen": 2027421696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922828282828283,
      "loss": 2.635,
      "theoretical_loss": 3.4264938254277166,
      "tokens_seen": 2027487232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922626262626263,
      "loss": 2.5788,
      "theoretical_loss": 3.426484615689688,
      "tokens_seen": 2027552768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039224242424242423,
      "loss": 2.5681,
      "theoretical_loss": 3.4264754063326865,
      "tokens_seen": 2027618304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039222222222222225,
      "loss": 2.441,
      "theoretical_loss": 3.426466197356685,
      "tokens_seen": 2027683840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922020202020202,
      "loss": 2.5216,
      "theoretical_loss": 3.426456988761654,
      "tokens_seen": 2027749376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921818181818182,
      "loss": 2.4985,
      "theoretical_loss": 3.4264477805475657,
      "tokens_seen": 2027814912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039216161616161617,
      "loss": 2.5328,
      "theoretical_loss": 3.426438572714393,
      "tokens_seen": 2027880448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921414141414141,
      "loss": 2.3552,
      "theoretical_loss": 3.4264293652621065,
      "tokens_seen": 2027945984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039212121212121214,
      "loss": 2.5907,
      "theoretical_loss": 3.4264201581906795,
      "tokens_seen": 2028011520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039210101010101015,
      "loss": 2.4451,
      "theoretical_loss": 3.426410951500083,
      "tokens_seen": 2028077056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920808080808081,
      "loss": 2.6231,
      "theoretical_loss": 3.4264017451902893,
      "tokens_seen": 2028142592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039206060606060606,
      "loss": 2.6823,
      "theoretical_loss": 3.4263925392612697,
      "tokens_seen": 2028208128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392040404040404,
      "loss": 2.5544,
      "theoretical_loss": 3.426383333712997,
      "tokens_seen": 2028273664
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039202020202020203,
      "loss": 2.832,
      "theoretical_loss": 3.4263741285454428,
      "tokens_seen": 2028339200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039200000000000004,
      "loss": 2.3735,
      "theoretical_loss": 3.4263649237585794,
      "tokens_seen": 2028404736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000391979797979798,
      "loss": 2.6263,
      "theoretical_loss": 3.426355719352378,
      "tokens_seen": 2028470272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039195959595959596,
      "loss": 2.4902,
      "theoretical_loss": 3.4263465153268116,
      "tokens_seen": 2028535808
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1117380,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6172401905059814,
      "objective/train/theoretical_loss": 3.426337311681851,
      "objective/train/tokens_used": 387460576,
      "theoretical_loss": 3.426337311681851,
      "tokens_seen": 2028601344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919393939393939,
      "loss": 2.4793,
      "theoretical_loss": 3.426337311681851,
      "tokens_seen": 2028601344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000391919191919192,
      "loss": 2.4009,
      "theoretical_loss": 3.4263281084174695,
      "tokens_seen": 2028666880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039189898989898993,
      "loss": 2.492,
      "theoretical_loss": 3.4263189055336376,
      "tokens_seen": 2028732416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003918787878787879,
      "loss": 2.4839,
      "theoretical_loss": 3.426309703030328,
      "tokens_seen": 2028797952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039185858585858585,
      "loss": 2.7173,
      "theoretical_loss": 3.426300500907513,
      "tokens_seen": 2028863488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003918383838383838,
      "loss": 2.5024,
      "theoretical_loss": 3.4262912991651637,
      "tokens_seen": 2028929024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039181818181818187,
      "loss": 2.6256,
      "theoretical_loss": 3.426282097803253,
      "tokens_seen": 2028994560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917979797979798,
      "loss": 2.638,
      "theoretical_loss": 3.426272896821753,
      "tokens_seen": 2029060096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917777777777778,
      "loss": 2.3982,
      "theoretical_loss": 3.4262636962206345,
      "tokens_seen": 2029125632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039175757575757574,
      "loss": 2.3516,
      "theoretical_loss": 3.42625449599987,
      "tokens_seen": 2029191168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039173737373737375,
      "loss": 2.4793,
      "theoretical_loss": 3.4262452961594323,
      "tokens_seen": 2029256704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039171717171717176,
      "loss": 2.5811,
      "theoretical_loss": 3.4262360966992924,
      "tokens_seen": 2029322240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916969696969697,
      "loss": 2.6168,
      "theoretical_loss": 3.426226897619423,
      "tokens_seen": 2029387776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916767676767677,
      "loss": 2.4779,
      "theoretical_loss": 3.426217698919795,
      "tokens_seen": 2029453312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039165656565656563,
      "loss": 2.4986,
      "theoretical_loss": 3.426208500600382,
      "tokens_seen": 2029518848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039163636363636364,
      "loss": 2.7759,
      "theoretical_loss": 3.426199302661155,
      "tokens_seen": 2029584384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039161616161616166,
      "loss": 2.4451,
      "theoretical_loss": 3.4261901051020858,
      "tokens_seen": 2029649920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915959595959596,
      "loss": 2.4326,
      "theoretical_loss": 3.426180907923147,
      "tokens_seen": 2029715456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039157575757575757,
      "loss": 2.4658,
      "theoretical_loss": 3.4261717111243106,
      "tokens_seen": 2029780992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915555555555556,
      "loss": 2.4223,
      "theoretical_loss": 3.426162514705548,
      "tokens_seen": 2029846528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039153535353535354,
      "loss": 2.7919,
      "theoretical_loss": 3.426153318666832,
      "tokens_seen": 2029912064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039151515151515155,
      "loss": 2.6321,
      "theoretical_loss": 3.426144123008134,
      "tokens_seen": 2029977600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003914949494949495,
      "loss": 2.5509,
      "theoretical_loss": 3.426134927729427,
      "tokens_seen": 2030043136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039147474747474746,
      "loss": 2.5512,
      "theoretical_loss": 3.426125732830682,
      "tokens_seen": 2030108672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003914545454545455,
      "loss": 2.5388,
      "theoretical_loss": 3.426116538311871,
      "tokens_seen": 2030174208
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 1118494,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7309558391571045,
      "objective/train/theoretical_loss": 3.4261073441729666,
      "objective/train/tokens_used": 389098976,
      "theoretical_loss": 3.4261073441729666,
      "tokens_seen": 2030239744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039143434343434343,
      "loss": 2.6567,
      "theoretical_loss": 3.4261073441729666,
      "tokens_seen": 2030239744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039141414141414144,
      "loss": 2.4831,
      "theoretical_loss": 3.426098150413941,
      "tokens_seen": 2030305280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913939393939394,
      "loss": 2.6372,
      "theoretical_loss": 3.4260889570347652,
      "tokens_seen": 2030370816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913737373737374,
      "loss": 2.5891,
      "theoretical_loss": 3.426079764035413,
      "tokens_seen": 2030436352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039135353535353537,
      "loss": 2.5801,
      "theoretical_loss": 3.4260705714158544,
      "tokens_seen": 2030501888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913333333333333,
      "loss": 2.8033,
      "theoretical_loss": 3.426061379176063,
      "tokens_seen": 2030567424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039131313131313133,
      "loss": 2.4674,
      "theoretical_loss": 3.42605218731601,
      "tokens_seen": 2030632960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912929292929293,
      "loss": 2.469,
      "theoretical_loss": 3.426042995835668,
      "tokens_seen": 2030698496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912727272727273,
      "loss": 2.672,
      "theoretical_loss": 3.4260338047350087,
      "tokens_seen": 2030764032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039125252525252526,
      "loss": 2.5594,
      "theoretical_loss": 3.426024614014004,
      "tokens_seen": 2030829568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912323232323232,
      "loss": 2.5454,
      "theoretical_loss": 3.426015423672627,
      "tokens_seen": 2030895104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912121212121212,
      "loss": 2.48,
      "theoretical_loss": 3.426006233710848,
      "tokens_seen": 2030960640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911919191919192,
      "loss": 2.5707,
      "theoretical_loss": 3.425997044128641,
      "tokens_seen": 2031026176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911717171717172,
      "loss": 2.5987,
      "theoretical_loss": 3.425987854925977,
      "tokens_seen": 2031091712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039115151515151515,
      "loss": 2.5819,
      "theoretical_loss": 3.4259786661028278,
      "tokens_seen": 2031157248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911313131313131,
      "loss": 2.6741,
      "theoretical_loss": 3.4259694776591663,
      "tokens_seen": 2031222784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911111111111111,
      "loss": 2.6763,
      "theoretical_loss": 3.425960289594964,
      "tokens_seen": 2031288320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039109090909090913,
      "loss": 2.6218,
      "theoretical_loss": 3.4259511019101936,
      "tokens_seen": 2031353856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910707070707071,
      "loss": 2.5182,
      "theoretical_loss": 3.4259419146048264,
      "tokens_seen": 2031419392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039105050505050504,
      "loss": 2.7424,
      "theoretical_loss": 3.425932727678835,
      "tokens_seen": 2031484928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000391030303030303,
      "loss": 2.5233,
      "theoretical_loss": 3.425923541132191,
      "tokens_seen": 2031550464
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000391010101010101,
      "loss": 2.8571,
      "theoretical_loss": 3.4259143549648674,
      "tokens_seen": 2031616000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390989898989899,
      "loss": 2.4963,
      "theoretical_loss": 3.4259051691768354,
      "tokens_seen": 2031681536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390969696969697,
      "loss": 2.6932,
      "theoretical_loss": 3.425895983768068,
      "tokens_seen": 2031747072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039094949494949494,
      "loss": 2.5079,
      "theoretical_loss": 3.425886798738537,
      "tokens_seen": 2031812608
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1119152,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6228444576263428,
      "objective/train/theoretical_loss": 3.425877614088213,
      "objective/train/tokens_used": 390737376,
      "theoretical_loss": 3.425877614088213,
      "tokens_seen": 2031878144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003909292929292929,
      "loss": 2.5856,
      "theoretical_loss": 3.425877614088213,
      "tokens_seen": 2031878144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039090909090909096,
      "loss": 2.6037,
      "theoretical_loss": 3.4258684298170703,
      "tokens_seen": 2031943680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908888888888889,
      "loss": 2.8355,
      "theoretical_loss": 3.4258592459250803,
      "tokens_seen": 2032009216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039086868686868687,
      "loss": 2.6795,
      "theoretical_loss": 3.4258500624122146,
      "tokens_seen": 2032074752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039084848484848483,
      "loss": 2.496,
      "theoretical_loss": 3.4258408792784456,
      "tokens_seen": 2032140288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908282828282828,
      "loss": 2.5187,
      "theoretical_loss": 3.425831696523746,
      "tokens_seen": 2032205824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039080808080808085,
      "loss": 2.5352,
      "theoretical_loss": 3.425822514148087,
      "tokens_seen": 2032271360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907878787878788,
      "loss": 2.571,
      "theoretical_loss": 3.4258133321514412,
      "tokens_seen": 2032336896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039076767676767677,
      "loss": 2.4906,
      "theoretical_loss": 3.4258041505337813,
      "tokens_seen": 2032402432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907474747474747,
      "loss": 2.6286,
      "theoretical_loss": 3.425794969295078,
      "tokens_seen": 2032467968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039072727272727273,
      "loss": 2.4928,
      "theoretical_loss": 3.425785788435305,
      "tokens_seen": 2032533504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039070707070707074,
      "loss": 2.7237,
      "theoretical_loss": 3.4257766079544334,
      "tokens_seen": 2032599040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906868686868687,
      "loss": 2.4682,
      "theoretical_loss": 3.4257674278524353,
      "tokens_seen": 2032664576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039066666666666666,
      "loss": 2.624,
      "theoretical_loss": 3.4257582481292834,
      "tokens_seen": 2032730112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906464646464646,
      "loss": 2.5292,
      "theoretical_loss": 3.42574906878495,
      "tokens_seen": 2032795648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906262626262626,
      "loss": 2.6357,
      "theoretical_loss": 3.4257398898194067,
      "tokens_seen": 2032861184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039060606060606064,
      "loss": 2.6523,
      "theoretical_loss": 3.425730711232626,
      "tokens_seen": 2032926720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003905858585858586,
      "loss": 2.5825,
      "theoretical_loss": 3.4257215330245794,
      "tokens_seen": 2032992256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039056565656565655,
      "loss": 2.5612,
      "theoretical_loss": 3.42571235519524,
      "tokens_seen": 2033057792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039054545454545456,
      "loss": 2.4326,
      "theoretical_loss": 3.4257031777445803,
      "tokens_seen": 2033123328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003905252525252525,
      "loss": 2.8784,
      "theoretical_loss": 3.425694000672571,
      "tokens_seen": 2033188864
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039050505050505053,
      "loss": 2.4251,
      "theoretical_loss": 3.425684823979185,
      "tokens_seen": 2033254400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904848484848485,
      "loss": 2.722,
      "theoretical_loss": 3.4256756476643946,
      "tokens_seen": 2033319936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039046464646464644,
      "loss": 2.7486,
      "theoretical_loss": 3.4256664717281717,
      "tokens_seen": 2033385472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039044444444444445,
      "loss": 2.4272,
      "theoretical_loss": 3.425657296170489,
      "tokens_seen": 2033451008
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1120297,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3779661655426025,
      "objective/train/theoretical_loss": 3.425648120991318,
      "objective/train/tokens_used": 392375776,
      "theoretical_loss": 3.425648120991318,
      "tokens_seen": 2033516544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039042424242424247,
      "loss": 2.5723,
      "theoretical_loss": 3.425648120991318,
      "tokens_seen": 2033516544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904040404040404,
      "loss": 2.9231,
      "theoretical_loss": 3.425638946190632,
      "tokens_seen": 2033582080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903838383838384,
      "loss": 2.6368,
      "theoretical_loss": 3.4256297717684014,
      "tokens_seen": 2033647616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903636363636364,
      "loss": 2.6167,
      "theoretical_loss": 3.4256205977246,
      "tokens_seen": 2033713152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039034343434343435,
      "loss": 2.5206,
      "theoretical_loss": 3.4256114240591993,
      "tokens_seen": 2033778688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039032323232323236,
      "loss": 2.6265,
      "theoretical_loss": 3.4256022507721715,
      "tokens_seen": 2033844224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903030303030303,
      "loss": 2.5845,
      "theoretical_loss": 3.425593077863489,
      "tokens_seen": 2033909760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039028282828282827,
      "loss": 2.7537,
      "theoretical_loss": 3.4255839053331236,
      "tokens_seen": 2033975296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902626262626263,
      "loss": 2.8376,
      "theoretical_loss": 3.4255747331810484,
      "tokens_seen": 2034040832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039024242424242424,
      "loss": 2.4043,
      "theoretical_loss": 3.4255655614072342,
      "tokens_seen": 2034106368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039022222222222225,
      "loss": 2.5765,
      "theoretical_loss": 3.4255563900116552,
      "tokens_seen": 2034171904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902020202020202,
      "loss": 2.7122,
      "theoretical_loss": 3.4255472189942813,
      "tokens_seen": 2034237440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901818181818182,
      "loss": 2.5745,
      "theoretical_loss": 3.4255380483550866,
      "tokens_seen": 2034302976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901616161616162,
      "loss": 2.4317,
      "theoretical_loss": 3.4255288780940427,
      "tokens_seen": 2034368512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039014141414141413,
      "loss": 2.5789,
      "theoretical_loss": 3.4255197082111213,
      "tokens_seen": 2034434048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039012121212121214,
      "loss": 2.5604,
      "theoretical_loss": 3.425510538706295,
      "tokens_seen": 2034499584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901010101010101,
      "loss": 2.6917,
      "theoretical_loss": 3.4255013695795364,
      "tokens_seen": 2034565120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003900808080808081,
      "loss": 2.6604,
      "theoretical_loss": 3.4254922008308175,
      "tokens_seen": 2034630656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039006060606060607,
      "loss": 2.5923,
      "theoretical_loss": 3.4254830324601104,
      "tokens_seen": 2034696192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390040404040404,
      "loss": 2.6612,
      "theoretical_loss": 3.425473864467387,
      "tokens_seen": 2034761728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039002020202020204,
      "loss": 2.7573,
      "theoretical_loss": 3.425464696852621,
      "tokens_seen": 2034827264
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039000000000000005,
      "loss": 2.567,
      "theoretical_loss": 3.4254555296157827,
      "tokens_seen": 2034892800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389979797979798,
      "loss": 2.8421,
      "theoretical_loss": 3.425446362756845,
      "tokens_seen": 2034958336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038995959595959596,
      "loss": 2.605,
      "theoretical_loss": 3.425437196275781,
      "tokens_seen": 2035023872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899393939393939,
      "loss": 2.7614,
      "theoretical_loss": 3.4254280301725624,
      "tokens_seen": 2035089408
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1121395,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6363847255706787,
      "objective/train/theoretical_loss": 3.425418864447161,
      "objective/train/tokens_used": 394014176,
      "theoretical_loss": 3.425418864447161,
      "tokens_seen": 2035154944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038991919191919193,
      "loss": 2.5577,
      "theoretical_loss": 3.425418864447161,
      "tokens_seen": 2035154944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038989898989898994,
      "loss": 2.5876,
      "theoretical_loss": 3.42540969909955,
      "tokens_seen": 2035220480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898787878787879,
      "loss": 2.4677,
      "theoretical_loss": 3.4254005341297007,
      "tokens_seen": 2035286016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038985858585858585,
      "loss": 2.5843,
      "theoretical_loss": 3.4253913695375857,
      "tokens_seen": 2035351552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898383838383838,
      "loss": 2.536,
      "theoretical_loss": 3.425382205323178,
      "tokens_seen": 2035417088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898181818181819,
      "loss": 2.59,
      "theoretical_loss": 3.4253730414864485,
      "tokens_seen": 2035482624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038979797979797983,
      "loss": 2.6873,
      "theoretical_loss": 3.4253638780273707,
      "tokens_seen": 2035548160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897777777777778,
      "loss": 2.8808,
      "theoretical_loss": 3.4253547149459163,
      "tokens_seen": 2035613696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038975757575757575,
      "loss": 2.7569,
      "theoretical_loss": 3.4253455522420575,
      "tokens_seen": 2035679232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897373737373737,
      "loss": 2.608,
      "theoretical_loss": 3.425336389915767,
      "tokens_seen": 2035744768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038971717171717177,
      "loss": 2.6626,
      "theoretical_loss": 3.4253272279670166,
      "tokens_seen": 2035810304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896969696969697,
      "loss": 2.7393,
      "theoretical_loss": 3.4253180663957794,
      "tokens_seen": 2035875840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896767676767677,
      "loss": 2.4374,
      "theoretical_loss": 3.4253089052020265,
      "tokens_seen": 2035941376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038965656565656564,
      "loss": 2.5749,
      "theoretical_loss": 3.425299744385731,
      "tokens_seen": 2036006912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896363636363636,
      "loss": 2.7259,
      "theoretical_loss": 3.4252905839468655,
      "tokens_seen": 2036072448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038961616161616166,
      "loss": 2.6453,
      "theoretical_loss": 3.425281423885401,
      "tokens_seen": 2036137984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003895959595959596,
      "loss": 2.7484,
      "theoretical_loss": 3.425272264201311,
      "tokens_seen": 2036203520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003895757575757576,
      "loss": 2.701,
      "theoretical_loss": 3.425263104894568,
      "tokens_seen": 2036269056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038955555555555553,
      "loss": 2.6245,
      "theoretical_loss": 3.425253945965143,
      "tokens_seen": 2036334592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038953535353535354,
      "loss": 2.8589,
      "theoretical_loss": 3.4252447874130096,
      "tokens_seen": 2036400128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038951515151515155,
      "loss": 2.7356,
      "theoretical_loss": 3.425235629238139,
      "tokens_seen": 2036465664
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894949494949495,
      "loss": 2.572,
      "theoretical_loss": 3.4252264714405047,
      "tokens_seen": 2036531200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038947474747474747,
      "loss": 2.6346,
      "theoretical_loss": 3.4252173140200783,
      "tokens_seen": 2036596736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894545454545454,
      "loss": 2.6631,
      "theoretical_loss": 3.425208156976832,
      "tokens_seen": 2036662272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038943434343434344,
      "loss": 2.7026,
      "theoretical_loss": 3.4251990003107387,
      "tokens_seen": 2036727808
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1121841,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7600228786468506,
      "objective/train/theoretical_loss": 3.4251898440217703,
      "objective/train/tokens_used": 395652576,
      "theoretical_loss": 3.4251898440217703,
      "tokens_seen": 2036793344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038941414141414145,
      "loss": 2.7747,
      "theoretical_loss": 3.4251898440217703,
      "tokens_seen": 2036793344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893939393939394,
      "loss": 2.5716,
      "theoretical_loss": 3.425180688109899,
      "tokens_seen": 2036858880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038937373737373736,
      "loss": 2.3788,
      "theoretical_loss": 3.4251715325750975,
      "tokens_seen": 2036924416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038935353535353537,
      "loss": 2.5345,
      "theoretical_loss": 3.4251623774173385,
      "tokens_seen": 2036989952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038933333333333333,
      "loss": 2.5381,
      "theoretical_loss": 3.4251532226365935,
      "tokens_seen": 2037055488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038931313131313134,
      "loss": 2.7236,
      "theoretical_loss": 3.425144068232835,
      "tokens_seen": 2037121024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892929292929293,
      "loss": 2.685,
      "theoretical_loss": 3.425134914206036,
      "tokens_seen": 2037186560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038927272727272725,
      "loss": 2.528,
      "theoretical_loss": 3.425125760556168,
      "tokens_seen": 2037252096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038925252525252526,
      "loss": 2.6716,
      "theoretical_loss": 3.425116607283204,
      "tokens_seen": 2037317632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892323232323232,
      "loss": 2.7653,
      "theoretical_loss": 3.425107454387116,
      "tokens_seen": 2037383168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038921212121212123,
      "loss": 2.6616,
      "theoretical_loss": 3.4250983018678767,
      "tokens_seen": 2037448704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891919191919192,
      "loss": 2.7058,
      "theoretical_loss": 3.425089149725458,
      "tokens_seen": 2037514240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891717171717172,
      "loss": 2.8981,
      "theoretical_loss": 3.4250799979598328,
      "tokens_seen": 2037579776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038915151515151516,
      "loss": 2.6551,
      "theoretical_loss": 3.4250708465709727,
      "tokens_seen": 2037645312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891313131313131,
      "loss": 2.4252,
      "theoretical_loss": 3.425061695558851,
      "tokens_seen": 2037710848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891111111111111,
      "loss": 2.7006,
      "theoretical_loss": 3.4250525449234397,
      "tokens_seen": 2037776384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890909090909091,
      "loss": 2.4131,
      "theoretical_loss": 3.4250433946647107,
      "tokens_seen": 2037841920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890707070707071,
      "loss": 2.5942,
      "theoretical_loss": 3.425034244782637,
      "tokens_seen": 2037907456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038905050505050505,
      "loss": 2.6065,
      "theoretical_loss": 3.4250250952771912,
      "tokens_seen": 2037972992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389030303030303,
      "loss": 2.6553,
      "theoretical_loss": 3.4250159461483447,
      "tokens_seen": 2038038528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389010101010101,
      "loss": 2.7689,
      "theoretical_loss": 3.4250067973960707,
      "tokens_seen": 2038104064
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038898989898989903,
      "loss": 2.6705,
      "theoretical_loss": 3.424997649020341,
      "tokens_seen": 2038169600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000388969696969697,
      "loss": 2.5966,
      "theoretical_loss": 3.4249885010211285,
      "tokens_seen": 2038235136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038894949494949494,
      "loss": 2.5179,
      "theoretical_loss": 3.4249793533984056,
      "tokens_seen": 2038300672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889292929292929,
      "loss": 2.5465,
      "theoretical_loss": 3.4249702061521443,
      "tokens_seen": 2038366208
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1123215,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.411255121231079,
      "objective/train/theoretical_loss": 3.424961059282318,
      "objective/train/tokens_used": 397290976,
      "theoretical_loss": 3.424961059282318,
      "tokens_seen": 2038431744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889090909090909,
      "loss": 2.5545,
      "theoretical_loss": 3.424961059282318,
      "tokens_seen": 2038431744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888888888888889,
      "loss": 2.5991,
      "theoretical_loss": 3.4249519127888974,
      "tokens_seen": 2038497280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888686868686869,
      "loss": 2.6284,
      "theoretical_loss": 3.4249427666718564,
      "tokens_seen": 2038562816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038884848484848483,
      "loss": 2.661,
      "theoretical_loss": 3.4249336209311667,
      "tokens_seen": 2038628352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888282828282828,
      "loss": 2.5987,
      "theoretical_loss": 3.424924475566801,
      "tokens_seen": 2038693888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038880808080808086,
      "loss": 2.6224,
      "theoretical_loss": 3.4249153305787314,
      "tokens_seen": 2038759424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887878787878788,
      "loss": 2.4543,
      "theoretical_loss": 3.4249061859669307,
      "tokens_seen": 2038824960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038876767676767677,
      "loss": 2.6108,
      "theoretical_loss": 3.424897041731371,
      "tokens_seen": 2038890496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038874747474747473,
      "loss": 2.4736,
      "theoretical_loss": 3.424887897872025,
      "tokens_seen": 2038956032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038872727272727274,
      "loss": 2.8445,
      "theoretical_loss": 3.4248787543888652,
      "tokens_seen": 2039021568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038870707070707075,
      "loss": 2.6671,
      "theoretical_loss": 3.4248696112818635,
      "tokens_seen": 2039087104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886868686868687,
      "loss": 2.6505,
      "theoretical_loss": 3.424860468550993,
      "tokens_seen": 2039152640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038866666666666666,
      "loss": 2.3294,
      "theoretical_loss": 3.4248513261962255,
      "tokens_seen": 2039218176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886464646464646,
      "loss": 2.5899,
      "theoretical_loss": 3.4248421842175336,
      "tokens_seen": 2039283712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886262626262627,
      "loss": 2.9221,
      "theoretical_loss": 3.4248330426148907,
      "tokens_seen": 2039349248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038860606060606064,
      "loss": 2.5795,
      "theoretical_loss": 3.424823901388268,
      "tokens_seen": 2039414784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885858585858586,
      "loss": 2.593,
      "theoretical_loss": 3.424814760537638,
      "tokens_seen": 2039480320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038856565656565656,
      "loss": 2.592,
      "theoretical_loss": 3.4248056200629744,
      "tokens_seen": 2039545856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885454545454545,
      "loss": 2.6182,
      "theoretical_loss": 3.424796479964248,
      "tokens_seen": 2039611392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885252525252526,
      "loss": 2.8057,
      "theoretical_loss": 3.4247873402414326,
      "tokens_seen": 2039676928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038850505050505054,
      "loss": 2.5074,
      "theoretical_loss": 3.4247782008945,
      "tokens_seen": 2039742464
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884848484848485,
      "loss": 2.8014,
      "theoretical_loss": 3.4247690619234223,
      "tokens_seen": 2039808000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038846464646464645,
      "loss": 2.5553,
      "theoretical_loss": 3.424759923328173,
      "tokens_seen": 2039873536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038844444444444446,
      "loss": 2.6588,
      "theoretical_loss": 3.424750785108724,
      "tokens_seen": 2039939072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038842424242424247,
      "loss": 2.6811,
      "theoretical_loss": 3.424741647265048,
      "tokens_seen": 2040004608
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1124379,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8337912559509277,
      "objective/train/theoretical_loss": 3.424732509797117,
      "objective/train/tokens_used": 398929376,
      "theoretical_loss": 3.424732509797117,
      "tokens_seen": 2040070144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038840404040404043,
      "loss": 2.8508,
      "theoretical_loss": 3.424732509797117,
      "tokens_seen": 2040070144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883838383838384,
      "loss": 2.7414,
      "theoretical_loss": 3.4247233727049036,
      "tokens_seen": 2040135680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038836363636363634,
      "loss": 2.5021,
      "theoretical_loss": 3.4247142359883807,
      "tokens_seen": 2040201216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038834343434343435,
      "loss": 2.6071,
      "theoretical_loss": 3.424705099647521,
      "tokens_seen": 2040266752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038832323232323236,
      "loss": 2.665,
      "theoretical_loss": 3.4246959636822956,
      "tokens_seen": 2040332288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883030303030303,
      "loss": 2.6956,
      "theoretical_loss": 3.4246868280926783,
      "tokens_seen": 2040397824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882828282828283,
      "loss": 2.6893,
      "theoretical_loss": 3.4246776928786415,
      "tokens_seen": 2040463360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882626262626263,
      "loss": 2.4344,
      "theoretical_loss": 3.4246685580401577,
      "tokens_seen": 2040528896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038824242424242425,
      "loss": 2.6557,
      "theoretical_loss": 3.4246594235771983,
      "tokens_seen": 2040594432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038822222222222226,
      "loss": 2.6863,
      "theoretical_loss": 3.424650289489737,
      "tokens_seen": 2040659968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882020202020202,
      "loss": 2.6083,
      "theoretical_loss": 3.4246411557777456,
      "tokens_seen": 2040725504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038818181818181817,
      "loss": 2.397,
      "theoretical_loss": 3.4246320224411972,
      "tokens_seen": 2040791040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881616161616162,
      "loss": 2.54,
      "theoretical_loss": 3.424622889480064,
      "tokens_seen": 2040856576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038814141414141414,
      "loss": 2.4706,
      "theoretical_loss": 3.4246137568943187,
      "tokens_seen": 2040922112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038812121212121215,
      "loss": 2.5738,
      "theoretical_loss": 3.4246046246839335,
      "tokens_seen": 2040987648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881010101010101,
      "loss": 2.5798,
      "theoretical_loss": 3.424595492848881,
      "tokens_seen": 2041053184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003880808080808081,
      "loss": 2.5736,
      "theoretical_loss": 3.4245863613891343,
      "tokens_seen": 2041118720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003880606060606061,
      "loss": 2.5231,
      "theoretical_loss": 3.424577230304665,
      "tokens_seen": 2041184256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038804040404040403,
      "loss": 2.4575,
      "theoretical_loss": 3.424568099595446,
      "tokens_seen": 2041249792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038802020202020204,
      "loss": 2.5567,
      "theoretical_loss": 3.42455896926145,
      "tokens_seen": 2041315328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000388,
      "loss": 2.5012,
      "theoretical_loss": 3.4245498393026494,
      "tokens_seen": 2041380864
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387979797979798,
      "loss": 2.6823,
      "theoretical_loss": 3.4245407097190173,
      "tokens_seen": 2041446400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038795959595959597,
      "loss": 2.5196,
      "theoretical_loss": 3.424531580510525,
      "tokens_seen": 2041511936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879393939393939,
      "loss": 2.5629,
      "theoretical_loss": 3.424522451677146,
      "tokens_seen": 2041577472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038791919191919193,
      "loss": 2.5199,
      "theoretical_loss": 3.4245133232188527,
      "tokens_seen": 2041643008
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1125100,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.047553062438965,
      "objective/train/theoretical_loss": 3.4245041951356177,
      "objective/train/tokens_used": 400567776,
      "theoretical_loss": 3.4245041951356177,
      "tokens_seen": 2041708544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878989898989899,
      "loss": 2.7422,
      "theoretical_loss": 3.4245041951356177,
      "tokens_seen": 2041708544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878787878787879,
      "loss": 2.6362,
      "theoretical_loss": 3.424495067427413,
      "tokens_seen": 2041774080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038785858585858586,
      "loss": 2.6407,
      "theoretical_loss": 3.4244859400942116,
      "tokens_seen": 2041839616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878383838383838,
      "loss": 2.5186,
      "theoretical_loss": 3.424476813135986,
      "tokens_seen": 2041905152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878181818181818,
      "loss": 2.421,
      "theoretical_loss": 3.424467686552709,
      "tokens_seen": 2041970688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038779797979797984,
      "loss": 2.5184,
      "theoretical_loss": 3.4244585603443527,
      "tokens_seen": 2042036224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877777777777778,
      "loss": 2.6667,
      "theoretical_loss": 3.4244494345108905,
      "tokens_seen": 2042101760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038775757575757575,
      "loss": 2.564,
      "theoretical_loss": 3.4244403090522937,
      "tokens_seen": 2042167296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877373737373737,
      "loss": 2.5341,
      "theoretical_loss": 3.4244311839685357,
      "tokens_seen": 2042232832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877171717171717,
      "loss": 2.4584,
      "theoretical_loss": 3.424422059259589,
      "tokens_seen": 2042298368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038769696969696973,
      "loss": 2.5476,
      "theoretical_loss": 3.424412934925426,
      "tokens_seen": 2042363904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876767676767677,
      "loss": 2.4415,
      "theoretical_loss": 3.4244038109660195,
      "tokens_seen": 2042429440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038765656565656564,
      "loss": 2.5881,
      "theoretical_loss": 3.4243946873813425,
      "tokens_seen": 2042494976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876363636363636,
      "loss": 2.6939,
      "theoretical_loss": 3.4243855641713665,
      "tokens_seen": 2042560512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038761616161616167,
      "loss": 2.2827,
      "theoretical_loss": 3.4243764413360642,
      "tokens_seen": 2042626048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875959595959596,
      "loss": 2.7188,
      "theoretical_loss": 3.4243673188754093,
      "tokens_seen": 2042691584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875757575757576,
      "loss": 2.4194,
      "theoretical_loss": 3.4243581967893735,
      "tokens_seen": 2042757120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038755555555555554,
      "loss": 2.6265,
      "theoretical_loss": 3.42434907507793,
      "tokens_seen": 2042822656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875353535353535,
      "loss": 2.5645,
      "theoretical_loss": 3.4243399537410504,
      "tokens_seen": 2042888192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038751515151515156,
      "loss": 2.5277,
      "theoretical_loss": 3.424330832778708,
      "tokens_seen": 2042953728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874949494949495,
      "loss": 2.7011,
      "theoretical_loss": 3.424321712190876,
      "tokens_seen": 2043019264
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874747474747475,
      "loss": 2.5774,
      "theoretical_loss": 3.4243125919775257,
      "tokens_seen": 2043084800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038745454545454543,
      "loss": 2.6009,
      "theoretical_loss": 3.4243034721386305,
      "tokens_seen": 2043150336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038743434343434344,
      "loss": 2.6962,
      "theoretical_loss": 3.424294352674163,
      "tokens_seen": 2043215872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038741414141414145,
      "loss": 2.8634,
      "theoretical_loss": 3.4242852335840954,
      "tokens_seen": 2043281408
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1126113,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.695091962814331,
      "objective/train/theoretical_loss": 3.424276114868401,
      "objective/train/tokens_used": 402206176,
      "theoretical_loss": 3.424276114868401,
      "tokens_seen": 2043346944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873939393939394,
      "loss": 2.7604,
      "theoretical_loss": 3.424276114868401,
      "tokens_seen": 2043346944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038737373737373737,
      "loss": 2.4762,
      "theoretical_loss": 3.424266996527052,
      "tokens_seen": 2043412480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873535353535353,
      "loss": 2.6634,
      "theoretical_loss": 3.424257878560021,
      "tokens_seen": 2043478016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038733333333333333,
      "loss": 2.4992,
      "theoretical_loss": 3.4242487609672807,
      "tokens_seen": 2043543552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038731313131313134,
      "loss": 2.8986,
      "theoretical_loss": 3.4242396437488036,
      "tokens_seen": 2043609088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872929292929293,
      "loss": 2.6341,
      "theoretical_loss": 3.4242305269045623,
      "tokens_seen": 2043674624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038727272727272726,
      "loss": 2.5154,
      "theoretical_loss": 3.42422141043453,
      "tokens_seen": 2043740160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038725252525252527,
      "loss": 2.5922,
      "theoretical_loss": 3.4242122943386786,
      "tokens_seen": 2043805696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872323232323232,
      "loss": 2.5824,
      "theoretical_loss": 3.4242031786169815,
      "tokens_seen": 2043871232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038721212121212124,
      "loss": 2.6942,
      "theoretical_loss": 3.424194063269411,
      "tokens_seen": 2043936768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871919191919192,
      "loss": 2.6585,
      "theoretical_loss": 3.424184948295939,
      "tokens_seen": 2044002304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038717171717171715,
      "loss": 2.4877,
      "theoretical_loss": 3.4241758336965393,
      "tokens_seen": 2044067840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038715151515151516,
      "loss": 2.5554,
      "theoretical_loss": 3.4241667194711845,
      "tokens_seen": 2044133376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871313131313131,
      "loss": 2.6856,
      "theoretical_loss": 3.424157605619846,
      "tokens_seen": 2044198912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038711111111111113,
      "loss": 2.4869,
      "theoretical_loss": 3.424148492142498,
      "tokens_seen": 2044264448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870909090909091,
      "loss": 2.5563,
      "theoretical_loss": 3.4241393790391124,
      "tokens_seen": 2044329984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870707070707071,
      "loss": 2.5721,
      "theoretical_loss": 3.4241302663096613,
      "tokens_seen": 2044395520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038705050505050505,
      "loss": 2.4328,
      "theoretical_loss": 3.4241211539541188,
      "tokens_seen": 2044461056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038703030303030307,
      "loss": 2.4588,
      "theoretical_loss": 3.4241120419724567,
      "tokens_seen": 2044526592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387010101010101,
      "loss": 2.4166,
      "theoretical_loss": 3.424102930364647,
      "tokens_seen": 2044592128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000386989898989899,
      "loss": 2.4133,
      "theoretical_loss": 3.424093819130664,
      "tokens_seen": 2044657664
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000386969696969697,
      "loss": 2.6739,
      "theoretical_loss": 3.4240847082704793,
      "tokens_seen": 2044723200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038694949494949495,
      "loss": 2.4544,
      "theoretical_loss": 3.4240755977840656,
      "tokens_seen": 2044788736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038692929292929296,
      "loss": 2.4786,
      "theoretical_loss": 3.4240664876713955,
      "tokens_seen": 2044854272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869090909090909,
      "loss": 2.6681,
      "theoretical_loss": 3.424057377932443,
      "tokens_seen": 2044919808
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1126722,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8349125385284424,
      "objective/train/theoretical_loss": 3.424048268567179,
      "objective/train/tokens_used": 403844576,
      "theoretical_loss": 3.424048268567179,
      "tokens_seen": 2044985344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868888888888889,
      "loss": 2.7018,
      "theoretical_loss": 3.424048268567179,
      "tokens_seen": 2044985344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868686868686869,
      "loss": 2.6153,
      "theoretical_loss": 3.4240391595755773,
      "tokens_seen": 2045050880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038684848484848484,
      "loss": 2.6144,
      "theoretical_loss": 3.42403005095761,
      "tokens_seen": 2045116416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038682828282828285,
      "loss": 2.3565,
      "theoretical_loss": 3.42402094271325,
      "tokens_seen": 2045181952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868080808080808,
      "loss": 2.513,
      "theoretical_loss": 3.424011834842471,
      "tokens_seen": 2045247488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867878787878788,
      "loss": 2.8146,
      "theoretical_loss": 3.424002727345244,
      "tokens_seen": 2045313024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867676767676768,
      "loss": 2.6684,
      "theoretical_loss": 3.4239936202215424,
      "tokens_seen": 2045378560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038674747474747473,
      "loss": 2.3647,
      "theoretical_loss": 3.423984513471339,
      "tokens_seen": 2045444096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038672727272727274,
      "loss": 2.5581,
      "theoretical_loss": 3.423975407094607,
      "tokens_seen": 2045509632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038670707070707076,
      "loss": 2.6833,
      "theoretical_loss": 3.423966301091318,
      "tokens_seen": 2045575168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003866868686868687,
      "loss": 2.6042,
      "theoretical_loss": 3.423957195461446,
      "tokens_seen": 2045640704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038666666666666667,
      "loss": 2.7104,
      "theoretical_loss": 3.423948090204963,
      "tokens_seen": 2045706240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003866464646464646,
      "loss": 2.6974,
      "theoretical_loss": 3.4239389853218416,
      "tokens_seen": 2045771776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038662626262626264,
      "loss": 2.5431,
      "theoretical_loss": 3.4239298808120546,
      "tokens_seen": 2045837312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038660606060606065,
      "loss": 2.4468,
      "theoretical_loss": 3.4239207766755753,
      "tokens_seen": 2045902848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865858585858586,
      "loss": 2.3854,
      "theoretical_loss": 3.423911672912376,
      "tokens_seen": 2045968384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038656565656565656,
      "loss": 2.5403,
      "theoretical_loss": 3.423902569522429,
      "tokens_seen": 2046033920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865454545454545,
      "loss": 2.4529,
      "theoretical_loss": 3.4238934665057084,
      "tokens_seen": 2046099456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865252525252526,
      "loss": 2.5285,
      "theoretical_loss": 3.4238843638621854,
      "tokens_seen": 2046164992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038650505050505054,
      "loss": 2.7578,
      "theoretical_loss": 3.4238752615918333,
      "tokens_seen": 2046230528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864848484848485,
      "loss": 2.6517,
      "theoretical_loss": 3.423866159694625,
      "tokens_seen": 2046296064
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038646464646464645,
      "loss": 2.595,
      "theoretical_loss": 3.4238570581705337,
      "tokens_seen": 2046361600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864444444444444,
      "loss": 2.5574,
      "theoretical_loss": 3.4238479570195315,
      "tokens_seen": 2046427136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864242424242425,
      "loss": 2.3707,
      "theoretical_loss": 3.423838856241591,
      "tokens_seen": 2046492672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038640404040404043,
      "loss": 2.7492,
      "theoretical_loss": 3.4238297558366853,
      "tokens_seen": 2046558208
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 1127892,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5316739082336426,
      "objective/train/theoretical_loss": 3.4238206558047874,
      "objective/train/tokens_used": 405482976,
      "theoretical_loss": 3.4238206558047874,
      "tokens_seen": 2046623744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003863838383838384,
      "loss": 2.4914,
      "theoretical_loss": 3.4238206558047874,
      "tokens_seen": 2046623744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038636363636363635,
      "loss": 2.6275,
      "theoretical_loss": 3.4238115561458695,
      "tokens_seen": 2046689280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863434343434343,
      "loss": 2.3453,
      "theoretical_loss": 3.423802456859905,
      "tokens_seen": 2046754816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038632323232323237,
      "loss": 2.355,
      "theoretical_loss": 3.423793357946866,
      "tokens_seen": 2046820352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863030303030303,
      "loss": 2.4346,
      "theoretical_loss": 3.423784259406726,
      "tokens_seen": 2046885888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862828282828283,
      "loss": 2.4405,
      "theoretical_loss": 3.423775161239457,
      "tokens_seen": 2046951424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038626262626262624,
      "loss": 2.6577,
      "theoretical_loss": 3.4237660634450324,
      "tokens_seen": 2047016960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038624242424242425,
      "loss": 2.4727,
      "theoretical_loss": 3.423756966023425,
      "tokens_seen": 2047082496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038622222222222226,
      "loss": 2.6148,
      "theoretical_loss": 3.4237478689746066,
      "tokens_seen": 2047148032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862020202020202,
      "loss": 2.5595,
      "theoretical_loss": 3.4237387722985515,
      "tokens_seen": 2047213568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861818181818182,
      "loss": 2.6433,
      "theoretical_loss": 3.4237296759952316,
      "tokens_seen": 2047279104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038616161616161613,
      "loss": 2.4541,
      "theoretical_loss": 3.4237205800646198,
      "tokens_seen": 2047344640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038614141414141414,
      "loss": 2.6971,
      "theoretical_loss": 3.423711484506689,
      "tokens_seen": 2047410176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038612121212121215,
      "loss": 2.3779,
      "theoretical_loss": 3.423702389321412,
      "tokens_seen": 2047475712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861010101010101,
      "loss": 2.7949,
      "theoretical_loss": 3.4236932945087615,
      "tokens_seen": 2047541248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038608080808080807,
      "loss": 2.5279,
      "theoretical_loss": 3.42368420006871,
      "tokens_seen": 2047606784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860606060606061,
      "loss": 2.692,
      "theoretical_loss": 3.423675106001231,
      "tokens_seen": 2047672320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038604040404040404,
      "loss": 2.4909,
      "theoretical_loss": 3.423666012306297,
      "tokens_seen": 2047737856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038602020202020205,
      "loss": 2.4489,
      "theoretical_loss": 3.423656918983881,
      "tokens_seen": 2047803392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000386,
      "loss": 2.6628,
      "theoretical_loss": 3.423647826033955,
      "tokens_seen": 2047868928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038597979797979796,
      "loss": 2.4424,
      "theoretical_loss": 3.4236387334564933,
      "tokens_seen": 2047934464
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038595959595959597,
      "loss": 2.7376,
      "theoretical_loss": 3.423629641251467,
      "tokens_seen": 2048000000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038593939393939393,
      "loss": 2.4449,
      "theoretical_loss": 3.42362054941885,
      "tokens_seen": 2048065536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038591919191919194,
      "loss": 2.4178,
      "theoretical_loss": 3.423611457958615,
      "tokens_seen": 2048131072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858989898989899,
      "loss": 2.3096,
      "theoretical_loss": 3.4236023668707354,
      "tokens_seen": 2048196608
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1128475,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.619576930999756,
      "objective/train/theoretical_loss": 3.4235932761551826,
      "objective/train/tokens_used": 407121376,
      "theoretical_loss": 3.4235932761551826,
      "tokens_seen": 2048262144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858787878787879,
      "loss": 2.5873,
      "theoretical_loss": 3.4235932761551826,
      "tokens_seen": 2048262144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038585858585858586,
      "loss": 2.8021,
      "theoretical_loss": 3.4235841858119302,
      "tokens_seen": 2048327680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858383838383838,
      "loss": 2.5011,
      "theoretical_loss": 3.4235750958409517,
      "tokens_seen": 2048393216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038581818181818183,
      "loss": 2.7487,
      "theoretical_loss": 3.4235660062422184,
      "tokens_seen": 2048458752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857979797979798,
      "loss": 2.5642,
      "theoretical_loss": 3.4235569170157047,
      "tokens_seen": 2048524288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857777777777778,
      "loss": 2.6144,
      "theoretical_loss": 3.423547828161383,
      "tokens_seen": 2048589824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038575757575757576,
      "loss": 2.6756,
      "theoretical_loss": 3.4235387396792256,
      "tokens_seen": 2048655360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857373737373737,
      "loss": 2.501,
      "theoretical_loss": 3.423529651569206,
      "tokens_seen": 2048720896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857171717171717,
      "loss": 2.4741,
      "theoretical_loss": 3.423520563831296,
      "tokens_seen": 2048786432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038569696969696974,
      "loss": 2.6386,
      "theoretical_loss": 3.42351147646547,
      "tokens_seen": 2048851968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856767676767677,
      "loss": 2.8198,
      "theoretical_loss": 3.4235023894717003,
      "tokens_seen": 2048917504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038565656565656565,
      "loss": 2.8276,
      "theoretical_loss": 3.423493302849959,
      "tokens_seen": 2048983040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856363636363636,
      "loss": 2.6402,
      "theoretical_loss": 3.4234842166002197,
      "tokens_seen": 2049048576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856161616161616,
      "loss": 2.4989,
      "theoretical_loss": 3.4234751307224554,
      "tokens_seen": 2049114112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038559595959595963,
      "loss": 2.5747,
      "theoretical_loss": 3.4234660452166383,
      "tokens_seen": 2049179648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855757575757576,
      "loss": 2.3174,
      "theoretical_loss": 3.423456960082742,
      "tokens_seen": 2049245184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038555555555555554,
      "loss": 2.5597,
      "theoretical_loss": 3.423447875320739,
      "tokens_seen": 2049310720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855353535353535,
      "loss": 2.396,
      "theoretical_loss": 3.423438790930602,
      "tokens_seen": 2049376256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038551515151515156,
      "loss": 2.3726,
      "theoretical_loss": 3.4234297069123043,
      "tokens_seen": 2049441792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854949494949495,
      "loss": 2.4606,
      "theoretical_loss": 3.4234206232658186,
      "tokens_seen": 2049507328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854747474747475,
      "loss": 2.6207,
      "theoretical_loss": 3.423411539991118,
      "tokens_seen": 2049572864
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038545454545454544,
      "loss": 2.3535,
      "theoretical_loss": 3.4234024570881747,
      "tokens_seen": 2049638400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854343434343434,
      "loss": 2.5528,
      "theoretical_loss": 3.4233933745569627,
      "tokens_seen": 2049703936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038541414141414146,
      "loss": 2.4746,
      "theoretical_loss": 3.4233842923974542,
      "tokens_seen": 2049769472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003853939393939394,
      "loss": 2.4524,
      "theoretical_loss": 3.423375210609622,
      "tokens_seen": 2049835008
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1129543,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6075730323791504,
      "objective/train/theoretical_loss": 3.423366129193439,
      "objective/train/tokens_used": 408759776,
      "theoretical_loss": 3.423366129193439,
      "tokens_seen": 2049900544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038537373737373737,
      "loss": 2.5068,
      "theoretical_loss": 3.423366129193439,
      "tokens_seen": 2049900544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038535353535353533,
      "loss": 2.8476,
      "theoretical_loss": 3.423357048148879,
      "tokens_seen": 2049966080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003853333333333334,
      "loss": 2.5966,
      "theoretical_loss": 3.4233479674759137,
      "tokens_seen": 2050031616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038531313131313135,
      "loss": 2.5537,
      "theoretical_loss": 3.4233388871745167,
      "tokens_seen": 2050097152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852929292929293,
      "loss": 2.5581,
      "theoretical_loss": 3.4233298072446607,
      "tokens_seen": 2050162688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038527272727272726,
      "loss": 2.5214,
      "theoretical_loss": 3.4233207276863187,
      "tokens_seen": 2050228224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852525252525252,
      "loss": 2.3995,
      "theoretical_loss": 3.423311648499464,
      "tokens_seen": 2050293760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852323232323233,
      "loss": 2.682,
      "theoretical_loss": 3.4233025696840684,
      "tokens_seen": 2050359296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038521212121212124,
      "loss": 2.5779,
      "theoretical_loss": 3.4232934912401065,
      "tokens_seen": 2050424832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851919191919192,
      "loss": 2.7086,
      "theoretical_loss": 3.4232844131675497,
      "tokens_seen": 2050490368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038517171717171716,
      "loss": 2.5114,
      "theoretical_loss": 3.4232753354663714,
      "tokens_seen": 2050555904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038515151515151517,
      "loss": 2.6217,
      "theoretical_loss": 3.4232662581365454,
      "tokens_seen": 2050621440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851313131313132,
      "loss": 2.7702,
      "theoretical_loss": 3.4232571811780437,
      "tokens_seen": 2050686976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038511111111111114,
      "loss": 2.4582,
      "theoretical_loss": 3.423248104590839,
      "tokens_seen": 2050752512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850909090909091,
      "loss": 2.4804,
      "theoretical_loss": 3.423239028374905,
      "tokens_seen": 2050818048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038507070707070705,
      "loss": 2.4453,
      "theoretical_loss": 3.423229952530215,
      "tokens_seen": 2050883584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038505050505050506,
      "loss": 2.5674,
      "theoretical_loss": 3.4232208770567403,
      "tokens_seen": 2050949120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038503030303030307,
      "loss": 2.5093,
      "theoretical_loss": 3.423211801954455,
      "tokens_seen": 2051014656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038501010101010103,
      "loss": 2.4003,
      "theoretical_loss": 3.4232027272233325,
      "tokens_seen": 2051080192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384989898989899,
      "loss": 2.4936,
      "theoretical_loss": 3.423193652863345,
      "tokens_seen": 2051145728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384969696969697,
      "loss": 2.6507,
      "theoretical_loss": 3.4231845788744657,
      "tokens_seen": 2051211264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038494949494949495,
      "loss": 2.6008,
      "theoretical_loss": 3.4231755052566673,
      "tokens_seen": 2051276800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038492929292929296,
      "loss": 2.6672,
      "theoretical_loss": 3.423166432009923,
      "tokens_seen": 2051342336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003849090909090909,
      "loss": 2.4185,
      "theoretical_loss": 3.4231573591342057,
      "tokens_seen": 2051407872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848888888888889,
      "loss": 2.4452,
      "theoretical_loss": 3.4231482866294884,
      "tokens_seen": 2051473408
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1129971,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5400750637054443,
      "objective/train/theoretical_loss": 3.4231392144957447,
      "objective/train/tokens_used": 410398176,
      "theoretical_loss": 3.4231392144957447,
      "tokens_seen": 2051538944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848686868686869,
      "loss": 2.544,
      "theoretical_loss": 3.4231392144957447,
      "tokens_seen": 2051538944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038484848484848485,
      "loss": 2.5103,
      "theoretical_loss": 3.423130142732947,
      "tokens_seen": 2051604480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038482828282828286,
      "loss": 2.6622,
      "theoretical_loss": 3.423121071341068,
      "tokens_seen": 2051670016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848080808080808,
      "loss": 2.6093,
      "theoretical_loss": 3.4231120003200806,
      "tokens_seen": 2051735552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038478787878787877,
      "loss": 2.7409,
      "theoretical_loss": 3.4231029296699584,
      "tokens_seen": 2051801088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847676767676768,
      "loss": 2.3119,
      "theoretical_loss": 3.4230938593906743,
      "tokens_seen": 2051866624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038474747474747474,
      "loss": 2.9422,
      "theoretical_loss": 3.423084789482201,
      "tokens_seen": 2051932160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038472727272727275,
      "loss": 2.5798,
      "theoretical_loss": 3.423075719944512,
      "tokens_seen": 2051997696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847070707070707,
      "loss": 2.6104,
      "theoretical_loss": 3.4230666507775798,
      "tokens_seen": 2052063232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846868686868687,
      "loss": 2.6882,
      "theoretical_loss": 3.4230575819813778,
      "tokens_seen": 2052128768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846666666666667,
      "loss": 2.7611,
      "theoretical_loss": 3.4230485135558784,
      "tokens_seen": 2052194304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038464646464646463,
      "loss": 2.6087,
      "theoretical_loss": 3.423039445501055,
      "tokens_seen": 2052259840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038462626262626264,
      "loss": 2.5599,
      "theoretical_loss": 3.4230303778168802,
      "tokens_seen": 2052325376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846060606060606,
      "loss": 2.5465,
      "theoretical_loss": 3.423021310503328,
      "tokens_seen": 2052390912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845858585858586,
      "loss": 2.5225,
      "theoretical_loss": 3.4230122435603705,
      "tokens_seen": 2052456448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038456565656565657,
      "loss": 2.5076,
      "theoretical_loss": 3.423003176987981,
      "tokens_seen": 2052521984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845454545454545,
      "loss": 2.4484,
      "theoretical_loss": 3.422994110786133,
      "tokens_seen": 2052587520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038452525252525254,
      "loss": 2.4406,
      "theoretical_loss": 3.4229850449547987,
      "tokens_seen": 2052653056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038450505050505055,
      "loss": 2.5912,
      "theoretical_loss": 3.422975979493952,
      "tokens_seen": 2052718592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844848484848485,
      "loss": 2.6731,
      "theoretical_loss": 3.422966914403565,
      "tokens_seen": 2052784128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038446464646464646,
      "loss": 2.5045,
      "theoretical_loss": 3.422957849683611,
      "tokens_seen": 2052849664
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844444444444444,
      "loss": 2.5084,
      "theoretical_loss": 3.4229487853340634,
      "tokens_seen": 2052915200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038442424242424243,
      "loss": 2.4632,
      "theoretical_loss": 3.4229397213548953,
      "tokens_seen": 2052980736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038440404040404044,
      "loss": 2.79,
      "theoretical_loss": 3.422930657746079,
      "tokens_seen": 2053046272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843838383838384,
      "loss": 2.653,
      "theoretical_loss": 3.4229215945075886,
      "tokens_seen": 2053111808
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1131091,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.304107666015625,
      "objective/train/theoretical_loss": 3.4229125316393967,
      "objective/train/tokens_used": 412036576,
      "theoretical_loss": 3.4229125316393967,
      "tokens_seen": 2053177344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038436363636363635,
      "loss": 2.3587,
      "theoretical_loss": 3.4229125316393967,
      "tokens_seen": 2053177344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843434343434343,
      "loss": 2.6467,
      "theoretical_loss": 3.4229034691414757,
      "tokens_seen": 2053242880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843232323232324,
      "loss": 2.6997,
      "theoretical_loss": 3.4228944070137994,
      "tokens_seen": 2053308416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038430303030303033,
      "loss": 2.6112,
      "theoretical_loss": 3.422885345256341,
      "tokens_seen": 2053373952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842828282828283,
      "loss": 2.9658,
      "theoretical_loss": 3.4228762838690727,
      "tokens_seen": 2053439488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038426262626262625,
      "loss": 2.5993,
      "theoretical_loss": 3.422867222851968,
      "tokens_seen": 2053505024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842424242424242,
      "loss": 2.3219,
      "theoretical_loss": 3.4228581622050007,
      "tokens_seen": 2053570560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038422222222222227,
      "loss": 2.2982,
      "theoretical_loss": 3.4228491019281426,
      "tokens_seen": 2053636096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842020202020202,
      "loss": 2.5316,
      "theoretical_loss": 3.422840042021368,
      "tokens_seen": 2053701632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841818181818182,
      "loss": 2.5537,
      "theoretical_loss": 3.4228309824846495,
      "tokens_seen": 2053767168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038416161616161614,
      "loss": 2.6209,
      "theoretical_loss": 3.422821923317959,
      "tokens_seen": 2053832704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038414141414141415,
      "loss": 2.3912,
      "theoretical_loss": 3.4228128645212714,
      "tokens_seen": 2053898240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038412121212121216,
      "loss": 2.6452,
      "theoretical_loss": 3.422803806094559,
      "tokens_seen": 2053963776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841010101010101,
      "loss": 2.6819,
      "theoretical_loss": 3.422794748037795,
      "tokens_seen": 2054029312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003840808080808081,
      "loss": 2.8093,
      "theoretical_loss": 3.4227856903509517,
      "tokens_seen": 2054094848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038406060606060603,
      "loss": 2.5679,
      "theoretical_loss": 3.4227766330340033,
      "tokens_seen": 2054160384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038404040404040404,
      "loss": 2.4692,
      "theoretical_loss": 3.4227675760869225,
      "tokens_seen": 2054225920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038402020202020205,
      "loss": 2.4967,
      "theoretical_loss": 3.4227585195096824,
      "tokens_seen": 2054291456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384,
      "loss": 2.5947,
      "theoretical_loss": 3.4227494633022557,
      "tokens_seen": 2054356992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038397979797979797,
      "loss": 2.4263,
      "theoretical_loss": 3.422740407464616,
      "tokens_seen": 2054422528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383959595959596,
      "loss": 2.5193,
      "theoretical_loss": 3.4227313519967364,
      "tokens_seen": 2054488064
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038393939393939393,
      "loss": 2.2884,
      "theoretical_loss": 3.4227222968985904,
      "tokens_seen": 2054553600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038391919191919195,
      "loss": 2.5641,
      "theoretical_loss": 3.4227132421701496,
      "tokens_seen": 2054619136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838989898989899,
      "loss": 2.3804,
      "theoretical_loss": 3.422704187811389,
      "tokens_seen": 2054684672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038387878787878786,
      "loss": 2.5944,
      "theoretical_loss": 3.4226951338222804,
      "tokens_seen": 2054750208
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1131564,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9922235012054443,
      "objective/train/theoretical_loss": 3.422686080202797,
      "objective/train/tokens_used": 413674976,
      "theoretical_loss": 3.422686080202797,
      "tokens_seen": 2054815744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038385858585858587,
      "loss": 2.3964,
      "theoretical_loss": 3.422686080202797,
      "tokens_seen": 2054815744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038383838383838383,
      "loss": 2.4571,
      "theoretical_loss": 3.422677026952913,
      "tokens_seen": 2054881280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038381818181818184,
      "loss": 2.6484,
      "theoretical_loss": 3.4226679740726,
      "tokens_seen": 2054946816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837979797979798,
      "loss": 2.5657,
      "theoretical_loss": 3.4226589215618324,
      "tokens_seen": 2055012352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837777777777778,
      "loss": 2.4732,
      "theoretical_loss": 3.4226498694205825,
      "tokens_seen": 2055077888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038375757575757576,
      "loss": 2.5445,
      "theoretical_loss": 3.4226408176488246,
      "tokens_seen": 2055143424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837373737373737,
      "loss": 2.3885,
      "theoretical_loss": 3.42263176624653,
      "tokens_seen": 2055208960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038371717171717173,
      "loss": 2.7094,
      "theoretical_loss": 3.4226227152136737,
      "tokens_seen": 2055274496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836969696969697,
      "loss": 2.5966,
      "theoretical_loss": 3.4226136645502274,
      "tokens_seen": 2055340032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836767676767677,
      "loss": 2.7365,
      "theoretical_loss": 3.422604614256165,
      "tokens_seen": 2055405568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038365656565656566,
      "loss": 2.5426,
      "theoretical_loss": 3.422595564331459,
      "tokens_seen": 2055471104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836363636363636,
      "loss": 2.4673,
      "theoretical_loss": 3.422586514776084,
      "tokens_seen": 2055536640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836161616161616,
      "loss": 2.4799,
      "theoretical_loss": 3.4225774655900114,
      "tokens_seen": 2055602176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038359595959595963,
      "loss": 2.804,
      "theoretical_loss": 3.422568416773216,
      "tokens_seen": 2055667712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835757575757576,
      "loss": 2.7354,
      "theoretical_loss": 3.422559368325669,
      "tokens_seen": 2055733248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038355555555555555,
      "loss": 2.6043,
      "theoretical_loss": 3.4225503202473453,
      "tokens_seen": 2055798784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038353535353535356,
      "loss": 2.5097,
      "theoretical_loss": 3.4225412725382176,
      "tokens_seen": 2055864320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835151515151515,
      "loss": 2.6064,
      "theoretical_loss": 3.4225322251982586,
      "tokens_seen": 2055929856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038349494949494953,
      "loss": 2.5131,
      "theoretical_loss": 3.4225231782274417,
      "tokens_seen": 2055995392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003834747474747475,
      "loss": 2.4516,
      "theoretical_loss": 3.42251413162574,
      "tokens_seen": 2056060928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038345454545454544,
      "loss": 2.7433,
      "theoretical_loss": 3.4225050853931274,
      "tokens_seen": 2056126464
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038343434343434345,
      "loss": 2.7684,
      "theoretical_loss": 3.4224960395295763,
      "tokens_seen": 2056192000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038341414141414146,
      "loss": 2.5327,
      "theoretical_loss": 3.4224869940350597,
      "tokens_seen": 2056257536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833939393939394,
      "loss": 2.4565,
      "theoretical_loss": 3.4224779489095516,
      "tokens_seen": 2056323072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833737373737374,
      "loss": 2.6484,
      "theoretical_loss": 3.4224689041530247,
      "tokens_seen": 2056388608
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1132626,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2589941024780273,
      "objective/train/theoretical_loss": 3.422459859765452,
      "objective/train/tokens_used": 415313376,
      "theoretical_loss": 3.422459859765452,
      "tokens_seen": 2056454144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038335353535353533,
      "loss": 2.3534,
      "theoretical_loss": 3.422459859765452,
      "tokens_seen": 2056454144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038333333333333334,
      "loss": 2.7033,
      "theoretical_loss": 3.422450815746807,
      "tokens_seen": 2056519680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038331313131313136,
      "loss": 2.8655,
      "theoretical_loss": 3.422441772097063,
      "tokens_seen": 2056585216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832929292929293,
      "loss": 2.6004,
      "theoretical_loss": 3.422432728816193,
      "tokens_seen": 2056650752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038327272727272727,
      "loss": 2.8118,
      "theoretical_loss": 3.42242368590417,
      "tokens_seen": 2056716288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832525252525252,
      "loss": 2.6264,
      "theoretical_loss": 3.422414643360967,
      "tokens_seen": 2056781824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832323232323233,
      "loss": 2.329,
      "theoretical_loss": 3.4224056011865587,
      "tokens_seen": 2056847360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038321212121212125,
      "loss": 2.7078,
      "theoretical_loss": 3.4223965593809167,
      "tokens_seen": 2056912896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831919191919192,
      "loss": 2.6651,
      "theoretical_loss": 3.4223875179440144,
      "tokens_seen": 2056978432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038317171717171716,
      "loss": 2.6018,
      "theoretical_loss": 3.4223784768758256,
      "tokens_seen": 2057043968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831515151515151,
      "loss": 2.7493,
      "theoretical_loss": 3.4223694361763233,
      "tokens_seen": 2057109504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831313131313132,
      "loss": 2.7182,
      "theoretical_loss": 3.422360395845481,
      "tokens_seen": 2057175040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038311111111111114,
      "loss": 2.656,
      "theoretical_loss": 3.422351355883271,
      "tokens_seen": 2057240576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003830909090909091,
      "loss": 2.6502,
      "theoretical_loss": 3.422342316289668,
      "tokens_seen": 2057306112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038307070707070706,
      "loss": 2.4645,
      "theoretical_loss": 3.4223332770646437,
      "tokens_seen": 2057371648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383050505050505,
      "loss": 2.7262,
      "theoretical_loss": 3.422324238208172,
      "tokens_seen": 2057437184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003830303030303031,
      "loss": 2.5653,
      "theoretical_loss": 3.422315199720227,
      "tokens_seen": 2057502720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038301010101010103,
      "loss": 2.7167,
      "theoretical_loss": 3.42230616160078,
      "tokens_seen": 2057568256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382989898989899,
      "loss": 2.6084,
      "theoretical_loss": 3.422297123849806,
      "tokens_seen": 2057633792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038296969696969695,
      "loss": 2.6191,
      "theoretical_loss": 3.4222880864672773,
      "tokens_seen": 2057699328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038294949494949496,
      "loss": 2.7294,
      "theoretical_loss": 3.4222790494531674,
      "tokens_seen": 2057764864
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038292929292929297,
      "loss": 2.4276,
      "theoretical_loss": 3.4222700128074495,
      "tokens_seen": 2057830400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829090909090909,
      "loss": 2.2369,
      "theoretical_loss": 3.4222609765300973,
      "tokens_seen": 2057895936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828888888888889,
      "loss": 2.7929,
      "theoretical_loss": 3.4222519406210834,
      "tokens_seen": 2057961472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038286868686868684,
      "loss": 2.5678,
      "theoretical_loss": 3.422242905080381,
      "tokens_seen": 2058027008
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1133082,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.546114683151245,
      "objective/train/theoretical_loss": 3.422233869907964,
      "objective/train/tokens_used": 416951776,
      "theoretical_loss": 3.422233869907964,
      "tokens_seen": 2058092544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038284848484848485,
      "loss": 2.5294,
      "theoretical_loss": 3.422233869907964,
      "tokens_seen": 2058092544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038282828282828286,
      "loss": 2.6326,
      "theoretical_loss": 3.4222248351038056,
      "tokens_seen": 2058158080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828080808080808,
      "loss": 2.7425,
      "theoretical_loss": 3.4222158006678787,
      "tokens_seen": 2058223616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827878787878788,
      "loss": 2.7415,
      "theoretical_loss": 3.4222067666001563,
      "tokens_seen": 2058289152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827676767676768,
      "loss": 2.4934,
      "theoretical_loss": 3.4221977329006124,
      "tokens_seen": 2058354688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038274747474747474,
      "loss": 2.7173,
      "theoretical_loss": 3.4221886995692197,
      "tokens_seen": 2058420224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038272727272727276,
      "loss": 2.5987,
      "theoretical_loss": 3.4221796666059516,
      "tokens_seen": 2058485760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827070707070707,
      "loss": 2.8897,
      "theoretical_loss": 3.422170634010782,
      "tokens_seen": 2058551296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038268686868686867,
      "loss": 2.9005,
      "theoretical_loss": 3.4221616017836833,
      "tokens_seen": 2058616832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003826666666666667,
      "loss": 2.8517,
      "theoretical_loss": 3.422152569924629,
      "tokens_seen": 2058682368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038264646464646464,
      "loss": 2.697,
      "theoretical_loss": 3.422143538433593,
      "tokens_seen": 2058747904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038262626262626265,
      "loss": 2.5326,
      "theoretical_loss": 3.422134507310548,
      "tokens_seen": 2058813440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003826060606060606,
      "loss": 2.5776,
      "theoretical_loss": 3.422125476555467,
      "tokens_seen": 2058878976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825858585858586,
      "loss": 2.4844,
      "theoretical_loss": 3.4221164461683236,
      "tokens_seen": 2058944512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038256565656565657,
      "loss": 2.6396,
      "theoretical_loss": 3.4221074161490916,
      "tokens_seen": 2059010048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038254545454545453,
      "loss": 2.9819,
      "theoretical_loss": 3.422098386497744,
      "tokens_seen": 2059075584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038252525252525254,
      "loss": 2.5535,
      "theoretical_loss": 3.422089357214254,
      "tokens_seen": 2059141120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825050505050505,
      "loss": 2.7637,
      "theoretical_loss": 3.4220803282985948,
      "tokens_seen": 2059206656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824848484848485,
      "loss": 2.4363,
      "theoretical_loss": 3.4220712997507396,
      "tokens_seen": 2059272192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038246464646464647,
      "loss": 2.6999,
      "theoretical_loss": 3.4220622715706623,
      "tokens_seen": 2059337728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824444444444444,
      "loss": 2.7858,
      "theoretical_loss": 3.4220532437583353,
      "tokens_seen": 2059403264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038242424242424243,
      "loss": 2.5095,
      "theoretical_loss": 3.4220442163137332,
      "tokens_seen": 2059468800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038240404040404044,
      "loss": 2.5913,
      "theoretical_loss": 3.422035189236828,
      "tokens_seen": 2059534336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823838383838384,
      "loss": 2.3923,
      "theoretical_loss": 3.4220261625275943,
      "tokens_seen": 2059599872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038236363636363636,
      "loss": 2.629,
      "theoretical_loss": 3.422017136186004,
      "tokens_seen": 2059665408
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1134328,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9984428882598877,
      "objective/train/theoretical_loss": 3.422008110212032,
      "objective/train/tokens_used": 418590176,
      "theoretical_loss": 3.422008110212032,
      "tokens_seen": 2059730944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823434343434343,
      "loss": 2.7817,
      "theoretical_loss": 3.422008110212032,
      "tokens_seen": 2059730944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823232323232323,
      "loss": 2.604,
      "theoretical_loss": 3.42199908460565,
      "tokens_seen": 2059796480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038230303030303034,
      "loss": 2.6303,
      "theoretical_loss": 3.4219900593668324,
      "tokens_seen": 2059862016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822828282828283,
      "loss": 2.594,
      "theoretical_loss": 3.421981034495553,
      "tokens_seen": 2059927552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038226262626262625,
      "loss": 2.6818,
      "theoretical_loss": 3.421972009991783,
      "tokens_seen": 2059993088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822424242424242,
      "loss": 2.5402,
      "theoretical_loss": 3.4219629858554983,
      "tokens_seen": 2060058624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822222222222223,
      "loss": 2.3345,
      "theoretical_loss": 3.4219539620866706,
      "tokens_seen": 2060124160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038220202020202023,
      "loss": 2.7212,
      "theoretical_loss": 3.421944938685274,
      "tokens_seen": 2060189696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821818181818182,
      "loss": 2.6767,
      "theoretical_loss": 3.4219359156512814,
      "tokens_seen": 2060255232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038216161616161614,
      "loss": 2.4343,
      "theoretical_loss": 3.4219268929846662,
      "tokens_seen": 2060320768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821414141414141,
      "loss": 2.4259,
      "theoretical_loss": 3.4219178706854025,
      "tokens_seen": 2060386304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038212121212121217,
      "loss": 2.6288,
      "theoretical_loss": 3.4219088487534624,
      "tokens_seen": 2060451840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821010101010101,
      "loss": 2.6574,
      "theoretical_loss": 3.4218998271888204,
      "tokens_seen": 2060517376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820808080808081,
      "loss": 2.5343,
      "theoretical_loss": 3.4218908059914495,
      "tokens_seen": 2060582912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038206060606060604,
      "loss": 2.6432,
      "theoretical_loss": 3.4218817851613226,
      "tokens_seen": 2060648448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038204040404040405,
      "loss": 2.6204,
      "theoretical_loss": 3.4218727646984135,
      "tokens_seen": 2060713984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038202020202020206,
      "loss": 2.6521,
      "theoretical_loss": 3.4218637446026956,
      "tokens_seen": 2060779520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382,
      "loss": 2.7458,
      "theoretical_loss": 3.4218547248741427,
      "tokens_seen": 2060845056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038197979797979797,
      "loss": 2.8135,
      "theoretical_loss": 3.4218457055127267,
      "tokens_seen": 2060910592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038195959595959593,
      "loss": 2.5271,
      "theoretical_loss": 3.4218366865184224,
      "tokens_seen": 2060976128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038193939393939394,
      "loss": 2.7021,
      "theoretical_loss": 3.421827667891203,
      "tokens_seen": 2061041664
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038191919191919195,
      "loss": 2.5729,
      "theoretical_loss": 3.4218186496310414,
      "tokens_seen": 2061107200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818989898989899,
      "loss": 2.416,
      "theoretical_loss": 3.4218096317379114,
      "tokens_seen": 2061172736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038187878787878786,
      "loss": 2.6395,
      "theoretical_loss": 3.4218006142117856,
      "tokens_seen": 2061238272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818585858585859,
      "loss": 2.6007,
      "theoretical_loss": 3.4217915970526387,
      "tokens_seen": 2061303808
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1135184,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6123404502868652,
      "objective/train/theoretical_loss": 3.421782580260443,
      "objective/train/tokens_used": 420228576,
      "theoretical_loss": 3.421782580260443,
      "tokens_seen": 2061369344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818383838383839,
      "loss": 2.6084,
      "theoretical_loss": 3.421782580260443,
      "tokens_seen": 2061369344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038181818181818184,
      "loss": 2.6643,
      "theoretical_loss": 3.4217735638351727,
      "tokens_seen": 2061434880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817979797979798,
      "loss": 2.6587,
      "theoretical_loss": 3.4217645477768004,
      "tokens_seen": 2061500416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038177777777777776,
      "loss": 2.6285,
      "theoretical_loss": 3.4217555320853004,
      "tokens_seen": 2061565952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038175757575757577,
      "loss": 2.5553,
      "theoretical_loss": 3.4217465167606456,
      "tokens_seen": 2061631488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817373737373738,
      "loss": 2.6293,
      "theoretical_loss": 3.4217375018028084,
      "tokens_seen": 2061697024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038171717171717174,
      "loss": 2.6729,
      "theoretical_loss": 3.421728487211764,
      "tokens_seen": 2061762560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816969696969697,
      "loss": 2.617,
      "theoretical_loss": 3.4217194729874856,
      "tokens_seen": 2061828096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816767676767677,
      "loss": 2.7,
      "theoretical_loss": 3.4217104591299456,
      "tokens_seen": 2061893632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038165656565656566,
      "loss": 2.7145,
      "theoretical_loss": 3.4217014456391177,
      "tokens_seen": 2061959168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038163636363636367,
      "loss": 2.501,
      "theoretical_loss": 3.421692432514976,
      "tokens_seen": 2062024704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038161616161616163,
      "loss": 2.4279,
      "theoretical_loss": 3.421683419757493,
      "tokens_seen": 2062090240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815959595959596,
      "loss": 2.4535,
      "theoretical_loss": 3.421674407366643,
      "tokens_seen": 2062155776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815757575757576,
      "loss": 2.5424,
      "theoretical_loss": 3.4216653953423988,
      "tokens_seen": 2062221312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038155555555555555,
      "loss": 2.5376,
      "theoretical_loss": 3.4216563836847342,
      "tokens_seen": 2062286848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038153535353535357,
      "loss": 2.6659,
      "theoretical_loss": 3.4216473723936223,
      "tokens_seen": 2062352384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815151515151515,
      "loss": 2.5682,
      "theoretical_loss": 3.421638361469037,
      "tokens_seen": 2062417920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814949494949495,
      "loss": 2.52,
      "theoretical_loss": 3.421629350910951,
      "tokens_seen": 2062483456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814747474747475,
      "loss": 2.4654,
      "theoretical_loss": 3.421620340719339,
      "tokens_seen": 2062548992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038145454545454545,
      "loss": 2.5279,
      "theoretical_loss": 3.421611330894174,
      "tokens_seen": 2062614528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038143434343434346,
      "loss": 2.3815,
      "theoretical_loss": 3.421602321435428,
      "tokens_seen": 2062680064
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814141414141414,
      "loss": 2.8111,
      "theoretical_loss": 3.4215933123430764,
      "tokens_seen": 2062745600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813939393939394,
      "loss": 2.4246,
      "theoretical_loss": 3.4215843036170916,
      "tokens_seen": 2062811136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813737373737374,
      "loss": 2.5496,
      "theoretical_loss": 3.421575295257447,
      "tokens_seen": 2062876672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038135353535353534,
      "loss": 2.5605,
      "theoretical_loss": 3.421566287264117,
      "tokens_seen": 2062942208
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 1135835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9515092372894287,
      "objective/train/theoretical_loss": 3.4215572796370743,
      "objective/train/tokens_used": 421866976,
      "theoretical_loss": 3.4215572796370743,
      "tokens_seen": 2063007744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038133333333333335,
      "loss": 2.7702,
      "theoretical_loss": 3.4215572796370743,
      "tokens_seen": 2063007744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003813131313131313,
      "loss": 2.7661,
      "theoretical_loss": 3.4215482723762927,
      "tokens_seen": 2063073280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812929292929293,
      "loss": 2.5597,
      "theoretical_loss": 3.421539265481745,
      "tokens_seen": 2063138816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812727272727273,
      "loss": 2.8065,
      "theoretical_loss": 3.421530258953406,
      "tokens_seen": 2063204352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038125252525252523,
      "loss": 2.5136,
      "theoretical_loss": 3.4215212527912477,
      "tokens_seen": 2063269888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038123232323232324,
      "loss": 2.638,
      "theoretical_loss": 3.4215122469952446,
      "tokens_seen": 2063335424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038121212121212125,
      "loss": 2.6101,
      "theoretical_loss": 3.4215032415653694,
      "tokens_seen": 2063400960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811919191919192,
      "loss": 2.5277,
      "theoretical_loss": 3.421494236501596,
      "tokens_seen": 2063466496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038117171717171717,
      "loss": 2.683,
      "theoretical_loss": 3.4214852318038984,
      "tokens_seen": 2063532032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811515151515151,
      "loss": 2.6695,
      "theoretical_loss": 3.421476227472249,
      "tokens_seen": 2063597568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038113131313131314,
      "loss": 2.6012,
      "theoretical_loss": 3.421467223506623,
      "tokens_seen": 2063663104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038111111111111115,
      "loss": 2.6724,
      "theoretical_loss": 3.421458219906992,
      "tokens_seen": 2063728640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810909090909091,
      "loss": 2.5402,
      "theoretical_loss": 3.4214492166733304,
      "tokens_seen": 2063794176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038107070707070706,
      "loss": 2.6403,
      "theoretical_loss": 3.421440213805612,
      "tokens_seen": 2063859712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000381050505050505,
      "loss": 2.5712,
      "theoretical_loss": 3.4214312113038092,
      "tokens_seen": 2063925248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810303030303031,
      "loss": 2.7039,
      "theoretical_loss": 3.4214222091678965,
      "tokens_seen": 2063990784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038101010101010104,
      "loss": 2.5423,
      "theoretical_loss": 3.421413207397847,
      "tokens_seen": 2064056320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380989898989899,
      "loss": 2.7038,
      "theoretical_loss": 3.4214042059936345,
      "tokens_seen": 2064121856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038096969696969695,
      "loss": 2.7188,
      "theoretical_loss": 3.4213952049552327,
      "tokens_seen": 2064187392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809494949494949,
      "loss": 2.5841,
      "theoretical_loss": 3.421386204282615,
      "tokens_seen": 2064252928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380929292929293,
      "loss": 2.9913,
      "theoretical_loss": 3.4213772039757537,
      "tokens_seen": 2064318464
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038090909090909093,
      "loss": 2.5172,
      "theoretical_loss": 3.421368204034624,
      "tokens_seen": 2064384000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808888888888889,
      "loss": 2.6884,
      "theoretical_loss": 3.4213592044591987,
      "tokens_seen": 2064449536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038086868686868685,
      "loss": 2.5919,
      "theoretical_loss": 3.421350205249451,
      "tokens_seen": 2064515072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038084848484848486,
      "loss": 2.545,
      "theoretical_loss": 3.421341206405355,
      "tokens_seen": 2064580608
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1136767,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7607131004333496,
      "objective/train/theoretical_loss": 3.421332207926884,
      "objective/train/tokens_used": 423505376,
      "theoretical_loss": 3.421332207926884,
      "tokens_seen": 2064646144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038082828282828287,
      "loss": 2.6132,
      "theoretical_loss": 3.421332207926884,
      "tokens_seen": 2064646144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808080808080808,
      "loss": 2.5181,
      "theoretical_loss": 3.421323209814012,
      "tokens_seen": 2064711680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807878787878788,
      "loss": 2.7202,
      "theoretical_loss": 3.4213142120667124,
      "tokens_seen": 2064777216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038076767676767674,
      "loss": 2.6562,
      "theoretical_loss": 3.421305214684958,
      "tokens_seen": 2064842752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038074747474747475,
      "loss": 2.6323,
      "theoretical_loss": 3.421296217668723,
      "tokens_seen": 2064908288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038072727272727276,
      "loss": 2.5255,
      "theoretical_loss": 3.4212872210179803,
      "tokens_seen": 2064973824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807070707070707,
      "loss": 2.6225,
      "theoretical_loss": 3.4212782247327045,
      "tokens_seen": 2065039360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806868686868687,
      "loss": 2.7296,
      "theoretical_loss": 3.4212692288128683,
      "tokens_seen": 2065104896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806666666666667,
      "loss": 2.5503,
      "theoretical_loss": 3.4212602332584456,
      "tokens_seen": 2065170432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038064646464646464,
      "loss": 2.7646,
      "theoretical_loss": 3.4212512380694102,
      "tokens_seen": 2065235968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038062626262626265,
      "loss": 2.461,
      "theoretical_loss": 3.421242243245735,
      "tokens_seen": 2065301504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806060606060606,
      "loss": 2.5117,
      "theoretical_loss": 3.421233248787394,
      "tokens_seen": 2065367040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038058585858585857,
      "loss": 2.7507,
      "theoretical_loss": 3.421224254694361,
      "tokens_seen": 2065432576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805656565656566,
      "loss": 2.4589,
      "theoretical_loss": 3.4212152609666093,
      "tokens_seen": 2065498112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038054545454545454,
      "loss": 2.7945,
      "theoretical_loss": 3.421206267604112,
      "tokens_seen": 2065563648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038052525252525255,
      "loss": 2.7459,
      "theoretical_loss": 3.4211972746068438,
      "tokens_seen": 2065629184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805050505050505,
      "loss": 2.5679,
      "theoretical_loss": 3.421188281974777,
      "tokens_seen": 2065694720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804848484848485,
      "loss": 2.7287,
      "theoretical_loss": 3.4211792897078865,
      "tokens_seen": 2065760256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038046464646464647,
      "loss": 2.8253,
      "theoretical_loss": 3.4211702978061447,
      "tokens_seen": 2065825792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038044444444444443,
      "loss": 2.938,
      "theoretical_loss": 3.421161306269526,
      "tokens_seen": 2065891328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038042424242424244,
      "loss": 2.7181,
      "theoretical_loss": 3.421152315098003,
      "tokens_seen": 2065956864
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804040404040404,
      "loss": 2.5896,
      "theoretical_loss": 3.4211433242915508,
      "tokens_seen": 2066022400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803838383838384,
      "loss": 2.5979,
      "theoretical_loss": 3.4211343338501417,
      "tokens_seen": 2066087936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038036363636363636,
      "loss": 2.7055,
      "theoretical_loss": 3.42112534377375,
      "tokens_seen": 2066153472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803434343434343,
      "loss": 2.8343,
      "theoretical_loss": 3.421116354062349,
      "tokens_seen": 2066219008
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1137905,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.325758457183838,
      "objective/train/theoretical_loss": 3.4211073647159123,
      "objective/train/tokens_used": 425143776,
      "theoretical_loss": 3.4211073647159123,
      "tokens_seen": 2066284544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038032323232323233,
      "loss": 2.7325,
      "theoretical_loss": 3.4211073647159123,
      "tokens_seen": 2066284544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038030303030303034,
      "loss": 2.5732,
      "theoretical_loss": 3.421098375734414,
      "tokens_seen": 2066350080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802828282828283,
      "loss": 2.7549,
      "theoretical_loss": 3.4210893871178265,
      "tokens_seen": 2066415616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038026262626262626,
      "loss": 2.768,
      "theoretical_loss": 3.421080398866125,
      "tokens_seen": 2066481152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802424242424242,
      "loss": 2.6855,
      "theoretical_loss": 3.4210714109792817,
      "tokens_seen": 2066546688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802222222222222,
      "loss": 2.7331,
      "theoretical_loss": 3.421062423457271,
      "tokens_seen": 2066612224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038020202020202024,
      "loss": 2.4095,
      "theoretical_loss": 3.421053436300067,
      "tokens_seen": 2066677760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003801818181818182,
      "loss": 2.695,
      "theoretical_loss": 3.421044449507642,
      "tokens_seen": 2066743296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038016161616161615,
      "loss": 2.7255,
      "theoretical_loss": 3.4210354630799706,
      "tokens_seen": 2066808832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038014141414141416,
      "loss": 2.5248,
      "theoretical_loss": 3.4210264770170262,
      "tokens_seen": 2066874368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038012121212121217,
      "loss": 2.6969,
      "theoretical_loss": 3.421017491318782,
      "tokens_seen": 2066939904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038010101010101013,
      "loss": 2.613,
      "theoretical_loss": 3.4210085059852124,
      "tokens_seen": 2067005440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800808080808081,
      "loss": 2.6488,
      "theoretical_loss": 3.4209995210162907,
      "tokens_seen": 2067070976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038006060606060604,
      "loss": 2.6582,
      "theoretical_loss": 3.42099053641199,
      "tokens_seen": 2067136512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038004040404040405,
      "loss": 2.4411,
      "theoretical_loss": 3.420981552172285,
      "tokens_seen": 2067202048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038002020202020206,
      "loss": 2.6109,
      "theoretical_loss": 3.420972568297149,
      "tokens_seen": 2067267584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038,
      "loss": 2.5581,
      "theoretical_loss": 3.4209635847865547,
      "tokens_seen": 2067333120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379979797979798,
      "loss": 2.7596,
      "theoretical_loss": 3.420954601640477,
      "tokens_seen": 2067398656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037995959595959593,
      "loss": 2.564,
      "theoretical_loss": 3.420945618858889,
      "tokens_seen": 2067464192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379939393939394,
      "loss": 2.7477,
      "theoretical_loss": 3.4209366364417644,
      "tokens_seen": 2067529728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037991919191919196,
      "loss": 2.4975,
      "theoretical_loss": 3.420927654389077,
      "tokens_seen": 2067595264
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798989898989899,
      "loss": 2.7735,
      "theoretical_loss": 3.4209186727008,
      "tokens_seen": 2067660800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037987878787878787,
      "loss": 2.5042,
      "theoretical_loss": 3.4209096913769073,
      "tokens_seen": 2067726336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037985858585858583,
      "loss": 2.7599,
      "theoretical_loss": 3.420900710417373,
      "tokens_seen": 2067791872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798383838383839,
      "loss": 2.7375,
      "theoretical_loss": 3.4208917298221704,
      "tokens_seen": 2067857408
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1138616,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.164562940597534,
      "objective/train/theoretical_loss": 3.4208827495912733,
      "objective/train/tokens_used": 426782176,
      "theoretical_loss": 3.4208827495912733,
      "tokens_seen": 2067922944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037981818181818185,
      "loss": 2.6866,
      "theoretical_loss": 3.4208827495912733,
      "tokens_seen": 2067922944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797979797979798,
      "loss": 2.5767,
      "theoretical_loss": 3.420873769724655,
      "tokens_seen": 2067988480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037977777777777776,
      "loss": 2.4466,
      "theoretical_loss": 3.4208647902222893,
      "tokens_seen": 2068054016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797575757575757,
      "loss": 2.6539,
      "theoretical_loss": 3.4208558110841505,
      "tokens_seen": 2068119552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797373737373738,
      "loss": 2.6817,
      "theoretical_loss": 3.420846832310212,
      "tokens_seen": 2068185088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037971717171717174,
      "loss": 2.6445,
      "theoretical_loss": 3.420837853900447,
      "tokens_seen": 2068250624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796969696969697,
      "loss": 2.6812,
      "theoretical_loss": 3.4208288758548293,
      "tokens_seen": 2068316160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037967676767676766,
      "loss": 2.5428,
      "theoretical_loss": 3.420819898173333,
      "tokens_seen": 2068381696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037965656565656567,
      "loss": 2.6958,
      "theoretical_loss": 3.420810920855932,
      "tokens_seen": 2068447232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796363636363637,
      "loss": 2.5625,
      "theoretical_loss": 3.4208019439025987,
      "tokens_seen": 2068512768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037961616161616163,
      "loss": 2.7598,
      "theoretical_loss": 3.4207929673133086,
      "tokens_seen": 2068578304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795959595959596,
      "loss": 2.6718,
      "theoretical_loss": 3.420783991088034,
      "tokens_seen": 2068643840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037957575757575755,
      "loss": 2.8103,
      "theoretical_loss": 3.4207750152267495,
      "tokens_seen": 2068709376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037955555555555556,
      "loss": 2.5705,
      "theoretical_loss": 3.420766039729428,
      "tokens_seen": 2068774912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037953535353535357,
      "loss": 2.6372,
      "theoretical_loss": 3.420757064596044,
      "tokens_seen": 2068840448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037951515151515153,
      "loss": 2.8598,
      "theoretical_loss": 3.4207480898265707,
      "tokens_seen": 2068905984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794949494949495,
      "loss": 2.3985,
      "theoretical_loss": 3.4207391154209823,
      "tokens_seen": 2068971520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794747474747475,
      "loss": 2.8237,
      "theoretical_loss": 3.4207301413792512,
      "tokens_seen": 2069037056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037945454545454545,
      "loss": 2.5604,
      "theoretical_loss": 3.420721167701353,
      "tokens_seen": 2069102592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037943434343434346,
      "loss": 2.4821,
      "theoretical_loss": 3.4207121943872605,
      "tokens_seen": 2069168128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794141414141414,
      "loss": 2.6009,
      "theoretical_loss": 3.4207032214369475,
      "tokens_seen": 2069233664
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793939393939394,
      "loss": 2.7229,
      "theoretical_loss": 3.4206942488503875,
      "tokens_seen": 2069299200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793737373737374,
      "loss": 2.7748,
      "theoretical_loss": 3.4206852766275544,
      "tokens_seen": 2069364736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037935353535353535,
      "loss": 2.6456,
      "theoretical_loss": 3.420676304768422,
      "tokens_seen": 2069430272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037933333333333336,
      "loss": 2.8453,
      "theoretical_loss": 3.420667333272964,
      "tokens_seen": 2069495808
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1139859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9761197566986084,
      "objective/train/theoretical_loss": 3.420658362141154,
      "objective/train/tokens_used": 428420576,
      "theoretical_loss": 3.420658362141154,
      "tokens_seen": 2069561344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793131313131313,
      "loss": 2.645,
      "theoretical_loss": 3.420658362141154,
      "tokens_seen": 2069561344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792929292929293,
      "loss": 2.4449,
      "theoretical_loss": 3.420649391372966,
      "tokens_seen": 2069626880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792727272727273,
      "loss": 2.6226,
      "theoretical_loss": 3.4206404209683736,
      "tokens_seen": 2069692416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037925252525252524,
      "loss": 2.5629,
      "theoretical_loss": 3.4206314509273508,
      "tokens_seen": 2069757952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037923232323232325,
      "loss": 2.7043,
      "theoretical_loss": 3.4206224812498713,
      "tokens_seen": 2069823488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792121212121212,
      "loss": 2.7883,
      "theoretical_loss": 3.4206135119359082,
      "tokens_seen": 2069889024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791919191919192,
      "loss": 2.4786,
      "theoretical_loss": 3.420604542985436,
      "tokens_seen": 2069954560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791717171717172,
      "loss": 2.811,
      "theoretical_loss": 3.420595574398429,
      "tokens_seen": 2070020096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037915151515151513,
      "loss": 2.5502,
      "theoretical_loss": 3.420586606174859,
      "tokens_seen": 2070085632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037913131313131314,
      "loss": 2.7804,
      "theoretical_loss": 3.4205776383147013,
      "tokens_seen": 2070151168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037911111111111115,
      "loss": 2.5411,
      "theoretical_loss": 3.4205686708179295,
      "tokens_seen": 2070216704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790909090909091,
      "loss": 2.6516,
      "theoretical_loss": 3.4205597036845177,
      "tokens_seen": 2070282240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037907070707070707,
      "loss": 2.5921,
      "theoretical_loss": 3.4205507369144383,
      "tokens_seen": 2070347776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379050505050505,
      "loss": 2.7129,
      "theoretical_loss": 3.4205417705076666,
      "tokens_seen": 2070413312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037903030303030303,
      "loss": 2.4977,
      "theoretical_loss": 3.420532804464175,
      "tokens_seen": 2070478848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037901010101010105,
      "loss": 2.6412,
      "theoretical_loss": 3.4205238387839385,
      "tokens_seen": 2070544384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378989898989899,
      "loss": 2.4797,
      "theoretical_loss": 3.4205148734669306,
      "tokens_seen": 2070609920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037896969696969696,
      "loss": 2.4803,
      "theoretical_loss": 3.4205059085131246,
      "tokens_seen": 2070675456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789494949494949,
      "loss": 2.5416,
      "theoretical_loss": 3.420496943922495,
      "tokens_seen": 2070740992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378929292929293,
      "loss": 2.7728,
      "theoretical_loss": 3.4204879796950145,
      "tokens_seen": 2070806528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037890909090909094,
      "loss": 2.6949,
      "theoretical_loss": 3.420479015830658,
      "tokens_seen": 2070872064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788888888888889,
      "loss": 2.6882,
      "theoretical_loss": 3.4204700523293985,
      "tokens_seen": 2070937600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037886868686868685,
      "loss": 2.7507,
      "theoretical_loss": 3.4204610891912104,
      "tokens_seen": 2071003136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788484848484848,
      "loss": 2.6074,
      "theoretical_loss": 3.4204521264160674,
      "tokens_seen": 2071068672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788282828282829,
      "loss": 2.6161,
      "theoretical_loss": 3.4204431640039434,
      "tokens_seen": 2071134208
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1140861,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.942692518234253,
      "objective/train/theoretical_loss": 3.4204342019548113,
      "objective/train/tokens_used": 430058976,
      "theoretical_loss": 3.4204342019548113,
      "tokens_seen": 2071199744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037880808080808083,
      "loss": 2.9214,
      "theoretical_loss": 3.4204342019548113,
      "tokens_seen": 2071199744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787878787878788,
      "loss": 2.4544,
      "theoretical_loss": 3.4204252402686466,
      "tokens_seen": 2071265280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037876767676767674,
      "loss": 2.8409,
      "theoretical_loss": 3.4204162789454213,
      "tokens_seen": 2071330816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037874747474747476,
      "loss": 2.4468,
      "theoretical_loss": 3.42040731798511,
      "tokens_seen": 2071396352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037872727272727277,
      "loss": 2.4653,
      "theoretical_loss": 3.420398357387687,
      "tokens_seen": 2071461888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787070707070707,
      "loss": 2.7628,
      "theoretical_loss": 3.4203893971531256,
      "tokens_seen": 2071527424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786868686868687,
      "loss": 2.4264,
      "theoretical_loss": 3.4203804372814,
      "tokens_seen": 2071592960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037866666666666664,
      "loss": 2.6893,
      "theoretical_loss": 3.420371477772483,
      "tokens_seen": 2071658496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037864646464646465,
      "loss": 2.8688,
      "theoretical_loss": 3.4203625186263498,
      "tokens_seen": 2071724032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037862626262626266,
      "loss": 2.7391,
      "theoretical_loss": 3.4203535598429737,
      "tokens_seen": 2071789568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786060606060606,
      "loss": 2.783,
      "theoretical_loss": 3.4203446014223284,
      "tokens_seen": 2071855104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785858585858586,
      "loss": 2.6383,
      "theoretical_loss": 3.4203356433643877,
      "tokens_seen": 2071920640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785656565656566,
      "loss": 2.61,
      "theoretical_loss": 3.4203266856691252,
      "tokens_seen": 2071986176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037854545454545454,
      "loss": 2.6615,
      "theoretical_loss": 3.4203177283365154,
      "tokens_seen": 2072051712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037852525252525255,
      "loss": 2.6566,
      "theoretical_loss": 3.420308771366532,
      "tokens_seen": 2072117248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785050505050505,
      "loss": 2.6085,
      "theoretical_loss": 3.420299814759148,
      "tokens_seen": 2072182784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037848484848484847,
      "loss": 2.6634,
      "theoretical_loss": 3.420290858514339,
      "tokens_seen": 2072248320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784646464646465,
      "loss": 2.6108,
      "theoretical_loss": 3.4202819026320768,
      "tokens_seen": 2072313856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784444444444445,
      "loss": 2.7928,
      "theoretical_loss": 3.4202729471123368,
      "tokens_seen": 2072379392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037842424242424244,
      "loss": 2.8207,
      "theoretical_loss": 3.4202639919550926,
      "tokens_seen": 2072444928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784040404040404,
      "loss": 2.9374,
      "theoretical_loss": 3.420255037160317,
      "tokens_seen": 2072510464
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783838383838384,
      "loss": 2.6872,
      "theoretical_loss": 3.420246082727985,
      "tokens_seen": 2072576000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037836363636363637,
      "loss": 2.6963,
      "theoretical_loss": 3.42023712865807,
      "tokens_seen": 2072641536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783434343434344,
      "loss": 2.674,
      "theoretical_loss": 3.4202281749505463,
      "tokens_seen": 2072707072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037832323232323234,
      "loss": 2.6523,
      "theoretical_loss": 3.4202192216053873,
      "tokens_seen": 2072772608
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1141535,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.150759696960449,
      "objective/train/theoretical_loss": 3.420210268622567,
      "objective/train/tokens_used": 431697376,
      "theoretical_loss": 3.420210268622567,
      "tokens_seen": 2072838144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783030303030303,
      "loss": 2.7664,
      "theoretical_loss": 3.420210268622567,
      "tokens_seen": 2072838144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782828282828283,
      "loss": 2.7415,
      "theoretical_loss": 3.4202013160020597,
      "tokens_seen": 2072903680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037826262626262626,
      "loss": 2.9062,
      "theoretical_loss": 3.4201923637438383,
      "tokens_seen": 2072969216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782424242424243,
      "loss": 2.8399,
      "theoretical_loss": 3.4201834118478773,
      "tokens_seen": 2073034752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037822222222222223,
      "loss": 2.7105,
      "theoretical_loss": 3.4201744603141515,
      "tokens_seen": 2073100288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782020202020202,
      "loss": 2.4727,
      "theoretical_loss": 3.420165509142633,
      "tokens_seen": 2073165824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781818181818182,
      "loss": 2.3444,
      "theoretical_loss": 3.420156558333297,
      "tokens_seen": 2073231360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037816161616161615,
      "loss": 2.5233,
      "theoretical_loss": 3.4201476078861166,
      "tokens_seen": 2073296896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037814141414141417,
      "loss": 2.4147,
      "theoretical_loss": 3.420138657801066,
      "tokens_seen": 2073362432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781212121212121,
      "loss": 2.7456,
      "theoretical_loss": 3.42012970807812,
      "tokens_seen": 2073427968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037810101010101013,
      "loss": 2.935,
      "theoretical_loss": 3.4201207587172506,
      "tokens_seen": 2073493504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003780808080808081,
      "loss": 2.6897,
      "theoretical_loss": 3.4201118097184335,
      "tokens_seen": 2073559040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037806060606060605,
      "loss": 2.645,
      "theoretical_loss": 3.4201028610816415,
      "tokens_seen": 2073624576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037804040404040406,
      "loss": 2.6722,
      "theoretical_loss": 3.4200939128068493,
      "tokens_seen": 2073690112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378020202020202,
      "loss": 2.544,
      "theoretical_loss": 3.4200849648940306,
      "tokens_seen": 2073755648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378,
      "loss": 2.7328,
      "theoretical_loss": 3.420076017343159,
      "tokens_seen": 2073821184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377979797979798,
      "loss": 3.0095,
      "theoretical_loss": 3.420067070154208,
      "tokens_seen": 2073886720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037795959595959594,
      "loss": 2.5682,
      "theoretical_loss": 3.4200581233271525,
      "tokens_seen": 2073952256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037793939393939395,
      "loss": 2.576,
      "theoretical_loss": 3.420049176861966,
      "tokens_seen": 2074017792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037791919191919196,
      "loss": 2.7552,
      "theoretical_loss": 3.420040230758622,
      "tokens_seen": 2074083328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778989898989899,
      "loss": 2.6943,
      "theoretical_loss": 3.420031285017096,
      "tokens_seen": 2074148864
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778787878787879,
      "loss": 2.6718,
      "theoretical_loss": 3.42002233963736,
      "tokens_seen": 2074214400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037785858585858583,
      "loss": 2.6333,
      "theoretical_loss": 3.4200133946193887,
      "tokens_seen": 2074279936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037783838383838384,
      "loss": 2.4596,
      "theoretical_loss": 3.420004449963156,
      "tokens_seen": 2074345472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037781818181818186,
      "loss": 2.4134,
      "theoretical_loss": 3.419995505668637,
      "tokens_seen": 2074411008
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1142675,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8896119594573975,
      "objective/train/theoretical_loss": 3.4199865617358034,
      "objective/train/tokens_used": 433335776,
      "theoretical_loss": 3.4199865617358034,
      "tokens_seen": 2074476544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777979797979798,
      "loss": 2.7336,
      "theoretical_loss": 3.4199865617358034,
      "tokens_seen": 2074476544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037777777777777777,
      "loss": 2.6664,
      "theoretical_loss": 3.419977618164631,
      "tokens_seen": 2074542080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777575757575757,
      "loss": 2.5695,
      "theoretical_loss": 3.4199686749550926,
      "tokens_seen": 2074607616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777373737373738,
      "loss": 2.7597,
      "theoretical_loss": 3.419959732107163,
      "tokens_seen": 2074673152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037771717171717175,
      "loss": 2.3906,
      "theoretical_loss": 3.419950789620816,
      "tokens_seen": 2074738688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776969696969697,
      "loss": 2.5256,
      "theoretical_loss": 3.419941847496025,
      "tokens_seen": 2074804224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037767676767676766,
      "loss": 2.6481,
      "theoretical_loss": 3.419932905732764,
      "tokens_seen": 2074869760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776565656565656,
      "loss": 2.6393,
      "theoretical_loss": 3.4199239643310078,
      "tokens_seen": 2074935296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776363636363637,
      "loss": 2.4312,
      "theoretical_loss": 3.41991502329073,
      "tokens_seen": 2075000832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037761616161616164,
      "loss": 2.6489,
      "theoretical_loss": 3.4199060826119037,
      "tokens_seen": 2075066368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775959595959596,
      "loss": 2.7394,
      "theoretical_loss": 3.419897142294504,
      "tokens_seen": 2075131904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037757575757575755,
      "loss": 2.649,
      "theoretical_loss": 3.4198882023385044,
      "tokens_seen": 2075197440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037755555555555557,
      "loss": 2.7409,
      "theoretical_loss": 3.419879262743879,
      "tokens_seen": 2075262976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775353535353536,
      "loss": 2.7992,
      "theoretical_loss": 3.419870323510602,
      "tokens_seen": 2075328512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037751515151515153,
      "loss": 2.7074,
      "theoretical_loss": 3.4198613846386468,
      "tokens_seen": 2075394048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003774949494949495,
      "loss": 2.4963,
      "theoretical_loss": 3.4198524461279876,
      "tokens_seen": 2075459584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037747474747474745,
      "loss": 2.8299,
      "theoretical_loss": 3.419843507978599,
      "tokens_seen": 2075525120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037745454545454546,
      "loss": 2.6015,
      "theoretical_loss": 3.419834570190454,
      "tokens_seen": 2075590656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037743434343434347,
      "loss": 2.6814,
      "theoretical_loss": 3.4198256327635272,
      "tokens_seen": 2075656192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003774141414141414,
      "loss": 2.8352,
      "theoretical_loss": 3.4198166956977927,
      "tokens_seen": 2075721728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773939393939394,
      "loss": 2.7546,
      "theoretical_loss": 3.4198077589932234,
      "tokens_seen": 2075787264
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773737373737374,
      "loss": 2.6698,
      "theoretical_loss": 3.4197988226497955,
      "tokens_seen": 2075852800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037735353535353535,
      "loss": 2.7075,
      "theoretical_loss": 3.4197898866674805,
      "tokens_seen": 2075918336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037733333333333336,
      "loss": 2.5608,
      "theoretical_loss": 3.4197809510462545,
      "tokens_seen": 2075983872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773131313131313,
      "loss": 2.3638,
      "theoretical_loss": 3.41977201578609,
      "tokens_seen": 2076049408
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1143849,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.351544141769409,
      "objective/train/theoretical_loss": 3.4197630808869617,
      "objective/train/tokens_used": 434974176,
      "theoretical_loss": 3.4197630808869617,
      "tokens_seen": 2076114944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772929292929293,
      "loss": 2.6191,
      "theoretical_loss": 3.4197630808869617,
      "tokens_seen": 2076114944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772727272727273,
      "loss": 2.8404,
      "theoretical_loss": 3.4197541463488434,
      "tokens_seen": 2076180480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037725252525252524,
      "loss": 2.7983,
      "theoretical_loss": 3.41974521217171,
      "tokens_seen": 2076246016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037723232323232325,
      "loss": 2.6373,
      "theoretical_loss": 3.419736278355534,
      "tokens_seen": 2076311552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772121212121212,
      "loss": 2.6044,
      "theoretical_loss": 3.4197273449002905,
      "tokens_seen": 2076377088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771919191919192,
      "loss": 2.3621,
      "theoretical_loss": 3.4197184118059534,
      "tokens_seen": 2076442624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771717171717172,
      "loss": 2.8032,
      "theoretical_loss": 3.419709479072496,
      "tokens_seen": 2076508160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037715151515151514,
      "loss": 2.7246,
      "theoretical_loss": 3.419700546699893,
      "tokens_seen": 2076573696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037713131313131315,
      "loss": 2.773,
      "theoretical_loss": 3.4196916146881184,
      "tokens_seen": 2076639232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771111111111111,
      "loss": 2.87,
      "theoretical_loss": 3.4196826830371467,
      "tokens_seen": 2076704768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770909090909091,
      "loss": 2.6496,
      "theoretical_loss": 3.419673751746951,
      "tokens_seen": 2076770304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037707070707070707,
      "loss": 2.5852,
      "theoretical_loss": 3.4196648208175056,
      "tokens_seen": 2076835840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037705050505050503,
      "loss": 2.7811,
      "theoretical_loss": 3.4196558902487846,
      "tokens_seen": 2076901376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037703030303030304,
      "loss": 2.5733,
      "theoretical_loss": 3.4196469600407626,
      "tokens_seen": 2076966912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037701010101010105,
      "loss": 2.5622,
      "theoretical_loss": 3.419638030193413,
      "tokens_seen": 2077032448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000376989898989899,
      "loss": 2.6142,
      "theoretical_loss": 3.4196291007067097,
      "tokens_seen": 2077097984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037696969696969696,
      "loss": 2.6837,
      "theoretical_loss": 3.419620171580627,
      "tokens_seen": 2077163520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003769494949494949,
      "loss": 2.5987,
      "theoretical_loss": 3.4196112428151397,
      "tokens_seen": 2077229056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037692929292929293,
      "loss": 2.7628,
      "theoretical_loss": 3.419602314410221,
      "tokens_seen": 2077294592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037690909090909094,
      "loss": 2.5048,
      "theoretical_loss": 3.419593386365845,
      "tokens_seen": 2077360128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768888888888889,
      "loss": 2.5455,
      "theoretical_loss": 3.4195844586819857,
      "tokens_seen": 2077425664
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037686868686868686,
      "loss": 2.6995,
      "theoretical_loss": 3.4195755313586176,
      "tokens_seen": 2077491200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768484848484848,
      "loss": 2.901,
      "theoretical_loss": 3.419566604395715,
      "tokens_seen": 2077556736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768282828282829,
      "loss": 2.6293,
      "theoretical_loss": 3.419557677793251,
      "tokens_seen": 2077622272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037680808080808084,
      "loss": 2.4327,
      "theoretical_loss": 3.4195487515512006,
      "tokens_seen": 2077687808
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1144478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4185636043548584,
      "objective/train/theoretical_loss": 3.4195398256695375,
      "objective/train/tokens_used": 436612576,
      "theoretical_loss": 3.4195398256695375,
      "tokens_seen": 2077753344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767878787878788,
      "loss": 2.756,
      "theoretical_loss": 3.4195398256695375,
      "tokens_seen": 2077753344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037676767676767675,
      "loss": 2.6662,
      "theoretical_loss": 3.4195309001482355,
      "tokens_seen": 2077818880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767474747474747,
      "loss": 2.5164,
      "theoretical_loss": 3.419521974987269,
      "tokens_seen": 2077884416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037672727272727277,
      "loss": 2.7099,
      "theoretical_loss": 3.4195130501866124,
      "tokens_seen": 2077949952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037670707070707073,
      "loss": 2.5342,
      "theoretical_loss": 3.4195041257462395,
      "tokens_seen": 2078015488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766868686868687,
      "loss": 2.5396,
      "theoretical_loss": 3.419495201666124,
      "tokens_seen": 2078081024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037666666666666664,
      "loss": 2.5388,
      "theoretical_loss": 3.4194862779462407,
      "tokens_seen": 2078146560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766464646464647,
      "loss": 2.6906,
      "theoretical_loss": 3.4194773545865633,
      "tokens_seen": 2078212096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037662626262626266,
      "loss": 2.7033,
      "theoretical_loss": 3.419468431587066,
      "tokens_seen": 2078277632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766060606060606,
      "loss": 2.6427,
      "theoretical_loss": 3.419459508947723,
      "tokens_seen": 2078343168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765858585858586,
      "loss": 2.6233,
      "theoretical_loss": 3.419450586668508,
      "tokens_seen": 2078408704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037656565656565654,
      "loss": 2.671,
      "theoretical_loss": 3.419441664749395,
      "tokens_seen": 2078474240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765454545454546,
      "loss": 2.5917,
      "theoretical_loss": 3.419432743190359,
      "tokens_seen": 2078539776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037652525252525256,
      "loss": 2.7356,
      "theoretical_loss": 3.4194238219913737,
      "tokens_seen": 2078605312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765050505050505,
      "loss": 2.5018,
      "theoretical_loss": 3.419414901152413,
      "tokens_seen": 2078670848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037648484848484847,
      "loss": 2.6064,
      "theoretical_loss": 3.4194059806734507,
      "tokens_seen": 2078736384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037646464646464643,
      "loss": 2.8587,
      "theoretical_loss": 3.419397060554462,
      "tokens_seen": 2078801920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764444444444445,
      "loss": 2.5207,
      "theoretical_loss": 3.41938814079542,
      "tokens_seen": 2078867456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037642424242424245,
      "loss": 2.644,
      "theoretical_loss": 3.4193792213962992,
      "tokens_seen": 2078932992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764040404040404,
      "loss": 2.4817,
      "theoretical_loss": 3.4193703023570743,
      "tokens_seen": 2078998528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037638383838383836,
      "loss": 2.4036,
      "theoretical_loss": 3.4193613836777184,
      "tokens_seen": 2079064064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763636363636364,
      "loss": 2.5825,
      "theoretical_loss": 3.419352465358206,
      "tokens_seen": 2079129600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763434343434344,
      "loss": 2.6791,
      "theoretical_loss": 3.419343547398512,
      "tokens_seen": 2079195136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037632323232323234,
      "loss": 2.6596,
      "theoretical_loss": 3.4193346297986094,
      "tokens_seen": 2079260672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763030303030303,
      "loss": 2.6833,
      "theoretical_loss": 3.419325712558473,
      "tokens_seen": 2079326208
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 1145756,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6733858585357666,
      "objective/train/theoretical_loss": 3.4193167956780766,
      "objective/train/tokens_used": 438250976,
      "theoretical_loss": 3.4193167956780766,
      "tokens_seen": 2079391744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037628282828282826,
      "loss": 2.4661,
      "theoretical_loss": 3.4193167956780766,
      "tokens_seen": 2079391744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037626262626262627,
      "loss": 2.744,
      "theoretical_loss": 3.4193078791573948,
      "tokens_seen": 2079457280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762424242424243,
      "loss": 2.611,
      "theoretical_loss": 3.4192989629964012,
      "tokens_seen": 2079522816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037622222222222224,
      "loss": 2.7055,
      "theoretical_loss": 3.4192900471950707,
      "tokens_seen": 2079588352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762020202020202,
      "loss": 2.6647,
      "theoretical_loss": 3.419281131753377,
      "tokens_seen": 2079653888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003761818181818182,
      "loss": 2.732,
      "theoretical_loss": 3.419272216671294,
      "tokens_seen": 2079719424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037616161616161616,
      "loss": 2.6474,
      "theoretical_loss": 3.4192633019487966,
      "tokens_seen": 2079784960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037614141414141417,
      "loss": 2.6054,
      "theoretical_loss": 3.4192543875858576,
      "tokens_seen": 2079850496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037612121212121213,
      "loss": 2.714,
      "theoretical_loss": 3.419245473582453,
      "tokens_seen": 2079916032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003761010101010101,
      "loss": 2.6577,
      "theoretical_loss": 3.4192365599385557,
      "tokens_seen": 2079981568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760808080808081,
      "loss": 2.7123,
      "theoretical_loss": 3.4192276466541403,
      "tokens_seen": 2080047104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037606060606060605,
      "loss": 2.9202,
      "theoretical_loss": 3.4192187337291804,
      "tokens_seen": 2080112640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037604040404040406,
      "loss": 2.7607,
      "theoretical_loss": 3.4192098211636512,
      "tokens_seen": 2080178176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000376020202020202,
      "loss": 2.6153,
      "theoretical_loss": 3.419200908957526,
      "tokens_seen": 2080243712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037600000000000003,
      "loss": 2.6651,
      "theoretical_loss": 3.41919199711078,
      "tokens_seen": 2080309248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375979797979798,
      "loss": 2.7883,
      "theoretical_loss": 3.419183085623386,
      "tokens_seen": 2080374784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037595959595959595,
      "loss": 2.5957,
      "theoretical_loss": 3.4191741744953195,
      "tokens_seen": 2080440320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037593939393939396,
      "loss": 2.4795,
      "theoretical_loss": 3.419165263726554,
      "tokens_seen": 2080505856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003759191919191919,
      "loss": 2.4591,
      "theoretical_loss": 3.4191563533170637,
      "tokens_seen": 2080571392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758989898989899,
      "loss": 2.5381,
      "theoretical_loss": 3.419147443266823,
      "tokens_seen": 2080636928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758787878787879,
      "loss": 2.76,
      "theoretical_loss": 3.4191385335758055,
      "tokens_seen": 2080702464
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037585858585858584,
      "loss": 2.8718,
      "theoretical_loss": 3.419129624243986,
      "tokens_seen": 2080768000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037583838383838385,
      "loss": 2.6215,
      "theoretical_loss": 3.4191207152713394,
      "tokens_seen": 2080833536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037581818181818186,
      "loss": 2.7257,
      "theoretical_loss": 3.4191118066578383,
      "tokens_seen": 2080899072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757979797979798,
      "loss": 2.7604,
      "theoretical_loss": 3.419102898403459,
      "tokens_seen": 2080964608
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1146429,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.32552433013916,
      "objective/train/theoretical_loss": 3.419093990508173,
      "objective/train/tokens_used": 439889376,
      "theoretical_loss": 3.419093990508173,
      "tokens_seen": 2081030144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757777777777778,
      "loss": 2.6839,
      "theoretical_loss": 3.419093990508173,
      "tokens_seen": 2081030144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037575757575757573,
      "loss": 2.6432,
      "theoretical_loss": 3.4190850829719563,
      "tokens_seen": 2081095680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037573737373737374,
      "loss": 2.557,
      "theoretical_loss": 3.4190761757947836,
      "tokens_seen": 2081161216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037571717171717175,
      "loss": 2.5485,
      "theoretical_loss": 3.4190672689766277,
      "tokens_seen": 2081226752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756969696969697,
      "loss": 2.7005,
      "theoretical_loss": 3.4190583625174633,
      "tokens_seen": 2081292288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037567676767676767,
      "loss": 2.6643,
      "theoretical_loss": 3.419049456417265,
      "tokens_seen": 2081357824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756565656565656,
      "loss": 2.5289,
      "theoretical_loss": 3.419040550676007,
      "tokens_seen": 2081423360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756363636363637,
      "loss": 2.8114,
      "theoretical_loss": 3.419031645293663,
      "tokens_seen": 2081488896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037561616161616165,
      "loss": 2.7906,
      "theoretical_loss": 3.4190227402702074,
      "tokens_seen": 2081554432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755959595959596,
      "loss": 2.9096,
      "theoretical_loss": 3.419013835605615,
      "tokens_seen": 2081619968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037557575757575756,
      "loss": 2.5331,
      "theoretical_loss": 3.4190049312998596,
      "tokens_seen": 2081685504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755555555555555,
      "loss": 2.7058,
      "theoretical_loss": 3.4189960273529154,
      "tokens_seen": 2081751040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755353535353536,
      "loss": 2.726,
      "theoretical_loss": 3.4189871237647567,
      "tokens_seen": 2081816576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037551515151515154,
      "loss": 2.562,
      "theoretical_loss": 3.4189782205353576,
      "tokens_seen": 2081882112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754949494949495,
      "loss": 2.5914,
      "theoretical_loss": 3.4189693176646925,
      "tokens_seen": 2081947648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037547474747474745,
      "loss": 2.7903,
      "theoretical_loss": 3.4189604151527364,
      "tokens_seen": 2082013184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037545454545454546,
      "loss": 2.6528,
      "theoretical_loss": 3.4189515129994623,
      "tokens_seen": 2082078720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754343434343435,
      "loss": 2.6828,
      "theoretical_loss": 3.418942611204845,
      "tokens_seen": 2082144256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037541414141414143,
      "loss": 2.5347,
      "theoretical_loss": 3.418933709768859,
      "tokens_seen": 2082209792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753939393939394,
      "loss": 2.6722,
      "theoretical_loss": 3.418924808691478,
      "tokens_seen": 2082275328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037537373737373735,
      "loss": 2.4734,
      "theoretical_loss": 3.4189159079726767,
      "tokens_seen": 2082340864
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037535353535353536,
      "loss": 2.6157,
      "theoretical_loss": 3.4189070076124293,
      "tokens_seen": 2082406400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037533333333333337,
      "loss": 2.7297,
      "theoretical_loss": 3.41889810761071,
      "tokens_seen": 2082471936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753131313131313,
      "loss": 2.6283,
      "theoretical_loss": 3.418889207967493,
      "tokens_seen": 2082537472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752929292929293,
      "loss": 2.3652,
      "theoretical_loss": 3.418880308682753,
      "tokens_seen": 2082603008
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1147399,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0194616317749023,
      "objective/train/theoretical_loss": 3.4188714097564636,
      "objective/train/tokens_used": 441527776,
      "theoretical_loss": 3.4188714097564636,
      "tokens_seen": 2082668544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752727272727273,
      "loss": 2.8635,
      "theoretical_loss": 3.4188714097564636,
      "tokens_seen": 2082668544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037525252525252525,
      "loss": 2.7142,
      "theoretical_loss": 3.4188625111886,
      "tokens_seen": 2082734080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037523232323232326,
      "loss": 2.5603,
      "theoretical_loss": 3.4188536129791354,
      "tokens_seen": 2082799616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752121212121212,
      "loss": 2.668,
      "theoretical_loss": 3.418844715128045,
      "tokens_seen": 2082865152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751919191919192,
      "loss": 2.9046,
      "theoretical_loss": 3.418835817635303,
      "tokens_seen": 2082930688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751717171717172,
      "loss": 2.5116,
      "theoretical_loss": 3.4188269205008828,
      "tokens_seen": 2082996224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037515151515151514,
      "loss": 2.5802,
      "theoretical_loss": 3.41881802372476,
      "tokens_seen": 2083061760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037513131313131315,
      "loss": 2.6405,
      "theoretical_loss": 3.4188091273069072,
      "tokens_seen": 2083127296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751111111111111,
      "loss": 2.613,
      "theoretical_loss": 3.4188002312473005,
      "tokens_seen": 2083192832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750909090909091,
      "loss": 2.6961,
      "theoretical_loss": 3.4187913355459134,
      "tokens_seen": 2083258368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750707070707071,
      "loss": 2.5561,
      "theoretical_loss": 3.41878244020272,
      "tokens_seen": 2083323904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037505050505050503,
      "loss": 2.6632,
      "theoretical_loss": 3.418773545217695,
      "tokens_seen": 2083389440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037503030303030305,
      "loss": 3.0064,
      "theoretical_loss": 3.418764650590812,
      "tokens_seen": 2083454976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375010101010101,
      "loss": 2.7303,
      "theoretical_loss": 3.4187557563220463,
      "tokens_seen": 2083520512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374989898989899,
      "loss": 2.7951,
      "theoretical_loss": 3.4187468624113717,
      "tokens_seen": 2083586048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037496969696969697,
      "loss": 2.4562,
      "theoretical_loss": 3.4187379688587627,
      "tokens_seen": 2083651584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374949494949495,
      "loss": 2.6969,
      "theoretical_loss": 3.4187290756641935,
      "tokens_seen": 2083717120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037492929292929294,
      "loss": 2.3717,
      "theoretical_loss": 3.4187201828276383,
      "tokens_seen": 2083782656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749090909090909,
      "loss": 2.6182,
      "theoretical_loss": 3.418711290349072,
      "tokens_seen": 2083848192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003748888888888889,
      "loss": 2.7406,
      "theoretical_loss": 3.418702398228468,
      "tokens_seen": 2083913728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037486868686868686,
      "loss": 2.6609,
      "theoretical_loss": 3.418693506465801,
      "tokens_seen": 2083979264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003748484848484849,
      "loss": 2.6927,
      "theoretical_loss": 3.4186846150610464,
      "tokens_seen": 2084044800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037482828282828283,
      "loss": 2.9106,
      "theoretical_loss": 3.418675724014177,
      "tokens_seen": 2084110336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037480808080808084,
      "loss": 2.7649,
      "theoretical_loss": 3.418666833325168,
      "tokens_seen": 2084175872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747878787878788,
      "loss": 2.6435,
      "theoretical_loss": 3.4186579429939927,
      "tokens_seen": 2084241408
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1148024,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3189051151275635,
      "objective/train/theoretical_loss": 3.4186490530206273,
      "objective/train/tokens_used": 443166176,
      "theoretical_loss": 3.4186490530206273,
      "tokens_seen": 2084306944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037476767676767676,
      "loss": 2.4753,
      "theoretical_loss": 3.4186490530206273,
      "tokens_seen": 2084306944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037474747474747477,
      "loss": 2.7101,
      "theoretical_loss": 3.4186401634050445,
      "tokens_seen": 2084372480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747272727272727,
      "loss": 2.6549,
      "theoretical_loss": 3.4186312741472196,
      "tokens_seen": 2084438016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037470707070707073,
      "loss": 2.6999,
      "theoretical_loss": 3.4186223852471262,
      "tokens_seen": 2084503552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003746868686868687,
      "loss": 2.6416,
      "theoretical_loss": 3.4186134967047392,
      "tokens_seen": 2084569088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037466666666666665,
      "loss": 2.6206,
      "theoretical_loss": 3.4186046085200332,
      "tokens_seen": 2084634624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037464646464646466,
      "loss": 2.6309,
      "theoretical_loss": 3.418595720692982,
      "tokens_seen": 2084700160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037462626262626267,
      "loss": 2.7098,
      "theoretical_loss": 3.41858683322356,
      "tokens_seen": 2084765696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037460606060606063,
      "loss": 2.3725,
      "theoretical_loss": 3.418577946111742,
      "tokens_seen": 2084831232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745858585858586,
      "loss": 2.7087,
      "theoretical_loss": 3.4185690593575018,
      "tokens_seen": 2084896768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037456565656565654,
      "loss": 2.7147,
      "theoretical_loss": 3.418560172960814,
      "tokens_seen": 2084962304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037454545454545455,
      "loss": 2.5222,
      "theoretical_loss": 3.418551286921653,
      "tokens_seen": 2085027840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037452525252525256,
      "loss": 2.4876,
      "theoretical_loss": 3.4185424012399936,
      "tokens_seen": 2085093376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745050505050505,
      "loss": 2.6135,
      "theoretical_loss": 3.4185335159158097,
      "tokens_seen": 2085158912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744848484848485,
      "loss": 2.5551,
      "theoretical_loss": 3.4185246309490758,
      "tokens_seen": 2085224448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037446464646464643,
      "loss": 2.5469,
      "theoretical_loss": 3.418515746339766,
      "tokens_seen": 2085289984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744444444444445,
      "loss": 2.536,
      "theoretical_loss": 3.418506862087855,
      "tokens_seen": 2085355520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037442424242424246,
      "loss": 2.4808,
      "theoretical_loss": 3.4184979781933174,
      "tokens_seen": 2085421056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744040404040404,
      "loss": 2.5116,
      "theoretical_loss": 3.418489094656127,
      "tokens_seen": 2085486592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037438383838383837,
      "loss": 2.597,
      "theoretical_loss": 3.418480211476259,
      "tokens_seen": 2085552128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743636363636363,
      "loss": 2.7988,
      "theoretical_loss": 3.4184713286536867,
      "tokens_seen": 2085617664
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743434343434344,
      "loss": 2.6326,
      "theoretical_loss": 3.4184624461883857,
      "tokens_seen": 2085683200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037432323232323235,
      "loss": 2.5852,
      "theoretical_loss": 3.41845356408033,
      "tokens_seen": 2085748736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743030303030303,
      "loss": 2.6487,
      "theoretical_loss": 3.4184446823294934,
      "tokens_seen": 2085814272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037428282828282826,
      "loss": 2.6216,
      "theoretical_loss": 3.418435800935851,
      "tokens_seen": 2085879808
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1149004,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6764867305755615,
      "objective/train/theoretical_loss": 3.4184269198993773,
      "objective/train/tokens_used": 444804576,
      "theoretical_loss": 3.4184269198993773,
      "tokens_seen": 2085945344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742626262626263,
      "loss": 2.7363,
      "theoretical_loss": 3.4184269198993773,
      "tokens_seen": 2085945344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742424242424243,
      "loss": 2.4787,
      "theoretical_loss": 3.418418039220046,
      "tokens_seen": 2086010880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037422222222222224,
      "loss": 2.5255,
      "theoretical_loss": 3.4184091588978314,
      "tokens_seen": 2086076416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742020202020202,
      "loss": 2.8832,
      "theoretical_loss": 3.418400278932709,
      "tokens_seen": 2086141952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037418181818181815,
      "loss": 2.6207,
      "theoretical_loss": 3.418391399324653,
      "tokens_seen": 2086207488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037416161616161617,
      "loss": 2.5225,
      "theoretical_loss": 3.418382520073637,
      "tokens_seen": 2086273024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741414141414142,
      "loss": 2.5866,
      "theoretical_loss": 3.4183736411796364,
      "tokens_seen": 2086338560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037412121212121213,
      "loss": 2.7921,
      "theoretical_loss": 3.4183647626426246,
      "tokens_seen": 2086404096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741010101010101,
      "loss": 2.6267,
      "theoretical_loss": 3.4183558844625765,
      "tokens_seen": 2086469632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740808080808081,
      "loss": 2.736,
      "theoretical_loss": 3.418347006639467,
      "tokens_seen": 2086535168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037406060606060606,
      "loss": 2.5893,
      "theoretical_loss": 3.41833812917327,
      "tokens_seen": 2086600704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037404040404040407,
      "loss": 2.4624,
      "theoretical_loss": 3.4183292520639603,
      "tokens_seen": 2086666240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374020202020202,
      "loss": 2.7353,
      "theoretical_loss": 3.418320375311512,
      "tokens_seen": 2086731776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374,
      "loss": 2.7195,
      "theoretical_loss": 3.4183114989159,
      "tokens_seen": 2086797312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373979797979798,
      "loss": 2.7483,
      "theoretical_loss": 3.4183026228770976,
      "tokens_seen": 2086862848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037395959595959595,
      "loss": 2.7238,
      "theoretical_loss": 3.4182937471950807,
      "tokens_seen": 2086928384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037393939393939396,
      "loss": 2.6748,
      "theoretical_loss": 3.4182848718698233,
      "tokens_seen": 2086993920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739191919191919,
      "loss": 2.4875,
      "theoretical_loss": 3.4182759969012992,
      "tokens_seen": 2087059456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037389898989898993,
      "loss": 2.6688,
      "theoretical_loss": 3.4182671222894836,
      "tokens_seen": 2087124992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738787878787879,
      "loss": 2.831,
      "theoretical_loss": 3.418258248034351,
      "tokens_seen": 2087190528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037385858585858584,
      "loss": 2.4438,
      "theoretical_loss": 3.4182493741358755,
      "tokens_seen": 2087256064
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037383838383838386,
      "loss": 2.6071,
      "theoretical_loss": 3.418240500594031,
      "tokens_seen": 2087321600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738181818181818,
      "loss": 2.7333,
      "theoretical_loss": 3.4182316274087934,
      "tokens_seen": 2087387136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737979797979798,
      "loss": 2.7107,
      "theoretical_loss": 3.4182227545801362,
      "tokens_seen": 2087452672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737777777777778,
      "loss": 2.491,
      "theoretical_loss": 3.418213882108034,
      "tokens_seen": 2087518208
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1149506,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7207868099212646,
      "objective/train/theoretical_loss": 3.4182050099924615,
      "objective/train/tokens_used": 446442976,
      "theoretical_loss": 3.4182050099924615,
      "tokens_seen": 2087583744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037375757575757574,
      "loss": 2.615,
      "theoretical_loss": 3.4182050099924615,
      "tokens_seen": 2087583744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037373737373737375,
      "loss": 2.6447,
      "theoretical_loss": 3.4181961382333927,
      "tokens_seen": 2087649280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037371717171717176,
      "loss": 2.7587,
      "theoretical_loss": 3.4181872668308024,
      "tokens_seen": 2087714816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003736969696969697,
      "loss": 2.8354,
      "theoretical_loss": 3.4181783957846656,
      "tokens_seen": 2087780352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037367676767676767,
      "loss": 2.4997,
      "theoretical_loss": 3.418169525094956,
      "tokens_seen": 2087845888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037365656565656563,
      "loss": 2.6017,
      "theoretical_loss": 3.4181606547616488,
      "tokens_seen": 2087911424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037363636363636364,
      "loss": 2.7573,
      "theoretical_loss": 3.4181517847847176,
      "tokens_seen": 2087976960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037361616161616165,
      "loss": 2.5815,
      "theoretical_loss": 3.418142915164138,
      "tokens_seen": 2088042496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735959595959596,
      "loss": 2.8231,
      "theoretical_loss": 3.4181340458998832,
      "tokens_seen": 2088108032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037357575757575757,
      "loss": 2.6708,
      "theoretical_loss": 3.4181251769919285,
      "tokens_seen": 2088173568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735555555555555,
      "loss": 2.6358,
      "theoretical_loss": 3.418116308440248,
      "tokens_seen": 2088239104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735353535353536,
      "loss": 2.6792,
      "theoretical_loss": 3.4181074402448175,
      "tokens_seen": 2088304640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037351515151515154,
      "loss": 2.8498,
      "theoretical_loss": 3.4180985724056097,
      "tokens_seen": 2088370176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734949494949495,
      "loss": 2.8895,
      "theoretical_loss": 3.4180897049226004,
      "tokens_seen": 2088435712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037347474747474746,
      "loss": 2.4677,
      "theoretical_loss": 3.418080837795763,
      "tokens_seen": 2088501248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734545454545454,
      "loss": 2.5268,
      "theoretical_loss": 3.418071971025073,
      "tokens_seen": 2088566784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734343434343435,
      "loss": 2.797,
      "theoretical_loss": 3.4180631046105048,
      "tokens_seen": 2088632320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037341414141414144,
      "loss": 2.8185,
      "theoretical_loss": 3.4180542385520325,
      "tokens_seen": 2088697856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733939393939394,
      "loss": 2.8718,
      "theoretical_loss": 3.4180453728496305,
      "tokens_seen": 2088763392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037337373737373735,
      "loss": 2.769,
      "theoretical_loss": 3.4180365075032744,
      "tokens_seen": 2088828928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733535353535353,
      "loss": 2.4134,
      "theoretical_loss": 3.4180276425129374,
      "tokens_seen": 2088894464
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733333333333334,
      "loss": 2.7009,
      "theoretical_loss": 3.4180187778785953,
      "tokens_seen": 2088960000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037331313131313133,
      "loss": 2.6487,
      "theoretical_loss": 3.418009913600221,
      "tokens_seen": 2089025536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732929292929293,
      "loss": 2.809,
      "theoretical_loss": 3.4180010496777906,
      "tokens_seen": 2089091072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037327272727272724,
      "loss": 2.6093,
      "theoretical_loss": 3.417992186111278,
      "tokens_seen": 2089156608
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1150478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.582503080368042,
      "objective/train/theoretical_loss": 3.4179833229006578,
      "objective/train/tokens_used": 448081376,
      "theoretical_loss": 3.4179833229006578,
      "tokens_seen": 2089222144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732525252525253,
      "loss": 2.5808,
      "theoretical_loss": 3.4179833229006578,
      "tokens_seen": 2089222144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037323232323232327,
      "loss": 2.5832,
      "theoretical_loss": 3.417974460045904,
      "tokens_seen": 2089287680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732121212121212,
      "loss": 2.8097,
      "theoretical_loss": 3.417965597546992,
      "tokens_seen": 2089353216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731919191919192,
      "loss": 2.8212,
      "theoretical_loss": 3.417956735403896,
      "tokens_seen": 2089418752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037317171717171714,
      "loss": 2.7588,
      "theoretical_loss": 3.4179478736165914,
      "tokens_seen": 2089484288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731515151515152,
      "loss": 2.6283,
      "theoretical_loss": 3.417939012185051,
      "tokens_seen": 2089549824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037313131313131316,
      "loss": 2.7115,
      "theoretical_loss": 3.4179301511092506,
      "tokens_seen": 2089615360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731111111111111,
      "loss": 2.4287,
      "theoretical_loss": 3.4179212903891645,
      "tokens_seen": 2089680896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037309090909090907,
      "loss": 2.5453,
      "theoretical_loss": 3.417912430024767,
      "tokens_seen": 2089746432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730707070707071,
      "loss": 2.5528,
      "theoretical_loss": 3.4179035700160334,
      "tokens_seen": 2089811968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730505050505051,
      "loss": 2.6557,
      "theoretical_loss": 3.417894710362937,
      "tokens_seen": 2089877504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037303030303030305,
      "loss": 2.5877,
      "theoretical_loss": 3.4178858510654537,
      "tokens_seen": 2089943040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373010101010101,
      "loss": 2.4979,
      "theoretical_loss": 3.4178769921235572,
      "tokens_seen": 2090008576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037298989898989896,
      "loss": 2.6663,
      "theoretical_loss": 3.4178681335372225,
      "tokens_seen": 2090074112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372969696969697,
      "loss": 2.735,
      "theoretical_loss": 3.4178592753064243,
      "tokens_seen": 2090139648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372949494949495,
      "loss": 2.4537,
      "theoretical_loss": 3.417850417431137,
      "tokens_seen": 2090205184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037292929292929294,
      "loss": 2.5839,
      "theoretical_loss": 3.417841559911335,
      "tokens_seen": 2090270720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729090909090909,
      "loss": 2.6842,
      "theoretical_loss": 3.417832702746993,
      "tokens_seen": 2090336256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728888888888889,
      "loss": 2.6375,
      "theoretical_loss": 3.4178238459380856,
      "tokens_seen": 2090401792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037286868686868687,
      "loss": 2.6688,
      "theoretical_loss": 3.4178149894845875,
      "tokens_seen": 2090467328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728484848484849,
      "loss": 2.6437,
      "theoretical_loss": 3.417806133386473,
      "tokens_seen": 2090532864
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037282828282828284,
      "loss": 2.6284,
      "theoretical_loss": 3.417797277643717,
      "tokens_seen": 2090598400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728080808080808,
      "loss": 2.5581,
      "theoretical_loss": 3.417788422256294,
      "tokens_seen": 2090663936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727878787878788,
      "loss": 2.6807,
      "theoretical_loss": 3.417779567224179,
      "tokens_seen": 2090729472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037276767676767676,
      "loss": 2.5903,
      "theoretical_loss": 3.417770712547346,
      "tokens_seen": 2090795008
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1151517,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.077552556991577,
      "objective/train/theoretical_loss": 3.41776185822577,
      "objective/train/tokens_used": 449719776,
      "theoretical_loss": 3.41776185822577,
      "tokens_seen": 2090860544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037274747474747477,
      "loss": 2.8183,
      "theoretical_loss": 3.41776185822577,
      "tokens_seen": 2090860544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037272727272727273,
      "loss": 2.757,
      "theoretical_loss": 3.417753004259425,
      "tokens_seen": 2090926080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037270707070707074,
      "loss": 2.7178,
      "theoretical_loss": 3.4177441506482866,
      "tokens_seen": 2090991616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726868686868687,
      "loss": 2.8016,
      "theoretical_loss": 3.417735297392328,
      "tokens_seen": 2091057152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037266666666666665,
      "loss": 2.603,
      "theoretical_loss": 3.417726444491526,
      "tokens_seen": 2091122688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037264646464646466,
      "loss": 2.6375,
      "theoretical_loss": 3.417717591945853,
      "tokens_seen": 2091188224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726262626262626,
      "loss": 2.7086,
      "theoretical_loss": 3.4177087397552848,
      "tokens_seen": 2091253760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037260606060606063,
      "loss": 2.4693,
      "theoretical_loss": 3.4176998879197957,
      "tokens_seen": 2091319296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725858585858586,
      "loss": 2.766,
      "theoretical_loss": 3.4176910364393605,
      "tokens_seen": 2091384832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037256565656565655,
      "loss": 2.6232,
      "theoretical_loss": 3.4176821853139536,
      "tokens_seen": 2091450368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037254545454545456,
      "loss": 2.5957,
      "theoretical_loss": 3.41767333454355,
      "tokens_seen": 2091515904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037252525252525257,
      "loss": 2.6096,
      "theoretical_loss": 3.417664484128124,
      "tokens_seen": 2091581440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725050505050505,
      "loss": 2.8325,
      "theoretical_loss": 3.4176556340676507,
      "tokens_seen": 2091646976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724848484848485,
      "loss": 2.8292,
      "theoretical_loss": 3.4176467843621037,
      "tokens_seen": 2091712512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037246464646464644,
      "loss": 2.7331,
      "theoretical_loss": 3.417637935011459,
      "tokens_seen": 2091778048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037244444444444445,
      "loss": 2.6483,
      "theoretical_loss": 3.4176290860156904,
      "tokens_seen": 2091843584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037242424242424246,
      "loss": 2.7649,
      "theoretical_loss": 3.4176202373747726,
      "tokens_seen": 2091909120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724040404040404,
      "loss": 2.4776,
      "theoretical_loss": 3.4176113890886803,
      "tokens_seen": 2091974656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723838383838384,
      "loss": 2.5944,
      "theoretical_loss": 3.4176025411573887,
      "tokens_seen": 2092040192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037236363636363633,
      "loss": 2.5043,
      "theoretical_loss": 3.417593693580872,
      "tokens_seen": 2092105728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723434343434344,
      "loss": 2.4376,
      "theoretical_loss": 3.4175848463591043,
      "tokens_seen": 2092171264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037232323232323235,
      "loss": 2.9385,
      "theoretical_loss": 3.4175759994920614,
      "tokens_seen": 2092236800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723030303030303,
      "loss": 2.5103,
      "theoretical_loss": 3.417567152979717,
      "tokens_seen": 2092302336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037228282828282827,
      "loss": 2.6269,
      "theoretical_loss": 3.4175583068220465,
      "tokens_seen": 2092367872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722626262626262,
      "loss": 2.7398,
      "theoretical_loss": 3.417549461019024,
      "tokens_seen": 2092433408
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1152178,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0097758769989014,
      "objective/train/theoretical_loss": 3.4175406155706245,
      "objective/train/tokens_used": 451358176,
      "theoretical_loss": 3.4175406155706245,
      "tokens_seen": 2092498944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722424242424243,
      "loss": 2.4834,
      "theoretical_loss": 3.4175406155706245,
      "tokens_seen": 2092498944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037222222222222225,
      "loss": 2.4869,
      "theoretical_loss": 3.417531770476823,
      "tokens_seen": 2092564480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722020202020202,
      "loss": 2.7087,
      "theoretical_loss": 3.4175229257375936,
      "tokens_seen": 2092630016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037218181818181816,
      "loss": 2.6086,
      "theoretical_loss": 3.417514081352911,
      "tokens_seen": 2092695552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037216161616161617,
      "loss": 2.666,
      "theoretical_loss": 3.41750523732275,
      "tokens_seen": 2092761088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721414141414142,
      "loss": 2.7373,
      "theoretical_loss": 3.4174963936470855,
      "tokens_seen": 2092826624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037212121212121214,
      "loss": 2.7108,
      "theoretical_loss": 3.417487550325892,
      "tokens_seen": 2092892160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721010101010101,
      "loss": 2.6518,
      "theoretical_loss": 3.417478707359144,
      "tokens_seen": 2092957696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037208080808080805,
      "loss": 2.5699,
      "theoretical_loss": 3.417469864746817,
      "tokens_seen": 2093023232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037206060606060606,
      "loss": 2.6461,
      "theoretical_loss": 3.417461022488885,
      "tokens_seen": 2093088768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720404040404041,
      "loss": 2.4359,
      "theoretical_loss": 3.417452180585322,
      "tokens_seen": 2093154304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037202020202020203,
      "loss": 2.5586,
      "theoretical_loss": 3.4174433390361045,
      "tokens_seen": 2093219840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372,
      "loss": 2.4325,
      "theoretical_loss": 3.4174344978412057,
      "tokens_seen": 2093285376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371979797979798,
      "loss": 2.569,
      "theoretical_loss": 3.417425657000601,
      "tokens_seen": 2093350912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037195959595959596,
      "loss": 2.6863,
      "theoretical_loss": 3.417416816514265,
      "tokens_seen": 2093416448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037193939393939397,
      "loss": 2.3295,
      "theoretical_loss": 3.4174079763821723,
      "tokens_seen": 2093481984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003719191919191919,
      "loss": 2.5572,
      "theoretical_loss": 3.4173991366042973,
      "tokens_seen": 2093547520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718989898989899,
      "loss": 2.4566,
      "theoretical_loss": 3.417390297180616,
      "tokens_seen": 2093613056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718787878787879,
      "loss": 2.4297,
      "theoretical_loss": 3.4173814581111017,
      "tokens_seen": 2093678592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037185858585858585,
      "loss": 2.8283,
      "theoretical_loss": 3.4173726193957297,
      "tokens_seen": 2093744128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037183838383838386,
      "loss": 2.6182,
      "theoretical_loss": 3.4173637810344744,
      "tokens_seen": 2093809664
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718181818181818,
      "loss": 2.4856,
      "theoretical_loss": 3.417354943027311,
      "tokens_seen": 2093875200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037179797979797983,
      "loss": 2.6159,
      "theoretical_loss": 3.417346105374214,
      "tokens_seen": 2093940736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717777777777778,
      "loss": 2.5668,
      "theoretical_loss": 3.4173372680751584,
      "tokens_seen": 2094006272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037175757575757574,
      "loss": 2.6343,
      "theoretical_loss": 3.4173284311301186,
      "tokens_seen": 2094071808
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1153480,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.500739574432373,
      "objective/train/theoretical_loss": 3.417319594539069,
      "objective/train/tokens_used": 452996576,
      "theoretical_loss": 3.417319594539069,
      "tokens_seen": 2094137344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037173737373737375,
      "loss": 2.6112,
      "theoretical_loss": 3.417319594539069,
      "tokens_seen": 2094137344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717171717171717,
      "loss": 2.6866,
      "theoretical_loss": 3.4173107583019853,
      "tokens_seen": 2094202880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716969696969697,
      "loss": 2.6229,
      "theoretical_loss": 3.417301922418842,
      "tokens_seen": 2094268416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716767676767677,
      "loss": 2.606,
      "theoretical_loss": 3.4172930868896128,
      "tokens_seen": 2094333952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037165656565656564,
      "loss": 2.5647,
      "theoretical_loss": 3.417284251714274,
      "tokens_seen": 2094399488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037163636363636365,
      "loss": 2.6784,
      "theoretical_loss": 3.4172754168927986,
      "tokens_seen": 2094465024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716161616161616,
      "loss": 2.5724,
      "theoretical_loss": 3.417266582425163,
      "tokens_seen": 2094530560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715959595959596,
      "loss": 2.3217,
      "theoretical_loss": 3.417257748311341,
      "tokens_seen": 2094596096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037157575757575757,
      "loss": 2.4304,
      "theoretical_loss": 3.4172489145513083,
      "tokens_seen": 2094661632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037155555555555553,
      "loss": 2.2137,
      "theoretical_loss": 3.4172400811450387,
      "tokens_seen": 2094727168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037153535353535354,
      "loss": 2.6197,
      "theoretical_loss": 3.4172312480925067,
      "tokens_seen": 2094792704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037151515151515155,
      "loss": 2.603,
      "theoretical_loss": 3.417222415393688,
      "tokens_seen": 2094858240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714949494949495,
      "loss": 2.5942,
      "theoretical_loss": 3.4172135830485573,
      "tokens_seen": 2094923776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037147474747474746,
      "loss": 2.5866,
      "theoretical_loss": 3.4172047510570893,
      "tokens_seen": 2094989312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714545454545455,
      "loss": 2.4817,
      "theoretical_loss": 3.417195919419258,
      "tokens_seen": 2095054848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037143434343434343,
      "loss": 2.6287,
      "theoretical_loss": 3.417187088135039,
      "tokens_seen": 2095120384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037141414141414144,
      "loss": 2.3498,
      "theoretical_loss": 3.417178257204407,
      "tokens_seen": 2095185920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713939393939394,
      "loss": 2.454,
      "theoretical_loss": 3.4171694266273365,
      "tokens_seen": 2095251456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037137373737373736,
      "loss": 2.5625,
      "theoretical_loss": 3.4171605964038023,
      "tokens_seen": 2095316992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037135353535353537,
      "loss": 2.6528,
      "theoretical_loss": 3.417151766533779,
      "tokens_seen": 2095382528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713333333333334,
      "loss": 2.5552,
      "theoretical_loss": 3.4171429370172426,
      "tokens_seen": 2095448064
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037131313131313134,
      "loss": 2.4329,
      "theoretical_loss": 3.4171341078541664,
      "tokens_seen": 2095513600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712929292929293,
      "loss": 2.6025,
      "theoretical_loss": 3.4171252790445257,
      "tokens_seen": 2095579136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037127272727272725,
      "loss": 2.6012,
      "theoretical_loss": 3.4171164505882956,
      "tokens_seen": 2095644672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037125252525252526,
      "loss": 2.6446,
      "theoretical_loss": 3.4171076224854504,
      "tokens_seen": 2095710208
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 1154065,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8171300888061523,
      "objective/train/theoretical_loss": 3.4170987947359652,
      "objective/train/tokens_used": 454634976,
      "theoretical_loss": 3.4170987947359652,
      "tokens_seen": 2095775744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037123232323232327,
      "loss": 2.7618,
      "theoretical_loss": 3.4170987947359652,
      "tokens_seen": 2095775744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037121212121212123,
      "loss": 2.3248,
      "theoretical_loss": 3.4170899673398147,
      "tokens_seen": 2095841280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711919191919192,
      "loss": 2.4875,
      "theoretical_loss": 3.4170811402969745,
      "tokens_seen": 2095906816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037117171717171714,
      "loss": 2.416,
      "theoretical_loss": 3.417072313607418,
      "tokens_seen": 2095972352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711515151515152,
      "loss": 2.5119,
      "theoretical_loss": 3.417063487271121,
      "tokens_seen": 2096037888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037113131313131316,
      "loss": 2.7506,
      "theoretical_loss": 3.4170546612880575,
      "tokens_seen": 2096103424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711111111111111,
      "loss": 2.6149,
      "theoretical_loss": 3.417045835658204,
      "tokens_seen": 2096168960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710909090909091,
      "loss": 2.7236,
      "theoretical_loss": 3.4170370103815335,
      "tokens_seen": 2096234496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037107070707070703,
      "loss": 2.5956,
      "theoretical_loss": 3.417028185458021,
      "tokens_seen": 2096300032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710505050505051,
      "loss": 2.506,
      "theoretical_loss": 3.4170193608876427,
      "tokens_seen": 2096365568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037103030303030306,
      "loss": 2.578,
      "theoretical_loss": 3.4170105366703725,
      "tokens_seen": 2096431104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000371010101010101,
      "loss": 2.5963,
      "theoretical_loss": 3.4170017128061847,
      "tokens_seen": 2096496640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037098989898989897,
      "loss": 2.7297,
      "theoretical_loss": 3.416992889295055,
      "tokens_seen": 2096562176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370969696969697,
      "loss": 2.8348,
      "theoretical_loss": 3.4169840661369584,
      "tokens_seen": 2096627712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370949494949495,
      "loss": 2.5569,
      "theoretical_loss": 3.4169752433318688,
      "tokens_seen": 2096693248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037092929292929295,
      "loss": 2.6674,
      "theoretical_loss": 3.4169664208797617,
      "tokens_seen": 2096758784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709090909090909,
      "loss": 2.9471,
      "theoretical_loss": 3.416957598780612,
      "tokens_seen": 2096824320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037088888888888886,
      "loss": 2.4803,
      "theoretical_loss": 3.416948777034394,
      "tokens_seen": 2096889856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708686868686869,
      "loss": 2.9403,
      "theoretical_loss": 3.416939955641083,
      "tokens_seen": 2096955392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708484848484849,
      "loss": 2.6036,
      "theoretical_loss": 3.416931134600654,
      "tokens_seen": 2097020928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037082828282828284,
      "loss": 2.6213,
      "theoretical_loss": 3.4169223139130813,
      "tokens_seen": 2097086464
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708080808080808,
      "loss": 2.5021,
      "theoretical_loss": 3.41691349357834,
      "tokens_seen": 2097152000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707878787878788,
      "loss": 2.5866,
      "theoretical_loss": 3.416904673596406,
      "tokens_seen": 2097217536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037076767676767677,
      "loss": 2.7012,
      "theoretical_loss": 3.416895853967252,
      "tokens_seen": 2097283072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707474747474748,
      "loss": 2.6695,
      "theoretical_loss": 3.4168870346908546,
      "tokens_seen": 2097348608
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1155055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3634607791900635,
      "objective/train/theoretical_loss": 3.416878215767188,
      "objective/train/tokens_used": 456273376,
      "theoretical_loss": 3.416878215767188,
      "tokens_seen": 2097414144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037072727272727273,
      "loss": 2.6518,
      "theoretical_loss": 3.416878215767188,
      "tokens_seen": 2097414144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707070707070707,
      "loss": 2.7283,
      "theoretical_loss": 3.416869397196227,
      "tokens_seen": 2097479680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706868686868687,
      "loss": 2.6649,
      "theoretical_loss": 3.416860578977947,
      "tokens_seen": 2097545216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037066666666666666,
      "loss": 2.6906,
      "theoretical_loss": 3.416851761112323,
      "tokens_seen": 2097610752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037064646464646467,
      "loss": 2.6143,
      "theoretical_loss": 3.416842943599329,
      "tokens_seen": 2097676288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037062626262626263,
      "loss": 2.3629,
      "theoretical_loss": 3.41683412643894,
      "tokens_seen": 2097741824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037060606060606064,
      "loss": 2.8845,
      "theoretical_loss": 3.4168253096311316,
      "tokens_seen": 2097807360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003705858585858586,
      "loss": 2.7976,
      "theoretical_loss": 3.4168164931758778,
      "tokens_seen": 2097872896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037056565656565655,
      "loss": 2.7928,
      "theoretical_loss": 3.4168076770731544,
      "tokens_seen": 2097938432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037054545454545456,
      "loss": 2.6009,
      "theoretical_loss": 3.416798861322936,
      "tokens_seen": 2098003968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003705252525252525,
      "loss": 2.7349,
      "theoretical_loss": 3.416790045925197,
      "tokens_seen": 2098069504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037050505050505053,
      "loss": 2.6537,
      "theoretical_loss": 3.416781230879913,
      "tokens_seen": 2098135040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704848484848485,
      "loss": 2.5315,
      "theoretical_loss": 3.416772416187058,
      "tokens_seen": 2098200576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037046464646464644,
      "loss": 2.3986,
      "theoretical_loss": 3.416763601846608,
      "tokens_seen": 2098266112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037044444444444446,
      "loss": 2.4969,
      "theoretical_loss": 3.4167547878585376,
      "tokens_seen": 2098331648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037042424242424247,
      "loss": 2.6538,
      "theoretical_loss": 3.416745974222821,
      "tokens_seen": 2098397184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704040404040404,
      "loss": 2.5273,
      "theoretical_loss": 3.4167371609394337,
      "tokens_seen": 2098462720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703838383838384,
      "loss": 2.7185,
      "theoretical_loss": 3.41672834800835,
      "tokens_seen": 2098528256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037036363636363634,
      "loss": 2.7387,
      "theoretical_loss": 3.416719535429546,
      "tokens_seen": 2098593792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037034343434343435,
      "loss": 2.6902,
      "theoretical_loss": 3.416710723202996,
      "tokens_seen": 2098659328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037032323232323236,
      "loss": 2.6166,
      "theoretical_loss": 3.4167019113286745,
      "tokens_seen": 2098724864
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703030303030303,
      "loss": 2.2139,
      "theoretical_loss": 3.4166930998065568,
      "tokens_seen": 2098790400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702828282828283,
      "loss": 2.6439,
      "theoretical_loss": 3.4166842886366178,
      "tokens_seen": 2098855936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037026262626262623,
      "loss": 2.8034,
      "theoretical_loss": 3.416675477818832,
      "tokens_seen": 2098921472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702424242424243,
      "loss": 2.5019,
      "theoretical_loss": 3.4166666673531756,
      "tokens_seen": 2098987008
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1156556,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7254951000213623,
      "objective/train/theoretical_loss": 3.4166578572396222,
      "objective/train/tokens_used": 457911776,
      "theoretical_loss": 3.4166578572396222,
      "tokens_seen": 2099052544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037022222222222225,
      "loss": 2.4085,
      "theoretical_loss": 3.4166578572396222,
      "tokens_seen": 2099052544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702020202020202,
      "loss": 2.739,
      "theoretical_loss": 3.4166490474781477,
      "tokens_seen": 2099118080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037018181818181817,
      "loss": 2.5515,
      "theoretical_loss": 3.416640238068726,
      "tokens_seen": 2099183616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701616161616161,
      "loss": 2.6694,
      "theoretical_loss": 3.416631429011333,
      "tokens_seen": 2099249152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701414141414142,
      "loss": 2.6967,
      "theoretical_loss": 3.416622620305943,
      "tokens_seen": 2099314688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037012121212121215,
      "loss": 2.5425,
      "theoretical_loss": 3.4166138119525313,
      "tokens_seen": 2099380224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701010101010101,
      "loss": 2.8095,
      "theoretical_loss": 3.416605003951073,
      "tokens_seen": 2099445760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037008080808080806,
      "loss": 2.5167,
      "theoretical_loss": 3.4165961963015423,
      "tokens_seen": 2099511296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370060606060606,
      "loss": 2.6135,
      "theoretical_loss": 3.416587389003915,
      "tokens_seen": 2099576832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003700404040404041,
      "loss": 2.635,
      "theoretical_loss": 3.4165785820581656,
      "tokens_seen": 2099642368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037002020202020204,
      "loss": 2.6785,
      "theoretical_loss": 3.416569775464269,
      "tokens_seen": 2099707904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037,
      "loss": 2.7011,
      "theoretical_loss": 3.4165609692222008,
      "tokens_seen": 2099773440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036997979797979795,
      "loss": 2.7845,
      "theoretical_loss": 3.416552163331935,
      "tokens_seen": 2099838976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036995959595959596,
      "loss": 2.5003,
      "theoretical_loss": 3.416543357793447,
      "tokens_seen": 2099904512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369939393939394,
      "loss": 2.5218,
      "theoretical_loss": 3.4165345526067123,
      "tokens_seen": 2099970048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036991919191919193,
      "loss": 2.5563,
      "theoretical_loss": 3.416525747771705,
      "tokens_seen": 2100035584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698989898989899,
      "loss": 2.7174,
      "theoretical_loss": 3.416516943288401,
      "tokens_seen": 2100101120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036987878787878784,
      "loss": 2.7711,
      "theoretical_loss": 3.4165081391567744,
      "tokens_seen": 2100166656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036985858585858586,
      "loss": 2.6654,
      "theoretical_loss": 3.4164993353768005,
      "tokens_seen": 2100232192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036983838383838387,
      "loss": 2.6669,
      "theoretical_loss": 3.4164905319484546,
      "tokens_seen": 2100297728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698181818181818,
      "loss": 2.6784,
      "theoretical_loss": 3.4164817288717106,
      "tokens_seen": 2100363264
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697979797979798,
      "loss": 2.4752,
      "theoretical_loss": 3.416472926146545,
      "tokens_seen": 2100428800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697777777777778,
      "loss": 2.4664,
      "theoretical_loss": 3.416464123772932,
      "tokens_seen": 2100494336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697575757575758,
      "loss": 2.6676,
      "theoretical_loss": 3.4164553217508464,
      "tokens_seen": 2100559872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036973737373737376,
      "loss": 2.5018,
      "theoretical_loss": 3.416446520080264,
      "tokens_seen": 2100625408
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1157195,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8589794635772705,
      "objective/train/theoretical_loss": 3.4164377187611588,
      "objective/train/tokens_used": 459550176,
      "theoretical_loss": 3.4164377187611588,
      "tokens_seen": 2100690944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697171717171717,
      "loss": 2.7354,
      "theoretical_loss": 3.4164377187611588,
      "tokens_seen": 2100690944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036969696969696967,
      "loss": 2.545,
      "theoretical_loss": 3.4164289177935063,
      "tokens_seen": 2100756480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696767676767677,
      "loss": 2.6809,
      "theoretical_loss": 3.4164201171772817,
      "tokens_seen": 2100822016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696565656565657,
      "loss": 2.7996,
      "theoretical_loss": 3.4164113169124595,
      "tokens_seen": 2100887552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036963636363636365,
      "loss": 2.4825,
      "theoretical_loss": 3.416402516999015,
      "tokens_seen": 2100953088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696161616161616,
      "loss": 2.5685,
      "theoretical_loss": 3.4163937174369234,
      "tokens_seen": 2101018624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695959595959596,
      "loss": 2.7909,
      "theoretical_loss": 3.416384918226159,
      "tokens_seen": 2101084160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695757575757576,
      "loss": 2.41,
      "theoretical_loss": 3.4163761193666975,
      "tokens_seen": 2101149696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695555555555556,
      "loss": 2.5918,
      "theoretical_loss": 3.416367320858514,
      "tokens_seen": 2101215232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036953535353535354,
      "loss": 2.5184,
      "theoretical_loss": 3.416358522701583,
      "tokens_seen": 2101280768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695151515151515,
      "loss": 2.5719,
      "theoretical_loss": 3.4163497248958805,
      "tokens_seen": 2101346304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694949494949495,
      "loss": 2.7546,
      "theoretical_loss": 3.4163409274413796,
      "tokens_seen": 2101411840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036947474747474747,
      "loss": 2.5471,
      "theoretical_loss": 3.4163321303380574,
      "tokens_seen": 2101477376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694545454545455,
      "loss": 2.6367,
      "theoretical_loss": 3.416323333585888,
      "tokens_seen": 2101542912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036943434343434344,
      "loss": 2.7109,
      "theoretical_loss": 3.416314537184846,
      "tokens_seen": 2101608448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036941414141414145,
      "loss": 2.8343,
      "theoretical_loss": 3.416305741134907,
      "tokens_seen": 2101673984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693939393939394,
      "loss": 2.5901,
      "theoretical_loss": 3.4162969454360463,
      "tokens_seen": 2101739520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036937373737373736,
      "loss": 2.6116,
      "theoretical_loss": 3.4162881500882385,
      "tokens_seen": 2101805056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693535353535354,
      "loss": 2.5429,
      "theoretical_loss": 3.4162793550914587,
      "tokens_seen": 2101870592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036933333333333333,
      "loss": 2.7821,
      "theoretical_loss": 3.416270560445682,
      "tokens_seen": 2101936128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036931313131313134,
      "loss": 2.7114,
      "theoretical_loss": 3.4162617661508836,
      "tokens_seen": 2102001664
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692929292929293,
      "loss": 2.6573,
      "theoretical_loss": 3.4162529722070385,
      "tokens_seen": 2102067200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036927272727272725,
      "loss": 2.5542,
      "theoretical_loss": 3.416244178614121,
      "tokens_seen": 2102132736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036925252525252527,
      "loss": 2.4011,
      "theoretical_loss": 3.416235385372107,
      "tokens_seen": 2102198272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692323232323233,
      "loss": 2.4574,
      "theoretical_loss": 3.416226592480972,
      "tokens_seen": 2102263808
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1158634,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4343600273132324,
      "objective/train/theoretical_loss": 3.41621779994069,
      "objective/train/tokens_used": 461188576,
      "theoretical_loss": 3.41621779994069,
      "tokens_seen": 2102329344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036921212121212123,
      "loss": 2.7046,
      "theoretical_loss": 3.41621779994069,
      "tokens_seen": 2102329344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691919191919192,
      "loss": 2.8016,
      "theoretical_loss": 3.4162090077512364,
      "tokens_seen": 2102394880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036917171717171715,
      "loss": 2.6305,
      "theoretical_loss": 3.4162002159125864,
      "tokens_seen": 2102460416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036915151515151516,
      "loss": 2.6215,
      "theoretical_loss": 3.416191424424715,
      "tokens_seen": 2102525952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036913131313131317,
      "loss": 2.7357,
      "theoretical_loss": 3.4161826332875975,
      "tokens_seen": 2102591488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691111111111111,
      "loss": 2.4573,
      "theoretical_loss": 3.4161738425012085,
      "tokens_seen": 2102657024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690909090909091,
      "loss": 2.7079,
      "theoretical_loss": 3.4161650520655233,
      "tokens_seen": 2102722560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036907070707070704,
      "loss": 2.522,
      "theoretical_loss": 3.4161562619805173,
      "tokens_seen": 2102788096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690505050505051,
      "loss": 2.8353,
      "theoretical_loss": 3.4161474722461653,
      "tokens_seen": 2102853632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036903030303030306,
      "loss": 2.5757,
      "theoretical_loss": 3.4161386828624423,
      "tokens_seen": 2102919168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369010101010101,
      "loss": 2.6205,
      "theoretical_loss": 3.416129893829323,
      "tokens_seen": 2102984704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368989898989899,
      "loss": 2.631,
      "theoretical_loss": 3.4161211051467837,
      "tokens_seen": 2103050240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036896969696969693,
      "loss": 2.4927,
      "theoretical_loss": 3.4161123168147984,
      "tokens_seen": 2103115776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368949494949495,
      "loss": 2.7321,
      "theoretical_loss": 3.4161035288333426,
      "tokens_seen": 2103181312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036892929292929295,
      "loss": 2.5872,
      "theoretical_loss": 3.416094741202391,
      "tokens_seen": 2103246848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003689090909090909,
      "loss": 2.9086,
      "theoretical_loss": 3.41608595392192,
      "tokens_seen": 2103312384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036888888888888887,
      "loss": 2.5979,
      "theoretical_loss": 3.416077166991903,
      "tokens_seen": 2103377920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688686868686869,
      "loss": 2.522,
      "theoretical_loss": 3.416068380412316,
      "tokens_seen": 2103443456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688484848484849,
      "loss": 2.62,
      "theoretical_loss": 3.416059594183134,
      "tokens_seen": 2103508992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036882828282828285,
      "loss": 2.7018,
      "theoretical_loss": 3.416050808304332,
      "tokens_seen": 2103574528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688080808080808,
      "loss": 2.4874,
      "theoretical_loss": 3.416042022775885,
      "tokens_seen": 2103640064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036878787878787876,
      "loss": 2.7253,
      "theoretical_loss": 3.4160332375977687,
      "tokens_seen": 2103705600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036876767676767677,
      "loss": 2.6079,
      "theoretical_loss": 3.4160244527699577,
      "tokens_seen": 2103771136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003687474747474748,
      "loss": 2.808,
      "theoretical_loss": 3.4160156682924274,
      "tokens_seen": 2103836672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036872727272727274,
      "loss": 2.5926,
      "theoretical_loss": 3.416006884165152,
      "tokens_seen": 2103902208
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1159301,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5771374702453613,
      "objective/train/theoretical_loss": 3.4159981003881086,
      "objective/train/tokens_used": 462826976,
      "theoretical_loss": 3.4159981003881086,
      "tokens_seen": 2103967744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003687070707070707,
      "loss": 2.724,
      "theoretical_loss": 3.4159981003881086,
      "tokens_seen": 2103967744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686868686868687,
      "loss": 2.6223,
      "theoretical_loss": 3.4159893169612703,
      "tokens_seen": 2104033280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036866666666666667,
      "loss": 2.5754,
      "theoretical_loss": 3.415980533884613,
      "tokens_seen": 2104098816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686464646464647,
      "loss": 2.6732,
      "theoretical_loss": 3.4159717511581116,
      "tokens_seen": 2104164352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036862626262626263,
      "loss": 2.7595,
      "theoretical_loss": 3.4159629687817423,
      "tokens_seen": 2104229888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686060606060606,
      "loss": 2.4892,
      "theoretical_loss": 3.415954186755479,
      "tokens_seen": 2104295424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685858585858586,
      "loss": 2.6573,
      "theoretical_loss": 3.415945405079297,
      "tokens_seen": 2104360960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036856565656565656,
      "loss": 2.7806,
      "theoretical_loss": 3.4159366237531725,
      "tokens_seen": 2104426496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036854545454545457,
      "loss": 2.5188,
      "theoretical_loss": 3.415927842777079,
      "tokens_seen": 2104492032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685252525252525,
      "loss": 2.4961,
      "theoretical_loss": 3.4159190621509934,
      "tokens_seen": 2104557568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036850505050505054,
      "loss": 2.5838,
      "theoretical_loss": 3.4159102818748894,
      "tokens_seen": 2104623104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684848484848485,
      "loss": 2.5448,
      "theoretical_loss": 3.415901501948743,
      "tokens_seen": 2104688640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036846464646464645,
      "loss": 2.6647,
      "theoretical_loss": 3.415892722372529,
      "tokens_seen": 2104754176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036844444444444446,
      "loss": 2.6084,
      "theoretical_loss": 3.4158839431462225,
      "tokens_seen": 2104819712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684242424242424,
      "loss": 2.6192,
      "theoretical_loss": 3.415875164269799,
      "tokens_seen": 2104885248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036840404040404043,
      "loss": 2.6547,
      "theoretical_loss": 3.415866385743233,
      "tokens_seen": 2104950784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683838383838384,
      "loss": 2.4461,
      "theoretical_loss": 3.4158576075665006,
      "tokens_seen": 2105016320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036836363636363634,
      "loss": 2.4437,
      "theoretical_loss": 3.415848829739576,
      "tokens_seen": 2105081856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036834343434343435,
      "loss": 2.6591,
      "theoretical_loss": 3.4158400522624355,
      "tokens_seen": 2105147392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683232323232323,
      "loss": 2.7838,
      "theoretical_loss": 3.4158312751350532,
      "tokens_seen": 2105212928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683030303030303,
      "loss": 2.541,
      "theoretical_loss": 3.415822498357405,
      "tokens_seen": 2105278464
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003682828282828283,
      "loss": 2.6542,
      "theoretical_loss": 3.4158137219294655,
      "tokens_seen": 2105344000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036826262626262624,
      "loss": 2.4815,
      "theoretical_loss": 3.4158049458512103,
      "tokens_seen": 2105409536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036824242424242425,
      "loss": 2.7231,
      "theoretical_loss": 3.415796170122615,
      "tokens_seen": 2105475072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036822222222222226,
      "loss": 2.5703,
      "theoretical_loss": 3.4157873947436537,
      "tokens_seen": 2105540608
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1160284,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0382537841796875,
      "objective/train/theoretical_loss": 3.415778619714302,
      "objective/train/tokens_used": 464465376,
      "theoretical_loss": 3.415778619714302,
      "tokens_seen": 2105606144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003682020202020202,
      "loss": 2.6972,
      "theoretical_loss": 3.415778619714302,
      "tokens_seen": 2105606144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036818181818181817,
      "loss": 2.4719,
      "theoretical_loss": 3.4157698450345353,
      "tokens_seen": 2105671680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036816161616161613,
      "loss": 2.6513,
      "theoretical_loss": 3.4157610707043284,
      "tokens_seen": 2105737216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036814141414141414,
      "loss": 2.6362,
      "theoretical_loss": 3.4157522967236575,
      "tokens_seen": 2105802752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036812121212121215,
      "loss": 2.4002,
      "theoretical_loss": 3.415743523092497,
      "tokens_seen": 2105868288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681010101010101,
      "loss": 2.7257,
      "theoretical_loss": 3.415734749810822,
      "tokens_seen": 2105933824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036808080808080806,
      "loss": 2.4798,
      "theoretical_loss": 3.4157259768786075,
      "tokens_seen": 2105999360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680606060606061,
      "loss": 2.3792,
      "theoretical_loss": 3.41571720429583,
      "tokens_seen": 2106064896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680404040404041,
      "loss": 2.7184,
      "theoretical_loss": 3.415708432062463,
      "tokens_seen": 2106130432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036802020202020204,
      "loss": 2.6301,
      "theoretical_loss": 3.415699660178483,
      "tokens_seen": 2106195968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368,
      "loss": 2.629,
      "theoretical_loss": 3.415690888643865,
      "tokens_seen": 2106261504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036797979797979796,
      "loss": 2.8607,
      "theoretical_loss": 3.415682117458583,
      "tokens_seen": 2106327040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036795959595959597,
      "loss": 2.4124,
      "theoretical_loss": 3.415673346622614,
      "tokens_seen": 2106392576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367939393939394,
      "loss": 2.5606,
      "theoretical_loss": 3.4156645761359323,
      "tokens_seen": 2106458112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036791919191919194,
      "loss": 2.471,
      "theoretical_loss": 3.415655805998513,
      "tokens_seen": 2106523648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003678989898989899,
      "loss": 2.4754,
      "theoretical_loss": 3.4156470362103315,
      "tokens_seen": 2106589184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036787878787878785,
      "loss": 2.6735,
      "theoretical_loss": 3.415638266771363,
      "tokens_seen": 2106654720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003678585858585859,
      "loss": 2.4972,
      "theoretical_loss": 3.415629497681583,
      "tokens_seen": 2106720256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036783838383838387,
      "loss": 2.6408,
      "theoretical_loss": 3.415620728940967,
      "tokens_seen": 2106785792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036781818181818183,
      "loss": 2.4854,
      "theoretical_loss": 3.4156119605494895,
      "tokens_seen": 2106851328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677979797979798,
      "loss": 2.6254,
      "theoretical_loss": 3.4156031925071257,
      "tokens_seen": 2106916864
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036777777777777774,
      "loss": 2.3775,
      "theoretical_loss": 3.4155944248138512,
      "tokens_seen": 2106982400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677575757575758,
      "loss": 2.688,
      "theoretical_loss": 3.4155856574696415,
      "tokens_seen": 2107047936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036773737373737376,
      "loss": 2.6889,
      "theoretical_loss": 3.4155768904744717,
      "tokens_seen": 2107113472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677171717171717,
      "loss": 2.2961,
      "theoretical_loss": 3.415568123828317,
      "tokens_seen": 2107179008
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1161731,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1398634910583496,
      "objective/train/theoretical_loss": 3.415559357531152,
      "objective/train/tokens_used": 466103776,
      "theoretical_loss": 3.415559357531152,
      "tokens_seen": 2107244544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676969696969697,
      "loss": 2.433,
      "theoretical_loss": 3.415559357531152,
      "tokens_seen": 2107244544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676767676767677,
      "loss": 2.4995,
      "theoretical_loss": 3.4155505915829525,
      "tokens_seen": 2107310080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676565656565657,
      "loss": 2.6416,
      "theoretical_loss": 3.415541825983694,
      "tokens_seen": 2107375616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036763636363636366,
      "loss": 2.4995,
      "theoretical_loss": 3.4155330607333516,
      "tokens_seen": 2107441152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676161616161616,
      "loss": 2.7959,
      "theoretical_loss": 3.4155242958319003,
      "tokens_seen": 2107506688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036759595959595957,
      "loss": 2.4846,
      "theoretical_loss": 3.4155155312793157,
      "tokens_seen": 2107572224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675757575757576,
      "loss": 2.7947,
      "theoretical_loss": 3.415506767075573,
      "tokens_seen": 2107637760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675555555555556,
      "loss": 2.6484,
      "theoretical_loss": 3.4154980032206472,
      "tokens_seen": 2107703296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036753535353535355,
      "loss": 2.5693,
      "theoretical_loss": 3.415489239714514,
      "tokens_seen": 2107768832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675151515151515,
      "loss": 2.754,
      "theoretical_loss": 3.415480476557148,
      "tokens_seen": 2107834368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674949494949495,
      "loss": 2.6904,
      "theoretical_loss": 3.4154717137485253,
      "tokens_seen": 2107899904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674747474747475,
      "loss": 2.6066,
      "theoretical_loss": 3.4154629512886205,
      "tokens_seen": 2107965440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674545454545455,
      "loss": 2.6974,
      "theoretical_loss": 3.4154541891774093,
      "tokens_seen": 2108030976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036743434343434344,
      "loss": 2.5926,
      "theoretical_loss": 3.415445427414867,
      "tokens_seen": 2108096512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674141414141414,
      "loss": 2.6391,
      "theoretical_loss": 3.4154366660009687,
      "tokens_seen": 2108162048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673939393939394,
      "loss": 2.5497,
      "theoretical_loss": 3.4154279049356897,
      "tokens_seen": 2108227584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036737373737373737,
      "loss": 2.6752,
      "theoretical_loss": 3.4154191442190056,
      "tokens_seen": 2108293120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673535353535354,
      "loss": 2.644,
      "theoretical_loss": 3.4154103838508907,
      "tokens_seen": 2108358656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036733333333333334,
      "loss": 2.6448,
      "theoretical_loss": 3.415401623831322,
      "tokens_seen": 2108424192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036731313131313135,
      "loss": 2.5945,
      "theoretical_loss": 3.415392864160273,
      "tokens_seen": 2108489728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672929292929293,
      "loss": 2.6826,
      "theoretical_loss": 3.41538410483772,
      "tokens_seen": 2108555264
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036727272727272726,
      "loss": 2.9419,
      "theoretical_loss": 3.4153753458636382,
      "tokens_seen": 2108620800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036725252525252527,
      "loss": 2.4951,
      "theoretical_loss": 3.415366587238003,
      "tokens_seen": 2108686336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036723232323232323,
      "loss": 2.6084,
      "theoretical_loss": 3.415357828960789,
      "tokens_seen": 2108751872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036721212121212124,
      "loss": 2.5701,
      "theoretical_loss": 3.4153490710319723,
      "tokens_seen": 2108817408
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1162303,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.828242063522339,
      "objective/train/theoretical_loss": 3.4153403134515283,
      "objective/train/tokens_used": 467742176,
      "theoretical_loss": 3.4153403134515283,
      "tokens_seen": 2108882944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671919191919192,
      "loss": 2.588,
      "theoretical_loss": 3.4153403134515283,
      "tokens_seen": 2108882944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036717171717171715,
      "loss": 2.5231,
      "theoretical_loss": 3.4153315562194315,
      "tokens_seen": 2108948480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036715151515151516,
      "loss": 2.8862,
      "theoretical_loss": 3.415322799335658,
      "tokens_seen": 2109014016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671313131313132,
      "loss": 2.7052,
      "theoretical_loss": 3.4153140428001825,
      "tokens_seen": 2109079552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036711111111111113,
      "loss": 2.4027,
      "theoretical_loss": 3.4153052866129814,
      "tokens_seen": 2109145088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670909090909091,
      "loss": 2.4914,
      "theoretical_loss": 3.4152965307740284,
      "tokens_seen": 2109210624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036707070707070705,
      "loss": 2.8417,
      "theoretical_loss": 3.4152877752833,
      "tokens_seen": 2109276160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036705050505050506,
      "loss": 2.6103,
      "theoretical_loss": 3.4152790201407717,
      "tokens_seen": 2109341696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036703030303030307,
      "loss": 2.529,
      "theoretical_loss": 3.4152702653464178,
      "tokens_seen": 2109407232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367010101010101,
      "loss": 2.6543,
      "theoretical_loss": 3.4152615109002142,
      "tokens_seen": 2109472768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366989898989899,
      "loss": 2.958,
      "theoretical_loss": 3.4152527568021362,
      "tokens_seen": 2109538304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036696969696969694,
      "loss": 2.6509,
      "theoretical_loss": 3.41524400305216,
      "tokens_seen": 2109603840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366949494949495,
      "loss": 2.5489,
      "theoretical_loss": 3.4152352496502596,
      "tokens_seen": 2109669376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036692929292929296,
      "loss": 2.6989,
      "theoretical_loss": 3.4152264965964108,
      "tokens_seen": 2109734912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003669090909090909,
      "loss": 2.483,
      "theoretical_loss": 3.4152177438905893,
      "tokens_seen": 2109800448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668888888888889,
      "loss": 2.5866,
      "theoretical_loss": 3.4152089915327695,
      "tokens_seen": 2109865984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036686868686868683,
      "loss": 2.5572,
      "theoretical_loss": 3.415200239522928,
      "tokens_seen": 2109931520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668484848484849,
      "loss": 2.5575,
      "theoretical_loss": 3.4151914878610397,
      "tokens_seen": 2109997056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036682828282828285,
      "loss": 2.4725,
      "theoretical_loss": 3.4151827365470795,
      "tokens_seen": 2110062592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668080808080808,
      "loss": 2.7051,
      "theoretical_loss": 3.4151739855810233,
      "tokens_seen": 2110128128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036678787878787877,
      "loss": 2.5453,
      "theoretical_loss": 3.4151652349628465,
      "tokens_seen": 2110193664
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667676767676767,
      "loss": 2.5881,
      "theoretical_loss": 3.415156484692524,
      "tokens_seen": 2110259200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667474747474748,
      "loss": 2.6439,
      "theoretical_loss": 3.415147734770031,
      "tokens_seen": 2110324736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036672727272727275,
      "loss": 2.6104,
      "theoretical_loss": 3.4151389851953438,
      "tokens_seen": 2110390272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667070707070707,
      "loss": 2.6512,
      "theoretical_loss": 3.4151302359684372,
      "tokens_seen": 2110455808
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1163684,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.875441312789917,
      "objective/train/theoretical_loss": 3.4151214870892868,
      "objective/train/tokens_used": 469380576,
      "theoretical_loss": 3.4151214870892868,
      "tokens_seen": 2110521344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036668686868686866,
      "loss": 2.6088,
      "theoretical_loss": 3.4151214870892868,
      "tokens_seen": 2110521344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036666666666666667,
      "loss": 2.6473,
      "theoretical_loss": 3.4151127385578675,
      "tokens_seen": 2110586880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666464646464647,
      "loss": 2.6037,
      "theoretical_loss": 3.4151039903741554,
      "tokens_seen": 2110652416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036662626262626264,
      "loss": 2.585,
      "theoretical_loss": 3.415095242538125,
      "tokens_seen": 2110717952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666060606060606,
      "loss": 2.4103,
      "theoretical_loss": 3.4150864950497524,
      "tokens_seen": 2110783488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036658585858585855,
      "loss": 2.4533,
      "theoretical_loss": 3.415077747909013,
      "tokens_seen": 2110849024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036656565656565656,
      "loss": 2.5017,
      "theoretical_loss": 3.4150690011158815,
      "tokens_seen": 2110914560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665454545454546,
      "loss": 2.5898,
      "theoretical_loss": 3.415060254670334,
      "tokens_seen": 2110980096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036652525252525253,
      "loss": 2.7279,
      "theoretical_loss": 3.415051508572346,
      "tokens_seen": 2111045632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665050505050505,
      "loss": 2.6679,
      "theoretical_loss": 3.415042762821892,
      "tokens_seen": 2111111168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664848484848485,
      "loss": 2.4494,
      "theoretical_loss": 3.4150340174189484,
      "tokens_seen": 2111176704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036646464646464646,
      "loss": 2.3438,
      "theoretical_loss": 3.41502527236349,
      "tokens_seen": 2111242240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036644444444444447,
      "loss": 2.6134,
      "theoretical_loss": 3.4150165276554922,
      "tokens_seen": 2111307776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664242424242424,
      "loss": 2.7989,
      "theoretical_loss": 3.4150077832949304,
      "tokens_seen": 2111373312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664040404040404,
      "loss": 2.7628,
      "theoretical_loss": 3.4149990392817804,
      "tokens_seen": 2111438848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663838383838384,
      "loss": 2.7186,
      "theoretical_loss": 3.4149902956160174,
      "tokens_seen": 2111504384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663636363636364,
      "loss": 2.6454,
      "theoretical_loss": 3.414981552297617,
      "tokens_seen": 2111569920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036634343434343436,
      "loss": 2.7228,
      "theoretical_loss": 3.4149728093265543,
      "tokens_seen": 2111635456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663232323232323,
      "loss": 2.5719,
      "theoretical_loss": 3.414964066702805,
      "tokens_seen": 2111700992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036630303030303033,
      "loss": 2.6716,
      "theoretical_loss": 3.4149553244263444,
      "tokens_seen": 2111766528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662828282828283,
      "loss": 2.6835,
      "theoretical_loss": 3.414946582497148,
      "tokens_seen": 2111832064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662626262626263,
      "loss": 2.7341,
      "theoretical_loss": 3.4149378409151905,
      "tokens_seen": 2111897600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036624242424242425,
      "loss": 2.7706,
      "theoretical_loss": 3.4149290996804487,
      "tokens_seen": 2111963136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662222222222222,
      "loss": 2.8759,
      "theoretical_loss": 3.414920358792897,
      "tokens_seen": 2112028672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662020202020202,
      "loss": 2.6096,
      "theoretical_loss": 3.414911618252511,
      "tokens_seen": 2112094208
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 1164108,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.35992169380188,
      "objective/train/theoretical_loss": 3.4149028780592667,
      "objective/train/tokens_used": 471018976,
      "theoretical_loss": 3.4149028780592667,
      "tokens_seen": 2112159744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661818181818182,
      "loss": 2.5137,
      "theoretical_loss": 3.4149028780592667,
      "tokens_seen": 2112159744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661616161616162,
      "loss": 2.7378,
      "theoretical_loss": 3.4148941382131386,
      "tokens_seen": 2112225280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036614141414141415,
      "loss": 2.695,
      "theoretical_loss": 3.414885398714103,
      "tokens_seen": 2112290816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036612121212121216,
      "loss": 2.3199,
      "theoretical_loss": 3.414876659562135,
      "tokens_seen": 2112356352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661010101010101,
      "loss": 2.6186,
      "theoretical_loss": 3.41486792075721,
      "tokens_seen": 2112421888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036608080808080807,
      "loss": 2.558,
      "theoretical_loss": 3.4148591822993035,
      "tokens_seen": 2112487424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660606060606061,
      "loss": 2.4394,
      "theoretical_loss": 3.4148504441883913,
      "tokens_seen": 2112552960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036604040404040404,
      "loss": 2.7204,
      "theoretical_loss": 3.4148417064244487,
      "tokens_seen": 2112618496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036602020202020205,
      "loss": 2.5691,
      "theoretical_loss": 3.414832969007451,
      "tokens_seen": 2112684032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000366,
      "loss": 2.4832,
      "theoretical_loss": 3.4148242319373727,
      "tokens_seen": 2112749568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036597979797979796,
      "loss": 2.6831,
      "theoretical_loss": 3.414815495214191,
      "tokens_seen": 2112815104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365959595959596,
      "loss": 2.7641,
      "theoretical_loss": 3.41480675883788,
      "tokens_seen": 2112880640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365939393939394,
      "loss": 2.7619,
      "theoretical_loss": 3.4147980228084167,
      "tokens_seen": 2112946176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036591919191919194,
      "loss": 2.6204,
      "theoretical_loss": 3.4147892871257746,
      "tokens_seen": 2113011712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658989898989899,
      "loss": 2.7348,
      "theoretical_loss": 3.414780551789931,
      "tokens_seen": 2113077248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036587878787878786,
      "loss": 2.4798,
      "theoretical_loss": 3.4147718168008607,
      "tokens_seen": 2113142784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036585858585858587,
      "loss": 2.4059,
      "theoretical_loss": 3.4147630821585384,
      "tokens_seen": 2113208320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658383838383839,
      "loss": 2.6539,
      "theoretical_loss": 3.4147543478629405,
      "tokens_seen": 2113273856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036581818181818183,
      "loss": 2.4668,
      "theoretical_loss": 3.414745613914042,
      "tokens_seen": 2113339392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657979797979798,
      "loss": 2.6112,
      "theoretical_loss": 3.4147368803118194,
      "tokens_seen": 2113404928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036577777777777775,
      "loss": 2.7575,
      "theoretical_loss": 3.4147281470562465,
      "tokens_seen": 2113470464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657575757575758,
      "loss": 2.5115,
      "theoretical_loss": 3.4147194141473003,
      "tokens_seen": 2113536000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036573737373737377,
      "loss": 2.756,
      "theoretical_loss": 3.414710681584956,
      "tokens_seen": 2113601536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036571717171717173,
      "loss": 2.88,
      "theoretical_loss": 3.4147019493691877,
      "tokens_seen": 2113667072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656969696969697,
      "loss": 2.5343,
      "theoretical_loss": 3.414693217499973,
      "tokens_seen": 2113732608
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1165368,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.723964214324951,
      "objective/train/theoretical_loss": 3.414684485977286,
      "objective/train/tokens_used": 472657376,
      "theoretical_loss": 3.414684485977286,
      "tokens_seen": 2113798144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036567676767676764,
      "loss": 2.5892,
      "theoretical_loss": 3.414684485977286,
      "tokens_seen": 2113798144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656565656565657,
      "loss": 2.8461,
      "theoretical_loss": 3.4146757548011024,
      "tokens_seen": 2113863680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036563636363636366,
      "loss": 2.599,
      "theoretical_loss": 3.4146670239713983,
      "tokens_seen": 2113929216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656161616161616,
      "loss": 2.7025,
      "theoretical_loss": 3.4146582934881486,
      "tokens_seen": 2113994752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655959595959596,
      "loss": 2.6218,
      "theoretical_loss": 3.414649563351329,
      "tokens_seen": 2114060288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655757575757576,
      "loss": 2.7486,
      "theoretical_loss": 3.4146408335609157,
      "tokens_seen": 2114125824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655555555555556,
      "loss": 2.6214,
      "theoretical_loss": 3.414632104116883,
      "tokens_seen": 2114191360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036553535353535356,
      "loss": 2.576,
      "theoretical_loss": 3.4146233750192065,
      "tokens_seen": 2114256896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655151515151515,
      "loss": 2.5162,
      "theoretical_loss": 3.4146146462678626,
      "tokens_seen": 2114322432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036549494949494947,
      "loss": 2.4917,
      "theoretical_loss": 3.4146059178628265,
      "tokens_seen": 2114387968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654747474747475,
      "loss": 2.6346,
      "theoretical_loss": 3.414597189804074,
      "tokens_seen": 2114453504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654545454545455,
      "loss": 2.5658,
      "theoretical_loss": 3.4145884620915803,
      "tokens_seen": 2114519040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036543434343434345,
      "loss": 2.7962,
      "theoretical_loss": 3.4145797347253204,
      "tokens_seen": 2114584576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654141414141414,
      "loss": 2.6246,
      "theoretical_loss": 3.4145710077052707,
      "tokens_seen": 2114650112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653939393939394,
      "loss": 2.6092,
      "theoretical_loss": 3.4145622810314062,
      "tokens_seen": 2114715648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653737373737374,
      "loss": 2.584,
      "theoretical_loss": 3.4145535547037023,
      "tokens_seen": 2114781184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653535353535354,
      "loss": 2.6047,
      "theoretical_loss": 3.4145448287221356,
      "tokens_seen": 2114846720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036533333333333334,
      "loss": 2.5933,
      "theoretical_loss": 3.4145361030866805,
      "tokens_seen": 2114912256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653131313131313,
      "loss": 2.4416,
      "theoretical_loss": 3.414527377797313,
      "tokens_seen": 2114977792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652929292929293,
      "loss": 2.6182,
      "theoretical_loss": 3.414518652854009,
      "tokens_seen": 2115043328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036527272727272727,
      "loss": 2.7091,
      "theoretical_loss": 3.4145099282567433,
      "tokens_seen": 2115108864
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652525252525253,
      "loss": 2.5143,
      "theoretical_loss": 3.414501204005492,
      "tokens_seen": 2115174400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036523232323232323,
      "loss": 2.54,
      "theoretical_loss": 3.41449248010023,
      "tokens_seen": 2115239936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036521212121212124,
      "loss": 2.72,
      "theoretical_loss": 3.4144837565409336,
      "tokens_seen": 2115305472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003651919191919192,
      "loss": 2.6686,
      "theoretical_loss": 3.4144750333275784,
      "tokens_seen": 2115371008
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1165914,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.618112087249756,
      "objective/train/theoretical_loss": 3.4144663104601394,
      "objective/train/tokens_used": 474295776,
      "theoretical_loss": 3.4144663104601394,
      "tokens_seen": 2115436544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036517171717171716,
      "loss": 2.641,
      "theoretical_loss": 3.4144663104601394,
      "tokens_seen": 2115436544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036515151515151517,
      "loss": 2.5092,
      "theoretical_loss": 3.4144575879385926,
      "tokens_seen": 2115502080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003651313131313131,
      "loss": 2.537,
      "theoretical_loss": 3.414448865762913,
      "tokens_seen": 2115567616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036511111111111114,
      "loss": 2.8961,
      "theoretical_loss": 3.414440143933077,
      "tokens_seen": 2115633152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650909090909091,
      "loss": 2.5186,
      "theoretical_loss": 3.4144314224490597,
      "tokens_seen": 2115698688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036507070707070705,
      "loss": 2.5034,
      "theoretical_loss": 3.414422701310837,
      "tokens_seen": 2115764224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036505050505050506,
      "loss": 2.5035,
      "theoretical_loss": 3.4144139805183835,
      "tokens_seen": 2115829760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365030303030303,
      "loss": 2.7573,
      "theoretical_loss": 3.4144052600716757,
      "tokens_seen": 2115895296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036501010101010103,
      "loss": 2.6071,
      "theoretical_loss": 3.4143965399706895,
      "tokens_seen": 2115960832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000364989898989899,
      "loss": 2.6751,
      "theoretical_loss": 3.414387820215399,
      "tokens_seen": 2116026368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036496969696969694,
      "loss": 2.771,
      "theoretical_loss": 3.4143791008057818,
      "tokens_seen": 2116091904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036494949494949496,
      "loss": 2.4936,
      "theoretical_loss": 3.4143703817418123,
      "tokens_seen": 2116157440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036492929292929297,
      "loss": 2.551,
      "theoretical_loss": 3.4143616630234654,
      "tokens_seen": 2116222976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649090909090909,
      "loss": 2.6326,
      "theoretical_loss": 3.4143529446507186,
      "tokens_seen": 2116288512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648888888888889,
      "loss": 2.6491,
      "theoretical_loss": 3.4143442266235455,
      "tokens_seen": 2116354048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036486868686868684,
      "loss": 2.6087,
      "theoretical_loss": 3.4143355089419227,
      "tokens_seen": 2116419584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036484848484848485,
      "loss": 2.692,
      "theoretical_loss": 3.414326791605826,
      "tokens_seen": 2116485120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036482828282828286,
      "loss": 2.2571,
      "theoretical_loss": 3.414318074615231,
      "tokens_seen": 2116550656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648080808080808,
      "loss": 2.6154,
      "theoretical_loss": 3.414309357970113,
      "tokens_seen": 2116616192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036478787878787877,
      "loss": 2.5167,
      "theoretical_loss": 3.4143006416704473,
      "tokens_seen": 2116681728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036476767676767673,
      "loss": 2.6605,
      "theoretical_loss": 3.41429192571621,
      "tokens_seen": 2116747264
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647474747474748,
      "loss": 2.6692,
      "theoretical_loss": 3.414283210107377,
      "tokens_seen": 2116812800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036472727272727275,
      "loss": 2.573,
      "theoretical_loss": 3.414274494843923,
      "tokens_seen": 2116878336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647070707070707,
      "loss": 2.8211,
      "theoretical_loss": 3.4142657799258243,
      "tokens_seen": 2116943872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036468686868686867,
      "loss": 2.7455,
      "theoretical_loss": 3.4142570653530564,
      "tokens_seen": 2117009408
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1167104,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4354066848754883,
      "objective/train/theoretical_loss": 3.414248351125595,
      "objective/train/tokens_used": 475934176,
      "theoretical_loss": 3.414248351125595,
      "tokens_seen": 2117074944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646666666666666,
      "loss": 2.5663,
      "theoretical_loss": 3.414248351125595,
      "tokens_seen": 2117074944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646464646464647,
      "loss": 2.4757,
      "theoretical_loss": 3.414239637243415,
      "tokens_seen": 2117140480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036462626262626264,
      "loss": 2.6941,
      "theoretical_loss": 3.4142309237064934,
      "tokens_seen": 2117206016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646060606060606,
      "loss": 2.6617,
      "theoretical_loss": 3.414222210514805,
      "tokens_seen": 2117271552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036458585858585856,
      "loss": 2.5227,
      "theoretical_loss": 3.4142134976683254,
      "tokens_seen": 2117337088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645656565656566,
      "loss": 2.5767,
      "theoretical_loss": 3.41420478516703,
      "tokens_seen": 2117402624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645454545454546,
      "loss": 2.6952,
      "theoretical_loss": 3.414196073010895,
      "tokens_seen": 2117468160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036452525252525254,
      "loss": 2.3163,
      "theoretical_loss": 3.414187361199896,
      "tokens_seen": 2117533696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645050505050505,
      "loss": 2.5076,
      "theoretical_loss": 3.414178649734008,
      "tokens_seen": 2117599232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036448484848484845,
      "loss": 2.5644,
      "theoretical_loss": 3.414169938613208,
      "tokens_seen": 2117664768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644646464646465,
      "loss": 2.6019,
      "theoretical_loss": 3.41416122783747,
      "tokens_seen": 2117730304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036444444444444447,
      "loss": 2.6005,
      "theoretical_loss": 3.4141525174067704,
      "tokens_seen": 2117795840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036442424242424243,
      "loss": 2.6914,
      "theoretical_loss": 3.4141438073210857,
      "tokens_seen": 2117861376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644040404040404,
      "loss": 2.4907,
      "theoretical_loss": 3.41413509758039,
      "tokens_seen": 2117926912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643838383838384,
      "loss": 2.7162,
      "theoretical_loss": 3.41412638818466,
      "tokens_seen": 2117992448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643636363636364,
      "loss": 2.4166,
      "theoretical_loss": 3.414117679133871,
      "tokens_seen": 2118057984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036434343434343437,
      "loss": 2.7053,
      "theoretical_loss": 3.414108970427999,
      "tokens_seen": 2118123520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643232323232323,
      "loss": 2.5607,
      "theoretical_loss": 3.414100262067019,
      "tokens_seen": 2118189056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643030303030303,
      "loss": 2.5588,
      "theoretical_loss": 3.4140915540509074,
      "tokens_seen": 2118254592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642828282828283,
      "loss": 2.6071,
      "theoretical_loss": 3.4140828463796398,
      "tokens_seen": 2118320128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642626262626263,
      "loss": 2.4847,
      "theoretical_loss": 3.4140741390531915,
      "tokens_seen": 2118385664
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036424242424242426,
      "loss": 2.636,
      "theoretical_loss": 3.4140654320715376,
      "tokens_seen": 2118451200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642222222222222,
      "loss": 2.6212,
      "theoretical_loss": 3.414056725434655,
      "tokens_seen": 2118516736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642020202020202,
      "loss": 2.7221,
      "theoretical_loss": 3.414048019142519,
      "tokens_seen": 2118582272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641818181818182,
      "loss": 2.551,
      "theoretical_loss": 3.4140393131951052,
      "tokens_seen": 2118647808
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1168338,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4486634731292725,
      "objective/train/theoretical_loss": 3.414030607592389,
      "objective/train/tokens_used": 477572576,
      "theoretical_loss": 3.414030607592389,
      "tokens_seen": 2118713344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641616161616162,
      "loss": 2.5584,
      "theoretical_loss": 3.414030607592389,
      "tokens_seen": 2118713344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036414141414141415,
      "loss": 2.6926,
      "theoretical_loss": 3.4140219023343468,
      "tokens_seen": 2118778880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641212121212121,
      "loss": 2.5152,
      "theoretical_loss": 3.414013197420953,
      "tokens_seen": 2118844416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641010101010101,
      "loss": 2.7168,
      "theoretical_loss": 3.414004492852185,
      "tokens_seen": 2118909952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640808080808081,
      "loss": 2.3342,
      "theoretical_loss": 3.413995788628017,
      "tokens_seen": 2118975488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640606060606061,
      "loss": 2.6007,
      "theoretical_loss": 3.4139870847484257,
      "tokens_seen": 2119041024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036404040404040404,
      "loss": 2.5816,
      "theoretical_loss": 3.4139783812133864,
      "tokens_seen": 2119106560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036402020202020205,
      "loss": 2.4268,
      "theoretical_loss": 3.4139696780228745,
      "tokens_seen": 2119172096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000364,
      "loss": 2.7488,
      "theoretical_loss": 3.4139609751768663,
      "tokens_seen": 2119237632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036397979797979797,
      "loss": 2.7172,
      "theoretical_loss": 3.413952272675337,
      "tokens_seen": 2119303168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363959595959596,
      "loss": 2.4971,
      "theoretical_loss": 3.413943570518263,
      "tokens_seen": 2119368704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036393939393939394,
      "loss": 2.6373,
      "theoretical_loss": 3.4139348687056192,
      "tokens_seen": 2119434240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036391919191919195,
      "loss": 2.6311,
      "theoretical_loss": 3.413926167237382,
      "tokens_seen": 2119499776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638989898989899,
      "loss": 2.621,
      "theoretical_loss": 3.4139174661135265,
      "tokens_seen": 2119565312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036387878787878786,
      "loss": 2.439,
      "theoretical_loss": 3.413908765334029,
      "tokens_seen": 2119630848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036385858585858587,
      "loss": 2.5993,
      "theoretical_loss": 3.413900064898865,
      "tokens_seen": 2119696384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638383838383839,
      "loss": 2.5193,
      "theoretical_loss": 3.41389136480801,
      "tokens_seen": 2119761920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036381818181818184,
      "loss": 2.5325,
      "theoretical_loss": 3.4138826650614402,
      "tokens_seen": 2119827456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637979797979798,
      "loss": 2.5291,
      "theoretical_loss": 3.413873965659131,
      "tokens_seen": 2119892992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036377777777777775,
      "loss": 2.6428,
      "theoretical_loss": 3.413865266601058,
      "tokens_seen": 2119958528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036375757575757576,
      "loss": 2.7483,
      "theoretical_loss": 3.4138565678871973,
      "tokens_seen": 2120024064
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637373737373738,
      "loss": 2.5676,
      "theoretical_loss": 3.4138478695175243,
      "tokens_seen": 2120089600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036371717171717173,
      "loss": 2.5161,
      "theoretical_loss": 3.4138391714920147,
      "tokens_seen": 2120155136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636969696969697,
      "loss": 2.5857,
      "theoretical_loss": 3.413830473810645,
      "tokens_seen": 2120220672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036367676767676765,
      "loss": 2.5532,
      "theoretical_loss": 3.41382177647339,
      "tokens_seen": 2120286208
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1168872,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.698695182800293,
      "objective/train/theoretical_loss": 3.413813079480226,
      "objective/train/tokens_used": 479210976,
      "theoretical_loss": 3.413813079480226,
      "tokens_seen": 2120351744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636565656565657,
      "loss": 2.5234,
      "theoretical_loss": 3.413813079480226,
      "tokens_seen": 2120351744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036363636363636367,
      "loss": 2.4915,
      "theoretical_loss": 3.413804382831129,
      "tokens_seen": 2120417280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636161616161616,
      "loss": 2.7313,
      "theoretical_loss": 3.4137956865260737,
      "tokens_seen": 2120482816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003635959595959596,
      "loss": 2.7314,
      "theoretical_loss": 3.413786990565037,
      "tokens_seen": 2120548352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036357575757575754,
      "loss": 2.8578,
      "theoretical_loss": 3.413778294947994,
      "tokens_seen": 2120613888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003635555555555556,
      "loss": 2.6198,
      "theoretical_loss": 3.4137695996749207,
      "tokens_seen": 2120679424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036353535353535356,
      "loss": 2.5495,
      "theoretical_loss": 3.413760904745793,
      "tokens_seen": 2120744960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003635151515151515,
      "loss": 2.5978,
      "theoretical_loss": 3.413752210160586,
      "tokens_seen": 2120810496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634949494949495,
      "loss": 2.6271,
      "theoretical_loss": 3.413743515919276,
      "tokens_seen": 2120876032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036347474747474743,
      "loss": 2.5787,
      "theoretical_loss": 3.413734822021839,
      "tokens_seen": 2120941568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634545454545455,
      "loss": 2.6924,
      "theoretical_loss": 3.4137261284682503,
      "tokens_seen": 2121007104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036343434343434345,
      "loss": 2.723,
      "theoretical_loss": 3.413717435258486,
      "tokens_seen": 2121072640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634141414141414,
      "loss": 2.3627,
      "theoretical_loss": 3.4137087423925214,
      "tokens_seen": 2121138176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036339393939393937,
      "loss": 2.3717,
      "theoretical_loss": 3.413700049870333,
      "tokens_seen": 2121203712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633737373737374,
      "loss": 2.6768,
      "theoretical_loss": 3.413691357691896,
      "tokens_seen": 2121269248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633535353535354,
      "loss": 2.5494,
      "theoretical_loss": 3.413682665857187,
      "tokens_seen": 2121334784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036333333333333335,
      "loss": 2.5297,
      "theoretical_loss": 3.4136739743661804,
      "tokens_seen": 2121400320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633131313131313,
      "loss": 2.4224,
      "theoretical_loss": 3.4136652832188528,
      "tokens_seen": 2121465856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036329292929292926,
      "loss": 2.5031,
      "theoretical_loss": 3.4136565924151805,
      "tokens_seen": 2121531392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036327272727272727,
      "loss": 2.5477,
      "theoretical_loss": 3.4136479019551382,
      "tokens_seen": 2121596928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632525252525253,
      "loss": 2.661,
      "theoretical_loss": 3.413639211838703,
      "tokens_seen": 2121662464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036323232323232324,
      "loss": 2.8096,
      "theoretical_loss": 3.4136305220658496,
      "tokens_seen": 2121728000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632121212121212,
      "loss": 2.763,
      "theoretical_loss": 3.413621832636554,
      "tokens_seen": 2121793536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631919191919192,
      "loss": 2.5383,
      "theoretical_loss": 3.4136131435507924,
      "tokens_seen": 2121859072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036317171717171716,
      "loss": 2.386,
      "theoretical_loss": 3.4136044548085405,
      "tokens_seen": 2121924608
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1169826,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2446186542510986,
      "objective/train/theoretical_loss": 3.413595766409774,
      "objective/train/tokens_used": 480849376,
      "theoretical_loss": 3.413595766409774,
      "tokens_seen": 2121990144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631515151515152,
      "loss": 2.4437,
      "theoretical_loss": 3.413595766409774,
      "tokens_seen": 2121990144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036313131313131313,
      "loss": 2.7176,
      "theoretical_loss": 3.413587078354468,
      "tokens_seen": 2122055680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631111111111111,
      "loss": 2.379,
      "theoretical_loss": 3.4135783906425994,
      "tokens_seen": 2122121216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003630909090909091,
      "loss": 2.5955,
      "theoretical_loss": 3.4135697032741437,
      "tokens_seen": 2122186752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036307070707070706,
      "loss": 2.6066,
      "theoretical_loss": 3.4135610162490773,
      "tokens_seen": 2122252288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036305050505050507,
      "loss": 2.5977,
      "theoretical_loss": 3.4135523295673744,
      "tokens_seen": 2122317824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363030303030303,
      "loss": 2.5323,
      "theoretical_loss": 3.4135436432290125,
      "tokens_seen": 2122383360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036301010101010104,
      "loss": 2.557,
      "theoretical_loss": 3.413534957233966,
      "tokens_seen": 2122448896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362989898989899,
      "loss": 2.6608,
      "theoretical_loss": 3.4135262715822123,
      "tokens_seen": 2122514432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036296969696969695,
      "loss": 2.428,
      "theoretical_loss": 3.413517586273726,
      "tokens_seen": 2122579968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036294949494949496,
      "loss": 2.7387,
      "theoretical_loss": 3.4135089013084836,
      "tokens_seen": 2122645504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629292929292929,
      "loss": 2.6331,
      "theoretical_loss": 3.4135002166864608,
      "tokens_seen": 2122711040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036290909090909093,
      "loss": 2.6378,
      "theoretical_loss": 3.413491532407633,
      "tokens_seen": 2122776576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628888888888889,
      "loss": 2.6081,
      "theoretical_loss": 3.4134828484719764,
      "tokens_seen": 2122842112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628686868686869,
      "loss": 2.6682,
      "theoretical_loss": 3.4134741648794664,
      "tokens_seen": 2122907648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036284848484848485,
      "loss": 2.5269,
      "theoretical_loss": 3.4134654816300802,
      "tokens_seen": 2122973184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036282828282828286,
      "loss": 2.6664,
      "theoretical_loss": 3.413456798723792,
      "tokens_seen": 2123038720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628080808080808,
      "loss": 2.7007,
      "theoretical_loss": 3.4134481161605787,
      "tokens_seen": 2123104256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627878787878788,
      "loss": 2.5572,
      "theoretical_loss": 3.4134394339404155,
      "tokens_seen": 2123169792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627676767676768,
      "loss": 2.7016,
      "theoretical_loss": 3.4134307520632787,
      "tokens_seen": 2123235328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036274747474747475,
      "loss": 2.5064,
      "theoretical_loss": 3.413422070529144,
      "tokens_seen": 2123300864
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036272727272727276,
      "loss": 2.6765,
      "theoretical_loss": 3.413413389337988,
      "tokens_seen": 2123366400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627070707070707,
      "loss": 2.5349,
      "theoretical_loss": 3.413404708489785,
      "tokens_seen": 2123431936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036268686868686867,
      "loss": 2.7782,
      "theoretical_loss": 3.4133960279845117,
      "tokens_seen": 2123497472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626666666666667,
      "loss": 2.6642,
      "theoretical_loss": 3.4133873478221446,
      "tokens_seen": 2123563008
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1170397,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.436032295227051,
      "objective/train/theoretical_loss": 3.4133786680026583,
      "objective/train/tokens_used": 482487776,
      "theoretical_loss": 3.4133786680026583,
      "tokens_seen": 2123628544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626464646464647,
      "loss": 2.5855,
      "theoretical_loss": 3.4133786680026583,
      "tokens_seen": 2123628544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036262626262626265,
      "loss": 2.5702,
      "theoretical_loss": 3.41336998852603,
      "tokens_seen": 2123694080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626060606060606,
      "loss": 2.4347,
      "theoretical_loss": 3.4133613093922346,
      "tokens_seen": 2123759616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036258585858585856,
      "loss": 2.8192,
      "theoretical_loss": 3.4133526306012483,
      "tokens_seen": 2123825152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625656565656566,
      "loss": 2.6402,
      "theoretical_loss": 3.413343952153047,
      "tokens_seen": 2123890688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625454545454546,
      "loss": 2.6956,
      "theoretical_loss": 3.4133352740476064,
      "tokens_seen": 2123956224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036252525252525254,
      "loss": 2.2476,
      "theoretical_loss": 3.413326596284903,
      "tokens_seen": 2124021760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625050505050505,
      "loss": 2.7057,
      "theoretical_loss": 3.413317918864912,
      "tokens_seen": 2124087296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036248484848484846,
      "loss": 2.4141,
      "theoretical_loss": 3.4133092417876094,
      "tokens_seen": 2124152832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624646464646465,
      "loss": 2.5951,
      "theoretical_loss": 3.4133005650529715,
      "tokens_seen": 2124218368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624444444444445,
      "loss": 2.6807,
      "theoretical_loss": 3.4132918886609733,
      "tokens_seen": 2124283904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036242424242424244,
      "loss": 2.5597,
      "theoretical_loss": 3.413283212611592,
      "tokens_seen": 2124349440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624040404040404,
      "loss": 2.4826,
      "theoretical_loss": 3.4132745369048028,
      "tokens_seen": 2124414976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036238383838383835,
      "loss": 2.698,
      "theoretical_loss": 3.413265861540582,
      "tokens_seen": 2124480512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623636363636364,
      "loss": 2.5928,
      "theoretical_loss": 3.413257186518904,
      "tokens_seen": 2124546048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036234343434343437,
      "loss": 2.7756,
      "theoretical_loss": 3.413248511839747,
      "tokens_seen": 2124611584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036232323232323233,
      "loss": 2.6898,
      "theoretical_loss": 3.413239837503085,
      "tokens_seen": 2124677120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623030303030303,
      "loss": 2.4804,
      "theoretical_loss": 3.413231163508895,
      "tokens_seen": 2124742656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622828282828283,
      "loss": 2.5586,
      "theoretical_loss": 3.4132224898571524,
      "tokens_seen": 2124808192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622626262626263,
      "loss": 2.4899,
      "theoretical_loss": 3.4132138165478336,
      "tokens_seen": 2124873728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036224242424242426,
      "loss": 2.7261,
      "theoretical_loss": 3.413205143580914,
      "tokens_seen": 2124939264
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622222222222222,
      "loss": 2.483,
      "theoretical_loss": 3.41319647095637,
      "tokens_seen": 2125004800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622020202020202,
      "loss": 2.6043,
      "theoretical_loss": 3.413187798674177,
      "tokens_seen": 2125070336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621818181818182,
      "loss": 2.6508,
      "theoretical_loss": 3.4131791267343115,
      "tokens_seen": 2125135872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621616161616162,
      "loss": 2.4916,
      "theoretical_loss": 3.4131704551367488,
      "tokens_seen": 2125201408
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1171543,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5023303031921387,
      "objective/train/theoretical_loss": 3.413161783881465,
      "objective/train/tokens_used": 484126176,
      "theoretical_loss": 3.413161783881465,
      "tokens_seen": 2125266944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036214141414141416,
      "loss": 2.5373,
      "theoretical_loss": 3.413161783881465,
      "tokens_seen": 2125266944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621212121212121,
      "loss": 2.7232,
      "theoretical_loss": 3.4131531129684367,
      "tokens_seen": 2125332480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621010101010101,
      "loss": 2.3724,
      "theoretical_loss": 3.4131444423976394,
      "tokens_seen": 2125398016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620808080808081,
      "loss": 2.594,
      "theoretical_loss": 3.4131357721690487,
      "tokens_seen": 2125463552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620606060606061,
      "loss": 2.4723,
      "theoretical_loss": 3.4131271022826413,
      "tokens_seen": 2125529088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036204040404040405,
      "loss": 2.6411,
      "theoretical_loss": 3.4131184327383917,
      "tokens_seen": 2125594624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362020202020202,
      "loss": 2.4801,
      "theoretical_loss": 3.4131097635362777,
      "tokens_seen": 2125660160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362,
      "loss": 2.819,
      "theoretical_loss": 3.413101094676274,
      "tokens_seen": 2125725696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361979797979798,
      "loss": 2.5704,
      "theoretical_loss": 3.4130924261583573,
      "tokens_seen": 2125791232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361959595959596,
      "loss": 2.5092,
      "theoretical_loss": 3.413083757982503,
      "tokens_seen": 2125856768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036193939393939394,
      "loss": 2.5521,
      "theoretical_loss": 3.413075090148687,
      "tokens_seen": 2125922304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003619191919191919,
      "loss": 2.7827,
      "theoretical_loss": 3.413066422656885,
      "tokens_seen": 2125987840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618989898989899,
      "loss": 2.3831,
      "theoretical_loss": 3.413057755507075,
      "tokens_seen": 2126053376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036187878787878787,
      "loss": 2.6201,
      "theoretical_loss": 3.41304908869923,
      "tokens_seen": 2126118912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618585858585859,
      "loss": 2.576,
      "theoretical_loss": 3.413040422233328,
      "tokens_seen": 2126184448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036183838383838383,
      "loss": 2.7195,
      "theoretical_loss": 3.413031756109344,
      "tokens_seen": 2126249984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036181818181818185,
      "loss": 2.4798,
      "theoretical_loss": 3.413023090327255,
      "tokens_seen": 2126315520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003617979797979798,
      "loss": 2.6043,
      "theoretical_loss": 3.413014424887036,
      "tokens_seen": 2126381056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036177777777777776,
      "loss": 2.5492,
      "theoretical_loss": 3.413005759788663,
      "tokens_seen": 2126446592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036175757575757577,
      "loss": 2.4745,
      "theoretical_loss": 3.4129970950321127,
      "tokens_seen": 2126512128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036173737373737373,
      "loss": 2.5508,
      "theoretical_loss": 3.412988430617361,
      "tokens_seen": 2126577664
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036171717171717174,
      "loss": 2.4879,
      "theoretical_loss": 3.4129797665443826,
      "tokens_seen": 2126643200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616969696969697,
      "loss": 2.5086,
      "theoretical_loss": 3.4129711028131546,
      "tokens_seen": 2126708736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036167676767676765,
      "loss": 2.6245,
      "theoretical_loss": 3.4129624394236533,
      "tokens_seen": 2126774272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036165656565656566,
      "loss": 2.6815,
      "theoretical_loss": 3.4129537763758537,
      "tokens_seen": 2126839808
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1172068,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6468193531036377,
      "objective/train/theoretical_loss": 3.412945113669733,
      "objective/train/tokens_used": 485764576,
      "theoretical_loss": 3.412945113669733,
      "tokens_seen": 2126905344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616363636363637,
      "loss": 2.75,
      "theoretical_loss": 3.412945113669733,
      "tokens_seen": 2126905344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036161616161616163,
      "loss": 2.6632,
      "theoretical_loss": 3.4129364513052654,
      "tokens_seen": 2126970880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615959595959596,
      "loss": 2.6412,
      "theoretical_loss": 3.4129277892824286,
      "tokens_seen": 2127036416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036157575757575754,
      "loss": 2.6567,
      "theoretical_loss": 3.4129191276011985,
      "tokens_seen": 2127101952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036155555555555556,
      "loss": 2.5087,
      "theoretical_loss": 3.41291046626155,
      "tokens_seen": 2127167488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036153535353535357,
      "loss": 2.4819,
      "theoretical_loss": 3.4129018052634597,
      "tokens_seen": 2127233024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615151515151515,
      "loss": 2.6294,
      "theoretical_loss": 3.4128931446069037,
      "tokens_seen": 2127298560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614949494949495,
      "loss": 2.5302,
      "theoretical_loss": 3.412884484291858,
      "tokens_seen": 2127364096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036147474747474744,
      "loss": 2.6261,
      "theoretical_loss": 3.4128758243182986,
      "tokens_seen": 2127429632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614545454545455,
      "loss": 2.4306,
      "theoretical_loss": 3.412867164686201,
      "tokens_seen": 2127495168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036143434343434346,
      "loss": 2.6665,
      "theoretical_loss": 3.4128585053955423,
      "tokens_seen": 2127560704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614141414141414,
      "loss": 2.6082,
      "theoretical_loss": 3.4128498464462975,
      "tokens_seen": 2127626240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613939393939394,
      "loss": 2.5125,
      "theoretical_loss": 3.412841187838443,
      "tokens_seen": 2127691776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036137373737373733,
      "loss": 2.8392,
      "theoretical_loss": 3.412832529571955,
      "tokens_seen": 2127757312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613535353535354,
      "loss": 2.4665,
      "theoretical_loss": 3.4128238716468093,
      "tokens_seen": 2127822848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036133333333333335,
      "loss": 2.4348,
      "theoretical_loss": 3.4128152140629817,
      "tokens_seen": 2127888384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613131313131313,
      "loss": 2.5413,
      "theoretical_loss": 3.412806556820449,
      "tokens_seen": 2127953920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036129292929292927,
      "loss": 2.6776,
      "theoretical_loss": 3.4127978999191866,
      "tokens_seen": 2128019456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612727272727273,
      "loss": 2.6288,
      "theoretical_loss": 3.412789243359171,
      "tokens_seen": 2128084992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612525252525253,
      "loss": 2.6286,
      "theoretical_loss": 3.412780587140378,
      "tokens_seen": 2128150528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036123232323232324,
      "loss": 2.4641,
      "theoretical_loss": 3.4127719312627827,
      "tokens_seen": 2128216064
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612121212121212,
      "loss": 2.5957,
      "theoretical_loss": 3.4127632757263626,
      "tokens_seen": 2128281600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036119191919191916,
      "loss": 2.7097,
      "theoretical_loss": 3.4127546205310932,
      "tokens_seen": 2128347136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003611717171717172,
      "loss": 2.5988,
      "theoretical_loss": 3.4127459656769505,
      "tokens_seen": 2128412672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003611515151515152,
      "loss": 2.4821,
      "theoretical_loss": 3.4127373111639105,
      "tokens_seen": 2128478208
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 1173260,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5667147636413574,
      "objective/train/theoretical_loss": 3.4127286569919493,
      "objective/train/tokens_used": 487402976,
      "theoretical_loss": 3.4127286569919493,
      "tokens_seen": 2128543744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036113131313131314,
      "loss": 2.6403,
      "theoretical_loss": 3.4127286569919493,
      "tokens_seen": 2128543744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003611111111111111,
      "loss": 2.5117,
      "theoretical_loss": 3.412720003161043,
      "tokens_seen": 2128609280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610909090909091,
      "loss": 2.5641,
      "theoretical_loss": 3.412711349671168,
      "tokens_seen": 2128674816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610707070707071,
      "loss": 2.5267,
      "theoretical_loss": 3.4127026965222997,
      "tokens_seen": 2128740352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610505050505051,
      "loss": 2.7131,
      "theoretical_loss": 3.4126940437144144,
      "tokens_seen": 2128805888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036103030303030303,
      "loss": 2.558,
      "theoretical_loss": 3.4126853912474884,
      "tokens_seen": 2128871424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000361010101010101,
      "loss": 2.5396,
      "theoretical_loss": 3.4126767391214976,
      "tokens_seen": 2128936960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360989898989899,
      "loss": 2.6273,
      "theoretical_loss": 3.412668087336418,
      "tokens_seen": 2129002496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360969696969697,
      "loss": 2.7662,
      "theoretical_loss": 3.412659435892226,
      "tokens_seen": 2129068032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036094949494949497,
      "loss": 2.6133,
      "theoretical_loss": 3.412650784788897,
      "tokens_seen": 2129133568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609292929292929,
      "loss": 2.8016,
      "theoretical_loss": 3.412642134026408,
      "tokens_seen": 2129199104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036090909090909093,
      "loss": 2.5226,
      "theoretical_loss": 3.4126334836047345,
      "tokens_seen": 2129264640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608888888888889,
      "loss": 2.4984,
      "theoretical_loss": 3.4126248335238527,
      "tokens_seen": 2129330176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608686868686869,
      "loss": 2.6015,
      "theoretical_loss": 3.4126161837837383,
      "tokens_seen": 2129395712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036084848484848486,
      "loss": 2.6194,
      "theoretical_loss": 3.412607534384368,
      "tokens_seen": 2129461248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608282828282828,
      "loss": 2.5166,
      "theoretical_loss": 3.4125988853257176,
      "tokens_seen": 2129526784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608080808080808,
      "loss": 2.4291,
      "theoretical_loss": 3.4125902366077634,
      "tokens_seen": 2129592320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607878787878788,
      "loss": 2.5511,
      "theoretical_loss": 3.4125815882304806,
      "tokens_seen": 2129657856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607676767676768,
      "loss": 2.5663,
      "theoretical_loss": 3.412572940193847,
      "tokens_seen": 2129723392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036074747474747475,
      "loss": 2.7557,
      "theoretical_loss": 3.412564292497837,
      "tokens_seen": 2129788928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036072727272727276,
      "loss": 2.5695,
      "theoretical_loss": 3.412555645142428,
      "tokens_seen": 2129854464
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607070707070707,
      "loss": 2.5838,
      "theoretical_loss": 3.4125469981275955,
      "tokens_seen": 2129920000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606868686868687,
      "loss": 2.6765,
      "theoretical_loss": 3.4125383514533154,
      "tokens_seen": 2129985536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606666666666667,
      "loss": 2.641,
      "theoretical_loss": 3.412529705119564,
      "tokens_seen": 2130051072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036064646464646464,
      "loss": 2.5517,
      "theoretical_loss": 3.4125210591263175,
      "tokens_seen": 2130116608
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1174166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.371098041534424,
      "objective/train/theoretical_loss": 3.4125124134735523,
      "objective/train/tokens_used": 489041376,
      "theoretical_loss": 3.4125124134735523,
      "tokens_seen": 2130182144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036062626262626266,
      "loss": 2.5198,
      "theoretical_loss": 3.4125124134735523,
      "tokens_seen": 2130182144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606060606060606,
      "loss": 2.5662,
      "theoretical_loss": 3.412503768161244,
      "tokens_seen": 2130247680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036058585858585857,
      "loss": 2.6296,
      "theoretical_loss": 3.4124951231893688,
      "tokens_seen": 2130313216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605656565656566,
      "loss": 2.5507,
      "theoretical_loss": 3.412486478557903,
      "tokens_seen": 2130378752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605454545454546,
      "loss": 2.799,
      "theoretical_loss": 3.4124778342668227,
      "tokens_seen": 2130444288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036052525252525255,
      "loss": 2.7239,
      "theoretical_loss": 3.412469190316104,
      "tokens_seen": 2130509824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605050505050505,
      "loss": 2.7608,
      "theoretical_loss": 3.412460546705723,
      "tokens_seen": 2130575360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036048484848484846,
      "loss": 2.6064,
      "theoretical_loss": 3.412451903435656,
      "tokens_seen": 2130640896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036046464646464647,
      "loss": 2.5552,
      "theoretical_loss": 3.4124432605058788,
      "tokens_seen": 2130706432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604444444444445,
      "loss": 2.9222,
      "theoretical_loss": 3.412434617916368,
      "tokens_seen": 2130771968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036042424242424244,
      "loss": 2.6447,
      "theoretical_loss": 3.412425975667099,
      "tokens_seen": 2130837504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604040404040404,
      "loss": 2.6339,
      "theoretical_loss": 3.4124173337580492,
      "tokens_seen": 2130903040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036038383838383835,
      "loss": 2.6131,
      "theoretical_loss": 3.412408692189193,
      "tokens_seen": 2130968576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003603636363636364,
      "loss": 2.716,
      "theoretical_loss": 3.412400050960508,
      "tokens_seen": 2131034112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003603434343434344,
      "loss": 2.7042,
      "theoretical_loss": 3.4123914100719697,
      "tokens_seen": 2131099648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036032323232323233,
      "loss": 2.5032,
      "theoretical_loss": 3.4123827695235542,
      "tokens_seen": 2131165184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003603030303030303,
      "loss": 2.6595,
      "theoretical_loss": 3.4123741293152383,
      "tokens_seen": 2131230720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036028282828282825,
      "loss": 2.715,
      "theoretical_loss": 3.412365489446998,
      "tokens_seen": 2131296256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602626262626263,
      "loss": 2.547,
      "theoretical_loss": 3.4123568499188086,
      "tokens_seen": 2131361792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036024242424242427,
      "loss": 2.7751,
      "theoretical_loss": 3.412348210730647,
      "tokens_seen": 2131427328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602222222222222,
      "loss": 2.4854,
      "theoretical_loss": 3.412339571882489,
      "tokens_seen": 2131492864
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602020202020202,
      "loss": 2.5917,
      "theoretical_loss": 3.412330933374311,
      "tokens_seen": 2131558400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036018181818181814,
      "loss": 2.6232,
      "theoretical_loss": 3.4123222952060894,
      "tokens_seen": 2131623936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601616161616162,
      "loss": 2.6344,
      "theoretical_loss": 3.4123136573777995,
      "tokens_seen": 2131689472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036014141414141416,
      "loss": 2.6682,
      "theoretical_loss": 3.412305019889419,
      "tokens_seen": 2131755008
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1174844,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.656928300857544,
      "objective/train/theoretical_loss": 3.412296382740922,
      "objective/train/tokens_used": 490679776,
      "theoretical_loss": 3.412296382740922,
      "tokens_seen": 2131820544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601212121212121,
      "loss": 2.5572,
      "theoretical_loss": 3.412296382740922,
      "tokens_seen": 2131820544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601010101010101,
      "loss": 2.5109,
      "theoretical_loss": 3.4122877459322867,
      "tokens_seen": 2131886080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600808080808081,
      "loss": 2.5999,
      "theoretical_loss": 3.412279109463488,
      "tokens_seen": 2131951616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600606060606061,
      "loss": 2.4027,
      "theoretical_loss": 3.4122704733345026,
      "tokens_seen": 2132017152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036004040404040405,
      "loss": 2.482,
      "theoretical_loss": 3.412261837545307,
      "tokens_seen": 2132082688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360020202020202,
      "loss": 2.4131,
      "theoretical_loss": 3.4122532020958762,
      "tokens_seen": 2132148224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035999999999999997,
      "loss": 2.6003,
      "theoretical_loss": 3.412244566986187,
      "tokens_seen": 2132213760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359979797979798,
      "loss": 2.3046,
      "theoretical_loss": 3.4122359322162166,
      "tokens_seen": 2132279296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359959595959596,
      "loss": 2.5533,
      "theoretical_loss": 3.4122272977859396,
      "tokens_seen": 2132344832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035993939393939395,
      "loss": 2.5854,
      "theoretical_loss": 3.4122186636953336,
      "tokens_seen": 2132410368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003599191919191919,
      "loss": 2.8106,
      "theoretical_loss": 3.412210029944374,
      "tokens_seen": 2132475904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598989898989899,
      "loss": 2.5694,
      "theoretical_loss": 3.4122013965330367,
      "tokens_seen": 2132541440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035987878787878787,
      "loss": 2.6185,
      "theoretical_loss": 3.4121927634612987,
      "tokens_seen": 2132606976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598585858585859,
      "loss": 2.4607,
      "theoretical_loss": 3.4121841307291354,
      "tokens_seen": 2132672512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035983838383838384,
      "loss": 2.7415,
      "theoretical_loss": 3.4121754983365236,
      "tokens_seen": 2132738048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598181818181818,
      "loss": 2.5386,
      "theoretical_loss": 3.4121668662834397,
      "tokens_seen": 2132803584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597979797979798,
      "loss": 2.5452,
      "theoretical_loss": 3.4121582345698593,
      "tokens_seen": 2132869120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035977777777777776,
      "loss": 2.4101,
      "theoretical_loss": 3.4121496031957586,
      "tokens_seen": 2132934656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597575757575758,
      "loss": 2.7979,
      "theoretical_loss": 3.4121409721611147,
      "tokens_seen": 2133000192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035973737373737373,
      "loss": 2.6733,
      "theoretical_loss": 3.412132341465903,
      "tokens_seen": 2133065728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035971717171717174,
      "loss": 2.5141,
      "theoretical_loss": 3.4121237111101,
      "tokens_seen": 2133131264
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596969696969697,
      "loss": 2.503,
      "theoretical_loss": 3.4121150810936816,
      "tokens_seen": 2133196800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035967676767676766,
      "loss": 2.5298,
      "theoretical_loss": 3.4121064514166246,
      "tokens_seen": 2133262336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035965656565656567,
      "loss": 2.6807,
      "theoretical_loss": 3.4120978220789047,
      "tokens_seen": 2133327872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596363636363636,
      "loss": 2.5674,
      "theoretical_loss": 3.412089193080498,
      "tokens_seen": 2133393408
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1175925,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5075652599334717,
      "objective/train/theoretical_loss": 3.412080564421382,
      "objective/train/tokens_used": 492318176,
      "theoretical_loss": 3.412080564421382,
      "tokens_seen": 2133458944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035961616161616164,
      "loss": 2.6865,
      "theoretical_loss": 3.412080564421382,
      "tokens_seen": 2133458944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595959595959596,
      "loss": 2.822,
      "theoretical_loss": 3.4120719361015315,
      "tokens_seen": 2133524480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035957575757575755,
      "loss": 2.541,
      "theoretical_loss": 3.4120633081209233,
      "tokens_seen": 2133590016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035955555555555556,
      "loss": 2.6752,
      "theoretical_loss": 3.4120546804795335,
      "tokens_seen": 2133655552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035953535353535357,
      "loss": 2.6471,
      "theoretical_loss": 3.412046053177339,
      "tokens_seen": 2133721088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035951515151515153,
      "loss": 2.6086,
      "theoretical_loss": 3.412037426214315,
      "tokens_seen": 2133786624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594949494949495,
      "loss": 2.5331,
      "theoretical_loss": 3.4120287995904386,
      "tokens_seen": 2133852160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594747474747475,
      "loss": 2.739,
      "theoretical_loss": 3.4120201733056854,
      "tokens_seen": 2133917696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035945454545454545,
      "loss": 2.6104,
      "theoretical_loss": 3.412011547360032,
      "tokens_seen": 2133983232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035943434343434347,
      "loss": 2.6294,
      "theoretical_loss": 3.412002921753455,
      "tokens_seen": 2134048768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594141414141414,
      "loss": 2.8417,
      "theoretical_loss": 3.41199429648593,
      "tokens_seen": 2134114304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593939393939394,
      "loss": 2.5628,
      "theoretical_loss": 3.4119856715574333,
      "tokens_seen": 2134179840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593737373737374,
      "loss": 2.6903,
      "theoretical_loss": 3.411977046967942,
      "tokens_seen": 2134245376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593535353535354,
      "loss": 2.7785,
      "theoretical_loss": 3.411968422717431,
      "tokens_seen": 2134310912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035933333333333336,
      "loss": 2.5786,
      "theoretical_loss": 3.411959798805878,
      "tokens_seen": 2134376448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593131313131313,
      "loss": 2.5277,
      "theoretical_loss": 3.4119511752332583,
      "tokens_seen": 2134441984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035929292929292927,
      "loss": 2.7237,
      "theoretical_loss": 3.4119425519995485,
      "tokens_seen": 2134507520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592727272727273,
      "loss": 2.6926,
      "theoretical_loss": 3.411933929104725,
      "tokens_seen": 2134573056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592525252525253,
      "loss": 2.6479,
      "theoretical_loss": 3.411925306548764,
      "tokens_seen": 2134638592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035923232323232325,
      "loss": 2.6249,
      "theoretical_loss": 3.4119166843316413,
      "tokens_seen": 2134704128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592121212121212,
      "loss": 2.6359,
      "theoretical_loss": 3.4119080624533336,
      "tokens_seen": 2134769664
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035919191919191916,
      "loss": 2.7272,
      "theoretical_loss": 3.411899440913818,
      "tokens_seen": 2134835200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035917171717171723,
      "loss": 2.7252,
      "theoretical_loss": 3.4118908197130695,
      "tokens_seen": 2134900736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591515151515152,
      "loss": 2.7684,
      "theoretical_loss": 3.411882198851065,
      "tokens_seen": 2134966272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035913131313131314,
      "loss": 2.5349,
      "theoretical_loss": 3.4118735783277803,
      "tokens_seen": 2135031808
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1176552,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6072494983673096,
      "objective/train/theoretical_loss": 3.411864958143192,
      "objective/train/tokens_used": 493956576,
      "theoretical_loss": 3.411864958143192,
      "tokens_seen": 2135097344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591111111111111,
      "loss": 2.4361,
      "theoretical_loss": 3.411864958143192,
      "tokens_seen": 2135097344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035909090909090906,
      "loss": 2.639,
      "theoretical_loss": 3.4118563382972766,
      "tokens_seen": 2135162880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590707070707071,
      "loss": 2.608,
      "theoretical_loss": 3.4118477187900105,
      "tokens_seen": 2135228416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590505050505051,
      "loss": 2.6096,
      "theoretical_loss": 3.4118390996213694,
      "tokens_seen": 2135293952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035903030303030304,
      "loss": 2.6053,
      "theoretical_loss": 3.41183048079133,
      "tokens_seen": 2135359488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359010101010101,
      "loss": 2.5846,
      "theoretical_loss": 3.4118218622998686,
      "tokens_seen": 2135425024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358989898989899,
      "loss": 2.4498,
      "theoretical_loss": 3.4118132441469617,
      "tokens_seen": 2135490560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358969696969697,
      "loss": 2.7093,
      "theoretical_loss": 3.411804626332585,
      "tokens_seen": 2135556096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035894949494949497,
      "loss": 2.5592,
      "theoretical_loss": 3.4117960088567156,
      "tokens_seen": 2135621632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035892929292929293,
      "loss": 2.6863,
      "theoretical_loss": 3.4117873917193293,
      "tokens_seen": 2135687168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589090909090909,
      "loss": 2.6119,
      "theoretical_loss": 3.4117787749204025,
      "tokens_seen": 2135752704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588888888888889,
      "loss": 2.8139,
      "theoretical_loss": 3.4117701584599116,
      "tokens_seen": 2135818240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588686868686869,
      "loss": 2.6701,
      "theoretical_loss": 3.4117615423378327,
      "tokens_seen": 2135883776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035884848484848486,
      "loss": 2.6039,
      "theoretical_loss": 3.4117529265541426,
      "tokens_seen": 2135949312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588282828282828,
      "loss": 2.5645,
      "theoretical_loss": 3.4117443111088175,
      "tokens_seen": 2136014848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035880808080808083,
      "loss": 2.7743,
      "theoretical_loss": 3.411735696001833,
      "tokens_seen": 2136080384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587878787878788,
      "loss": 2.654,
      "theoretical_loss": 3.411727081233166,
      "tokens_seen": 2136145920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587676767676768,
      "loss": 2.6771,
      "theoretical_loss": 3.4117184668027933,
      "tokens_seen": 2136211456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035874747474747476,
      "loss": 2.7149,
      "theoretical_loss": 3.41170985271069,
      "tokens_seen": 2136276992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587272727272727,
      "loss": 2.7189,
      "theoretical_loss": 3.4117012389568337,
      "tokens_seen": 2136342528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587070707070707,
      "loss": 2.5876,
      "theoretical_loss": 3.4116926255412006,
      "tokens_seen": 2136408064
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586868686868687,
      "loss": 2.6411,
      "theoretical_loss": 3.411684012463766,
      "tokens_seen": 2136473600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586666666666667,
      "loss": 2.5305,
      "theoretical_loss": 3.4116753997245075,
      "tokens_seen": 2136539136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035864646464646465,
      "loss": 2.5586,
      "theoretical_loss": 3.4116667873234006,
      "tokens_seen": 2136604672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586262626262626,
      "loss": 2.6443,
      "theoretical_loss": 3.4116581752604223,
      "tokens_seen": 2136670208
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1177415,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7517342567443848,
      "objective/train/theoretical_loss": 3.4116495635355486,
      "objective/train/tokens_used": 495594976,
      "theoretical_loss": 3.4116495635355486,
      "tokens_seen": 2136735744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586060606060606,
      "loss": 2.6526,
      "theoretical_loss": 3.4116495635355486,
      "tokens_seen": 2136735744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585858585858586,
      "loss": 2.8409,
      "theoretical_loss": 3.411640952148755,
      "tokens_seen": 2136801280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585656565656566,
      "loss": 2.6958,
      "theoretical_loss": 3.4116323411000193,
      "tokens_seen": 2136866816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035854545454545454,
      "loss": 2.6911,
      "theoretical_loss": 3.4116237303893175,
      "tokens_seen": 2136932352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035852525252525255,
      "loss": 2.5204,
      "theoretical_loss": 3.4116151200166254,
      "tokens_seen": 2136997888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585050505050505,
      "loss": 2.4892,
      "theoretical_loss": 3.4116065099819197,
      "tokens_seen": 2137063424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035848484848484847,
      "loss": 2.7625,
      "theoretical_loss": 3.4115979002851775,
      "tokens_seen": 2137128960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584646464646465,
      "loss": 2.748,
      "theoretical_loss": 3.411589290926374,
      "tokens_seen": 2137194496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035844444444444444,
      "loss": 2.8012,
      "theoretical_loss": 3.4115806819054857,
      "tokens_seen": 2137260032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035842424242424245,
      "loss": 2.4286,
      "theoretical_loss": 3.411572073222489,
      "tokens_seen": 2137325568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584040404040404,
      "loss": 2.7613,
      "theoretical_loss": 3.411563464877361,
      "tokens_seen": 2137391104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035838383838383836,
      "loss": 2.521,
      "theoretical_loss": 3.4115548568700778,
      "tokens_seen": 2137456640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035836363636363637,
      "loss": 2.5064,
      "theoretical_loss": 3.4115462492006157,
      "tokens_seen": 2137522176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583434343434344,
      "loss": 2.687,
      "theoretical_loss": 3.411537641868951,
      "tokens_seen": 2137587712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035832323232323234,
      "loss": 2.6783,
      "theoretical_loss": 3.41152903487506,
      "tokens_seen": 2137653248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583030303030303,
      "loss": 2.6258,
      "theoretical_loss": 3.411520428218919,
      "tokens_seen": 2137718784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035828282828282825,
      "loss": 2.4896,
      "theoretical_loss": 3.411511821900505,
      "tokens_seen": 2137784320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035826262626262626,
      "loss": 2.8056,
      "theoretical_loss": 3.4115032159197938,
      "tokens_seen": 2137849856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582424242424243,
      "loss": 2.712,
      "theoretical_loss": 3.411494610276762,
      "tokens_seen": 2137915392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035822222222222223,
      "loss": 2.7067,
      "theoretical_loss": 3.4114860049713855,
      "tokens_seen": 2137980928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582020202020202,
      "loss": 2.8186,
      "theoretical_loss": 3.4114774000036414,
      "tokens_seen": 2138046464
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035818181818181815,
      "loss": 2.7648,
      "theoretical_loss": 3.4114687953735063,
      "tokens_seen": 2138112000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581616161616162,
      "loss": 2.7007,
      "theoretical_loss": 3.411460191080956,
      "tokens_seen": 2138177536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035814141414141417,
      "loss": 2.6543,
      "theoretical_loss": 3.411451587125967,
      "tokens_seen": 2138243072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581212121212121,
      "loss": 2.6273,
      "theoretical_loss": 3.4114429835085156,
      "tokens_seen": 2138308608
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1177948,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.514267921447754,
      "objective/train/theoretical_loss": 3.411434380228579,
      "objective/train/tokens_used": 497233376,
      "theoretical_loss": 3.411434380228579,
      "tokens_seen": 2138374144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581010101010101,
      "loss": 2.5898,
      "theoretical_loss": 3.411434380228579,
      "tokens_seen": 2138374144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035808080808080804,
      "loss": 2.7825,
      "theoretical_loss": 3.4114257772861327,
      "tokens_seen": 2138439680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580606060606061,
      "loss": 2.7285,
      "theoretical_loss": 3.4114171746811537,
      "tokens_seen": 2138505216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035804040404040406,
      "loss": 2.5128,
      "theoretical_loss": 3.411408572413618,
      "tokens_seen": 2138570752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358020202020202,
      "loss": 2.5706,
      "theoretical_loss": 3.4113999704835023,
      "tokens_seen": 2138636288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358,
      "loss": 2.8487,
      "theoretical_loss": 3.411391368890783,
      "tokens_seen": 2138701824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357979797979798,
      "loss": 2.7602,
      "theoretical_loss": 3.4113827676354362,
      "tokens_seen": 2138767360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357959595959596,
      "loss": 2.5372,
      "theoretical_loss": 3.4113741667174384,
      "tokens_seen": 2138832896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035793939393939395,
      "loss": 2.7173,
      "theoretical_loss": 3.411365566136767,
      "tokens_seen": 2138898432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579191919191919,
      "loss": 2.5033,
      "theoretical_loss": 3.411356965893397,
      "tokens_seen": 2138963968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035789898989898987,
      "loss": 2.7777,
      "theoretical_loss": 3.4113483659873056,
      "tokens_seen": 2139029504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578787878787879,
      "loss": 2.8195,
      "theoretical_loss": 3.4113397664184695,
      "tokens_seen": 2139095040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578585858585859,
      "loss": 2.4594,
      "theoretical_loss": 3.4113311671868645,
      "tokens_seen": 2139160576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035783838383838385,
      "loss": 2.7956,
      "theoretical_loss": 3.4113225682924675,
      "tokens_seen": 2139226112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578181818181818,
      "loss": 2.8989,
      "theoretical_loss": 3.4113139697352546,
      "tokens_seen": 2139291648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577979797979798,
      "loss": 2.5859,
      "theoretical_loss": 3.411305371515202,
      "tokens_seen": 2139357184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035777777777777777,
      "loss": 2.6077,
      "theoretical_loss": 3.4112967736322872,
      "tokens_seen": 2139422720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577575757575758,
      "loss": 2.5769,
      "theoretical_loss": 3.411288176086486,
      "tokens_seen": 2139488256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035773737373737374,
      "loss": 2.6175,
      "theoretical_loss": 3.4112795788777746,
      "tokens_seen": 2139553792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577171717171717,
      "loss": 2.4836,
      "theoretical_loss": 3.41127098200613,
      "tokens_seen": 2139619328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576969696969697,
      "loss": 2.6395,
      "theoretical_loss": 3.4112623854715283,
      "tokens_seen": 2139684864
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576767676767677,
      "loss": 2.7262,
      "theoretical_loss": 3.411253789273946,
      "tokens_seen": 2139750400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576565656565657,
      "loss": 2.6886,
      "theoretical_loss": 3.41124519341336,
      "tokens_seen": 2139815936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035763636363636363,
      "loss": 2.6317,
      "theoretical_loss": 3.4112365978897463,
      "tokens_seen": 2139881472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035761616161616164,
      "loss": 2.5842,
      "theoretical_loss": 3.411228002703081,
      "tokens_seen": 2139947008
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1178547,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.098909378051758,
      "objective/train/theoretical_loss": 3.4112194078533413,
      "objective/train/tokens_used": 498871776,
      "theoretical_loss": 3.4112194078533413,
      "tokens_seen": 2140012544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575959595959596,
      "loss": 2.8142,
      "theoretical_loss": 3.4112194078533413,
      "tokens_seen": 2140012544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575757575757576,
      "loss": 2.6792,
      "theoretical_loss": 3.4112108133405035,
      "tokens_seen": 2140078080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035755555555555557,
      "loss": 2.9019,
      "theoretical_loss": 3.4112022191645437,
      "tokens_seen": 2140143616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575353535353535,
      "loss": 2.6421,
      "theoretical_loss": 3.4111936253254393,
      "tokens_seen": 2140209152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035751515151515153,
      "loss": 2.5492,
      "theoretical_loss": 3.4111850318231656,
      "tokens_seen": 2140274688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574949494949495,
      "loss": 2.6782,
      "theoretical_loss": 3.4111764386576997,
      "tokens_seen": 2140340224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574747474747475,
      "loss": 2.6779,
      "theoretical_loss": 3.411167845829018,
      "tokens_seen": 2140405760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035745454545454546,
      "loss": 2.6727,
      "theoretical_loss": 3.4111592533370976,
      "tokens_seen": 2140471296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035743434343434347,
      "loss": 2.4204,
      "theoretical_loss": 3.411150661181914,
      "tokens_seen": 2140536832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035741414141414143,
      "loss": 2.8483,
      "theoretical_loss": 3.4111420693634438,
      "tokens_seen": 2140602368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573939393939394,
      "loss": 2.5644,
      "theoretical_loss": 3.4111334778816644,
      "tokens_seen": 2140667904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573737373737374,
      "loss": 2.618,
      "theoretical_loss": 3.411124886736551,
      "tokens_seen": 2140733440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035735353535353535,
      "loss": 2.5871,
      "theoretical_loss": 3.4111162959280814,
      "tokens_seen": 2140798976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035733333333333336,
      "loss": 2.865,
      "theoretical_loss": 3.4111077054562315,
      "tokens_seen": 2140864512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573131313131313,
      "loss": 2.8388,
      "theoretical_loss": 3.4110991153209778,
      "tokens_seen": 2140930048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572929292929293,
      "loss": 2.8237,
      "theoretical_loss": 3.4110905255222965,
      "tokens_seen": 2140995584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572727272727273,
      "loss": 2.5144,
      "theoretical_loss": 3.4110819360601647,
      "tokens_seen": 2141061120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572525252525253,
      "loss": 2.5991,
      "theoretical_loss": 3.4110733469345584,
      "tokens_seen": 2141126656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035723232323232326,
      "loss": 2.4743,
      "theoretical_loss": 3.4110647581454545,
      "tokens_seen": 2141192192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572121212121212,
      "loss": 2.6059,
      "theoretical_loss": 3.4110561696928294,
      "tokens_seen": 2141257728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035719191919191917,
      "loss": 2.6451,
      "theoretical_loss": 3.4110475815766597,
      "tokens_seen": 2141323264
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571717171717172,
      "loss": 2.6165,
      "theoretical_loss": 3.4110389937969217,
      "tokens_seen": 2141388800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571515151515152,
      "loss": 2.5672,
      "theoretical_loss": 3.411030406353592,
      "tokens_seen": 2141454336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035713131313131315,
      "loss": 2.6528,
      "theoretical_loss": 3.4110218192466473,
      "tokens_seen": 2141519872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571111111111111,
      "loss": 2.8488,
      "theoretical_loss": 3.411013232476064,
      "tokens_seen": 2141585408
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1179742,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.639477014541626,
      "objective/train/theoretical_loss": 3.4110046460418184,
      "objective/train/tokens_used": 500510176,
      "theoretical_loss": 3.4110046460418184,
      "tokens_seen": 2141650944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035709090909090906,
      "loss": 2.6017,
      "theoretical_loss": 3.4110046460418184,
      "tokens_seen": 2141650944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035707070707070713,
      "loss": 2.7004,
      "theoretical_loss": 3.410996059943887,
      "tokens_seen": 2141716480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570505050505051,
      "loss": 2.6781,
      "theoretical_loss": 3.4109874741822477,
      "tokens_seen": 2141782016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035703030303030304,
      "loss": 2.8444,
      "theoretical_loss": 3.4109788887568753,
      "tokens_seen": 2141847552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357010101010101,
      "loss": 2.8455,
      "theoretical_loss": 3.410970303667747,
      "tokens_seen": 2141913088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035698989898989896,
      "loss": 2.8306,
      "theoretical_loss": 3.410961718914839,
      "tokens_seen": 2141978624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000356969696969697,
      "loss": 2.7306,
      "theoretical_loss": 3.4109531344981283,
      "tokens_seen": 2142044160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000356949494949495,
      "loss": 2.6625,
      "theoretical_loss": 3.4109445504175917,
      "tokens_seen": 2142109696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035692929292929293,
      "loss": 2.5678,
      "theoretical_loss": 3.4109359666732053,
      "tokens_seen": 2142175232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569090909090909,
      "loss": 2.6553,
      "theoretical_loss": 3.4109273832649456,
      "tokens_seen": 2142240768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035688888888888885,
      "loss": 2.6125,
      "theoretical_loss": 3.410918800192789,
      "tokens_seen": 2142306304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568686868686869,
      "loss": 2.6823,
      "theoretical_loss": 3.4109102174567125,
      "tokens_seen": 2142371840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035684848484848487,
      "loss": 2.6649,
      "theoretical_loss": 3.4109016350566934,
      "tokens_seen": 2142437376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568282828282828,
      "loss": 2.7469,
      "theoretical_loss": 3.410893052992706,
      "tokens_seen": 2142502912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568080808080808,
      "loss": 2.5226,
      "theoretical_loss": 3.410884471264729,
      "tokens_seen": 2142568448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567878787878788,
      "loss": 2.3986,
      "theoretical_loss": 3.410875889872738,
      "tokens_seen": 2142633984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567676767676768,
      "loss": 2.6273,
      "theoretical_loss": 3.41086730881671,
      "tokens_seen": 2142699520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035674747474747476,
      "loss": 2.5577,
      "theoretical_loss": 3.4108587280966214,
      "tokens_seen": 2142765056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567272727272727,
      "loss": 2.5092,
      "theoretical_loss": 3.4108501477124484,
      "tokens_seen": 2142830592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567070707070707,
      "loss": 2.4898,
      "theoretical_loss": 3.4108415676641677,
      "tokens_seen": 2142896128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566868686868687,
      "loss": 2.7255,
      "theoretical_loss": 3.4108329879517565,
      "tokens_seen": 2142961664
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566666666666667,
      "loss": 2.6781,
      "theoretical_loss": 3.4108244085751904,
      "tokens_seen": 2143027200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035664646464646466,
      "loss": 2.6827,
      "theoretical_loss": 3.4108158295344473,
      "tokens_seen": 2143092736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566262626262626,
      "loss": 2.7293,
      "theoretical_loss": 3.4108072508295026,
      "tokens_seen": 2143158272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566060606060606,
      "loss": 2.4671,
      "theoretical_loss": 3.4107986724603334,
      "tokens_seen": 2143223808
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1180436,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.763584613800049,
      "objective/train/theoretical_loss": 3.410790094426916,
      "objective/train/tokens_used": 502148576,
      "theoretical_loss": 3.410790094426916,
      "tokens_seen": 2143289344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565858585858586,
      "loss": 2.6251,
      "theoretical_loss": 3.410790094426916,
      "tokens_seen": 2143289344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565656565656566,
      "loss": 2.7291,
      "theoretical_loss": 3.4107815167292275,
      "tokens_seen": 2143354880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035654545454545455,
      "loss": 2.4953,
      "theoretical_loss": 3.410772939367244,
      "tokens_seen": 2143420416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565252525252525,
      "loss": 2.654,
      "theoretical_loss": 3.4107643623409425,
      "tokens_seen": 2143485952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565050505050505,
      "loss": 2.6565,
      "theoretical_loss": 3.410755785650299,
      "tokens_seen": 2143551488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564848484848485,
      "loss": 2.6607,
      "theoretical_loss": 3.410747209295291,
      "tokens_seen": 2143617024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564646464646465,
      "loss": 2.7669,
      "theoretical_loss": 3.410738633275894,
      "tokens_seen": 2143682560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035644444444444444,
      "loss": 2.7056,
      "theoretical_loss": 3.4107300575920854,
      "tokens_seen": 2143748096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035642424242424245,
      "loss": 2.8577,
      "theoretical_loss": 3.410721482243842,
      "tokens_seen": 2143813632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564040404040404,
      "loss": 2.6011,
      "theoretical_loss": 3.4107129072311393,
      "tokens_seen": 2143879168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035638383838383837,
      "loss": 2.4695,
      "theoretical_loss": 3.410704332553955,
      "tokens_seen": 2143944704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563636363636364,
      "loss": 2.62,
      "theoretical_loss": 3.410695758212266,
      "tokens_seen": 2144010240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035634343434343433,
      "loss": 2.5418,
      "theoretical_loss": 3.410687184206047,
      "tokens_seen": 2144075776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035632323232323234,
      "loss": 2.6616,
      "theoretical_loss": 3.4106786105352773,
      "tokens_seen": 2144141312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563030303030303,
      "loss": 2.6573,
      "theoretical_loss": 3.410670037199931,
      "tokens_seen": 2144206848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035628282828282826,
      "loss": 2.7227,
      "theoretical_loss": 3.410661464199986,
      "tokens_seen": 2144272384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035626262626262627,
      "loss": 2.9286,
      "theoretical_loss": 3.4106528915354195,
      "tokens_seen": 2144337920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562424242424243,
      "loss": 2.7204,
      "theoretical_loss": 3.410644319206207,
      "tokens_seen": 2144403456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035622222222222224,
      "loss": 2.7795,
      "theoretical_loss": 3.410635747212325,
      "tokens_seen": 2144468992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562020202020202,
      "loss": 2.6493,
      "theoretical_loss": 3.4106271755537514,
      "tokens_seen": 2144534528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035618181818181815,
      "loss": 2.7353,
      "theoretical_loss": 3.4106186042304616,
      "tokens_seen": 2144600064
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035616161616161616,
      "loss": 2.6767,
      "theoretical_loss": 3.410610033242433,
      "tokens_seen": 2144665600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561414141414142,
      "loss": 2.6527,
      "theoretical_loss": 3.410601462589642,
      "tokens_seen": 2144731136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035612121212121213,
      "loss": 2.5647,
      "theoretical_loss": 3.4105928922720654,
      "tokens_seen": 2144796672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561010101010101,
      "loss": 2.6571,
      "theoretical_loss": 3.4105843222896794,
      "tokens_seen": 2144862208
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 1181645,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3639755249023438,
      "objective/train/theoretical_loss": 3.410575752642461,
      "objective/train/tokens_used": 503786976,
      "theoretical_loss": 3.410575752642461,
      "tokens_seen": 2144927744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035608080808080804,
      "loss": 2.7949,
      "theoretical_loss": 3.410575752642461,
      "tokens_seen": 2144927744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003560606060606061,
      "loss": 2.5403,
      "theoretical_loss": 3.4105671833303868,
      "tokens_seen": 2144993280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035604040404040407,
      "loss": 2.71,
      "theoretical_loss": 3.4105586143534334,
      "tokens_seen": 2145058816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000356020202020202,
      "loss": 2.5801,
      "theoretical_loss": 3.4105500457115774,
      "tokens_seen": 2145124352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000356,
      "loss": 2.5069,
      "theoretical_loss": 3.410541477404796,
      "tokens_seen": 2145189888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355979797979798,
      "loss": 2.815,
      "theoretical_loss": 3.4105329094330648,
      "tokens_seen": 2145255424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355959595959596,
      "loss": 2.5818,
      "theoretical_loss": 3.4105243417963615,
      "tokens_seen": 2145320960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035593939393939396,
      "loss": 2.4357,
      "theoretical_loss": 3.4105157744946624,
      "tokens_seen": 2145386496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003559191919191919,
      "loss": 2.6195,
      "theoretical_loss": 3.4105072075279432,
      "tokens_seen": 2145452032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035589898989898987,
      "loss": 2.7188,
      "theoretical_loss": 3.410498640896183,
      "tokens_seen": 2145517568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035587878787878794,
      "loss": 2.6192,
      "theoretical_loss": 3.410490074599356,
      "tokens_seen": 2145583104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558585858585859,
      "loss": 2.4677,
      "theoretical_loss": 3.4104815086374396,
      "tokens_seen": 2145648640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035583838383838385,
      "loss": 2.7138,
      "theoretical_loss": 3.4104729430104115,
      "tokens_seen": 2145714176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558181818181818,
      "loss": 2.8676,
      "theoretical_loss": 3.410464377718247,
      "tokens_seen": 2145779712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035579797979797977,
      "loss": 2.7443,
      "theoretical_loss": 3.4104558127609232,
      "tokens_seen": 2145845248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035577777777777783,
      "loss": 2.9617,
      "theoretical_loss": 3.4104472481384174,
      "tokens_seen": 2145910784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557575757575758,
      "loss": 2.5756,
      "theoretical_loss": 3.410438683850706,
      "tokens_seen": 2145976320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035573737373737374,
      "loss": 2.8851,
      "theoretical_loss": 3.4104301198977653,
      "tokens_seen": 2146041856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557171717171717,
      "loss": 2.8399,
      "theoretical_loss": 3.410421556279572,
      "tokens_seen": 2146107392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556969696969697,
      "loss": 2.5007,
      "theoretical_loss": 3.410412992996103,
      "tokens_seen": 2146172928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556767676767677,
      "loss": 2.5867,
      "theoretical_loss": 3.4104044300473353,
      "tokens_seen": 2146238464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556565656565657,
      "loss": 2.6169,
      "theoretical_loss": 3.410395867433245,
      "tokens_seen": 2146304000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035563636363636364,
      "loss": 2.5714,
      "theoretical_loss": 3.4103873051538094,
      "tokens_seen": 2146369536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556161616161616,
      "loss": 2.6238,
      "theoretical_loss": 3.410378743209005,
      "tokens_seen": 2146435072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555959595959596,
      "loss": 2.685,
      "theoretical_loss": 3.4103701815988083,
      "tokens_seen": 2146500608
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1182191,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.627549886703491,
      "objective/train/theoretical_loss": 3.4103616203231955,
      "objective/train/tokens_used": 505425376,
      "theoretical_loss": 3.4103616203231955,
      "tokens_seen": 2146566144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555757575757576,
      "loss": 2.7703,
      "theoretical_loss": 3.4103616203231955,
      "tokens_seen": 2146566144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035555555555555557,
      "loss": 2.5253,
      "theoretical_loss": 3.4103530593821447,
      "tokens_seen": 2146631680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035553535353535353,
      "loss": 2.9456,
      "theoretical_loss": 3.4103444987756317,
      "tokens_seen": 2146697216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035551515151515154,
      "loss": 2.5637,
      "theoretical_loss": 3.4103359385036334,
      "tokens_seen": 2146762752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554949494949495,
      "loss": 2.5552,
      "theoretical_loss": 3.4103273785661266,
      "tokens_seen": 2146828288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554747474747475,
      "loss": 2.7312,
      "theoretical_loss": 3.4103188189630878,
      "tokens_seen": 2146893824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035545454545454547,
      "loss": 2.7148,
      "theoretical_loss": 3.4103102596944934,
      "tokens_seen": 2146959360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554343434343434,
      "loss": 2.9333,
      "theoretical_loss": 3.4103017007603205,
      "tokens_seen": 2147024896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035541414141414143,
      "loss": 2.5406,
      "theoretical_loss": 3.4102931421605467,
      "tokens_seen": 2147090432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553939393939394,
      "loss": 2.8978,
      "theoretical_loss": 3.410284583895147,
      "tokens_seen": 2147155968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553737373737374,
      "loss": 2.7002,
      "theoretical_loss": 3.4102760259641,
      "tokens_seen": 2147221504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035535353535353536,
      "loss": 2.7219,
      "theoretical_loss": 3.4102674683673806,
      "tokens_seen": 2147287040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553333333333333,
      "loss": 2.5009,
      "theoretical_loss": 3.4102589111049664,
      "tokens_seen": 2147352576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553131313131313,
      "loss": 2.491,
      "theoretical_loss": 3.4102503541768345,
      "tokens_seen": 2147418112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552929292929293,
      "loss": 2.7664,
      "theoretical_loss": 3.4102417975829615,
      "tokens_seen": 2147483648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552727272727273,
      "loss": 2.6139,
      "theoretical_loss": 3.4102332413233234,
      "tokens_seen": 2147549184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035525252525252525,
      "loss": 2.7421,
      "theoretical_loss": 3.410224685397898,
      "tokens_seen": 2147614720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035523232323232326,
      "loss": 2.6755,
      "theoretical_loss": 3.4102161298066607,
      "tokens_seen": 2147680256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552121212121212,
      "loss": 2.634,
      "theoretical_loss": 3.41020757454959,
      "tokens_seen": 2147745792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551919191919192,
      "loss": 2.6799,
      "theoretical_loss": 3.410199019626661,
      "tokens_seen": 2147811328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551717171717172,
      "loss": 2.7408,
      "theoretical_loss": 3.410190465037852,
      "tokens_seen": 2147876864
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035515151515151514,
      "loss": 2.5752,
      "theoretical_loss": 3.410181910783138,
      "tokens_seen": 2147942400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035513131313131315,
      "loss": 2.582,
      "theoretical_loss": 3.4101733568624972,
      "tokens_seen": 2148007936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551111111111111,
      "loss": 2.6301,
      "theoretical_loss": 3.4101648032759053,
      "tokens_seen": 2148073472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035509090909090907,
      "loss": 2.8678,
      "theoretical_loss": 3.4101562500233404,
      "tokens_seen": 2148139008
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1183255,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5413014888763428,
      "objective/train/theoretical_loss": 3.4101476971047777,
      "objective/train/tokens_used": 507063776,
      "theoretical_loss": 3.4101476971047777,
      "tokens_seen": 2148204544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550707070707071,
      "loss": 2.5406,
      "theoretical_loss": 3.4101476971047777,
      "tokens_seen": 2148204544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550505050505051,
      "loss": 2.511,
      "theoretical_loss": 3.4101391445201954,
      "tokens_seen": 2148270080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035503030303030305,
      "loss": 2.6562,
      "theoretical_loss": 3.410130592269569,
      "tokens_seen": 2148335616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355010101010101,
      "loss": 2.7914,
      "theoretical_loss": 3.410122040352877,
      "tokens_seen": 2148401152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035498989898989896,
      "loss": 2.6844,
      "theoretical_loss": 3.4101134887700937,
      "tokens_seen": 2148466688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035496969696969697,
      "loss": 2.5422,
      "theoretical_loss": 3.410104937521198,
      "tokens_seen": 2148532224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354949494949495,
      "loss": 2.7876,
      "theoretical_loss": 3.410096386606166,
      "tokens_seen": 2148597760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035492929292929294,
      "loss": 2.7034,
      "theoretical_loss": 3.4100878360249736,
      "tokens_seen": 2148663296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549090909090909,
      "loss": 2.4995,
      "theoretical_loss": 3.4100792857775994,
      "tokens_seen": 2148728832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035488888888888885,
      "loss": 2.6483,
      "theoretical_loss": 3.4100707358640188,
      "tokens_seen": 2148794368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548686868686869,
      "loss": 2.775,
      "theoretical_loss": 3.4100621862842084,
      "tokens_seen": 2148859904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548484848484849,
      "loss": 2.5338,
      "theoretical_loss": 3.4100536370381462,
      "tokens_seen": 2148925440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035482828282828283,
      "loss": 2.9031,
      "theoretical_loss": 3.410045088125808,
      "tokens_seen": 2148990976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548080808080808,
      "loss": 2.5603,
      "theoretical_loss": 3.4100365395471712,
      "tokens_seen": 2149056512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035478787878787875,
      "loss": 2.5336,
      "theoretical_loss": 3.4100279913022122,
      "tokens_seen": 2149122048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547676767676768,
      "loss": 2.7277,
      "theoretical_loss": 3.410019443390908,
      "tokens_seen": 2149187584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035474747474747477,
      "loss": 2.7138,
      "theoretical_loss": 3.410010895813236,
      "tokens_seen": 2149253120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547272727272727,
      "loss": 2.5253,
      "theoretical_loss": 3.4100023485691713,
      "tokens_seen": 2149318656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547070707070707,
      "loss": 2.7073,
      "theoretical_loss": 3.4099938016586924,
      "tokens_seen": 2149384192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546868686868687,
      "loss": 2.5695,
      "theoretical_loss": 3.409985255081775,
      "tokens_seen": 2149449728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546666666666667,
      "loss": 2.5718,
      "theoretical_loss": 3.4099767088383968,
      "tokens_seen": 2149515264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035464646464646466,
      "loss": 2.6058,
      "theoretical_loss": 3.409968162928534,
      "tokens_seen": 2149580800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546262626262626,
      "loss": 2.6631,
      "theoretical_loss": 3.409959617352164,
      "tokens_seen": 2149646336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546060606060606,
      "loss": 2.7089,
      "theoretical_loss": 3.4099510721092625,
      "tokens_seen": 2149711872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545858585858586,
      "loss": 2.356,
      "theoretical_loss": 3.4099425271998074,
      "tokens_seen": 2149777408
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1183881,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.526425361633301,
      "objective/train/theoretical_loss": 3.4099339826237753,
      "objective/train/tokens_used": 508702176,
      "theoretical_loss": 3.4099339826237753,
      "tokens_seen": 2149842944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545656565656566,
      "loss": 2.6612,
      "theoretical_loss": 3.4099339826237753,
      "tokens_seen": 2149842944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035454545454545455,
      "loss": 2.6441,
      "theoretical_loss": 3.409925438381143,
      "tokens_seen": 2149908480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545252525252525,
      "loss": 2.4833,
      "theoretical_loss": 3.4099168944718867,
      "tokens_seen": 2149974016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545050505050505,
      "loss": 2.6153,
      "theoretical_loss": 3.409908350895984,
      "tokens_seen": 2150039552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544848484848485,
      "loss": 2.5933,
      "theoretical_loss": 3.409899807653412,
      "tokens_seen": 2150105088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544646464646465,
      "loss": 2.4394,
      "theoretical_loss": 3.4098912647441466,
      "tokens_seen": 2150170624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035444444444444445,
      "loss": 2.4475,
      "theoretical_loss": 3.409882722168165,
      "tokens_seen": 2150236160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544242424242424,
      "loss": 2.7372,
      "theoretical_loss": 3.4098741799254446,
      "tokens_seen": 2150301696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544040404040404,
      "loss": 2.642,
      "theoretical_loss": 3.409865638015961,
      "tokens_seen": 2150367232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035438383838383837,
      "loss": 2.6953,
      "theoretical_loss": 3.4098570964396924,
      "tokens_seen": 2150432768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543636363636364,
      "loss": 2.8193,
      "theoretical_loss": 3.4098485551966147,
      "tokens_seen": 2150498304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035434343434343434,
      "loss": 2.6868,
      "theoretical_loss": 3.409840014286705,
      "tokens_seen": 2150563840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035432323232323235,
      "loss": 2.6162,
      "theoretical_loss": 3.409831473709941,
      "tokens_seen": 2150629376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543030303030303,
      "loss": 2.8,
      "theoretical_loss": 3.409822933466298,
      "tokens_seen": 2150694912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542828282828283,
      "loss": 2.6104,
      "theoretical_loss": 3.409814393555754,
      "tokens_seen": 2150760448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542626262626263,
      "loss": 2.7018,
      "theoretical_loss": 3.4098058539782854,
      "tokens_seen": 2150825984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035424242424242423,
      "loss": 2.5429,
      "theoretical_loss": 3.4097973147338694,
      "tokens_seen": 2150891520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035422222222222224,
      "loss": 2.4372,
      "theoretical_loss": 3.4097887758224825,
      "tokens_seen": 2150957056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542020202020202,
      "loss": 2.6098,
      "theoretical_loss": 3.4097802372441013,
      "tokens_seen": 2151022592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541818181818182,
      "loss": 2.6754,
      "theoretical_loss": 3.4097716989987035,
      "tokens_seen": 2151088128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035416161616161617,
      "loss": 2.4958,
      "theoretical_loss": 3.409763161086266,
      "tokens_seen": 2151153664
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541414141414142,
      "loss": 2.6487,
      "theoretical_loss": 3.4097546235067644,
      "tokens_seen": 2151219200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035412121212121214,
      "loss": 2.8716,
      "theoretical_loss": 3.409746086260177,
      "tokens_seen": 2151284736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541010101010101,
      "loss": 2.6091,
      "theoretical_loss": 3.409737549346479,
      "tokens_seen": 2151350272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540808080808081,
      "loss": 2.71,
      "theoretical_loss": 3.4097290127656494,
      "tokens_seen": 2151415808
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1184815,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6544196605682373,
      "objective/train/theoretical_loss": 3.409720476517664,
      "objective/train/tokens_used": 510340576,
      "theoretical_loss": 3.409720476517664,
      "tokens_seen": 2151481344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035406060606060606,
      "loss": 2.6166,
      "theoretical_loss": 3.409720476517664,
      "tokens_seen": 2151481344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035404040404040407,
      "loss": 2.7207,
      "theoretical_loss": 3.4097119406024996,
      "tokens_seen": 2151546880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035402020202020203,
      "loss": 2.5904,
      "theoretical_loss": 3.4097034050201334,
      "tokens_seen": 2151612416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354,
      "loss": 2.6907,
      "theoretical_loss": 3.4096948697705414,
      "tokens_seen": 2151677952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353979797979798,
      "loss": 2.6082,
      "theoretical_loss": 3.4096863348537014,
      "tokens_seen": 2151743488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353959595959596,
      "loss": 2.7023,
      "theoretical_loss": 3.4096778002695904,
      "tokens_seen": 2151809024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035393939393939396,
      "loss": 2.6989,
      "theoretical_loss": 3.409669266018185,
      "tokens_seen": 2151874560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003539191919191919,
      "loss": 2.5781,
      "theoretical_loss": 3.4096607320994616,
      "tokens_seen": 2151940096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538989898989899,
      "loss": 2.4732,
      "theoretical_loss": 3.409652198513398,
      "tokens_seen": 2152005632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538787878787879,
      "loss": 2.632,
      "theoretical_loss": 3.4096436652599706,
      "tokens_seen": 2152071168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538585858585859,
      "loss": 2.7633,
      "theoretical_loss": 3.4096351323391563,
      "tokens_seen": 2152136704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035383838383838386,
      "loss": 2.5427,
      "theoretical_loss": 3.409626599750932,
      "tokens_seen": 2152202240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538181818181818,
      "loss": 2.627,
      "theoretical_loss": 3.4096180674952747,
      "tokens_seen": 2152267776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035379797979797977,
      "loss": 2.4213,
      "theoretical_loss": 3.4096095355721614,
      "tokens_seen": 2152333312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035377777777777784,
      "loss": 2.4559,
      "theoretical_loss": 3.409601003981569,
      "tokens_seen": 2152398848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537575757575758,
      "loss": 2.4325,
      "theoretical_loss": 3.409592472723474,
      "tokens_seen": 2152464384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035373737373737375,
      "loss": 2.6978,
      "theoretical_loss": 3.4095839417978544,
      "tokens_seen": 2152529920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537171717171717,
      "loss": 2.6345,
      "theoretical_loss": 3.4095754112046857,
      "tokens_seen": 2152595456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035369696969696966,
      "loss": 2.704,
      "theoretical_loss": 3.409566880943945,
      "tokens_seen": 2152660992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035367676767676773,
      "loss": 2.4888,
      "theoretical_loss": 3.409558351015611,
      "tokens_seen": 2152726528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536565656565657,
      "loss": 2.4636,
      "theoretical_loss": 3.4095498214196587,
      "tokens_seen": 2152792064
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035363636363636364,
      "loss": 2.5985,
      "theoretical_loss": 3.4095412921560655,
      "tokens_seen": 2152857600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536161616161616,
      "loss": 2.547,
      "theoretical_loss": 3.409532763224809,
      "tokens_seen": 2152923136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035359595959595956,
      "loss": 2.6535,
      "theoretical_loss": 3.4095242346258656,
      "tokens_seen": 2152988672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535757575757576,
      "loss": 2.6403,
      "theoretical_loss": 3.409515706359212,
      "tokens_seen": 2153054208
    },
    {
      "debugging/Self-BLEU-5": 0.39580322091234643,
      "debugging/distinct-1-grams": 0.797421315948946,
      "debugging/distinct-2-grams": 0.9612328752077172,
      "debugging/entropy-1-grams": 5.484895349701254,
      "debugging/entropy-2-grams": 6.169724221975992,
      "debugging/length": 483.25,
      "debugging/num_segments": 8,
      "epoch": 0.3,
      "objective/train/docs_used": 1185854,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8254241943359375,
      "objective/train/theoretical_loss": 3.4095071784248256,
      "objective/train/tokens_used": 511978976,
      "theoretical_loss": 3.4095071784248256,
      "tokens_seen": 2153119744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535555555555556,
      "loss": 2.5479,
      "theoretical_loss": 3.4095071784248256,
      "tokens_seen": 2153119744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035353535353535354,
      "loss": 2.6056,
      "theoretical_loss": 3.409498650822683,
      "tokens_seen": 2153185280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535151515151515,
      "loss": 2.6974,
      "theoretical_loss": 3.4094901235527613,
      "tokens_seen": 2153250816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534949494949495,
      "loss": 2.3753,
      "theoretical_loss": 3.409481596615038,
      "tokens_seen": 2153316352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534747474747475,
      "loss": 2.5776,
      "theoretical_loss": 3.4094730700094886,
      "tokens_seen": 2153381888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035345454545454547,
      "loss": 2.5597,
      "theoretical_loss": 3.4094645437360915,
      "tokens_seen": 2153447424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035343434343434343,
      "loss": 2.773,
      "theoretical_loss": 3.409456017794823,
      "tokens_seen": 2153512960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534141414141414,
      "loss": 2.6986,
      "theoretical_loss": 3.4094474921856603,
      "tokens_seen": 2153578496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533939393939394,
      "loss": 2.6894,
      "theoretical_loss": 3.40943896690858,
      "tokens_seen": 2153644032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533737373737374,
      "loss": 2.916,
      "theoretical_loss": 3.4094304419635595,
      "tokens_seen": 2153709568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035335353535353536,
      "loss": 2.6452,
      "theoretical_loss": 3.409421917350575,
      "tokens_seen": 2153775104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533333333333333,
      "loss": 2.6145,
      "theoretical_loss": 3.409413393069605,
      "tokens_seen": 2153840640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035331313131313133,
      "loss": 2.8686,
      "theoretical_loss": 3.4094048691206247,
      "tokens_seen": 2153906176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532929292929293,
      "loss": 2.8178,
      "theoretical_loss": 3.4093963455036125,
      "tokens_seen": 2153971712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532727272727273,
      "loss": 2.6879,
      "theoretical_loss": 3.4093878222185445,
      "tokens_seen": 2154037248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035325252525252526,
      "loss": 2.623,
      "theoretical_loss": 3.4093792992653977,
      "tokens_seen": 2154102784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532323232323232,
      "loss": 2.7944,
      "theoretical_loss": 3.409370776644149,
      "tokens_seen": 2154168320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532121212121212,
      "loss": 2.7522,
      "theoretical_loss": 3.409362254354776,
      "tokens_seen": 2154233856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531919191919192,
      "loss": 2.5473,
      "theoretical_loss": 3.409353732397255,
      "tokens_seen": 2154299392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531717171717172,
      "loss": 2.6862,
      "theoretical_loss": 3.409345210771564,
      "tokens_seen": 2154364928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035315151515151515,
      "loss": 2.6817,
      "theoretical_loss": 3.409336689477679,
      "tokens_seen": 2154430464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035313131313131316,
      "loss": 2.5026,
      "theoretical_loss": 3.409328168515577,
      "tokens_seen": 2154496000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531111111111111,
      "loss": 2.7158,
      "theoretical_loss": 3.4093196478852352,
      "tokens_seen": 2154561536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530909090909091,
      "loss": 2.5761,
      "theoretical_loss": 3.4093111275866312,
      "tokens_seen": 2154627072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530707070707071,
      "loss": 2.7904,
      "theoretical_loss": 3.4093026076197415,
      "tokens_seen": 2154692608
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1186317,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.778775691986084,
      "objective/train/theoretical_loss": 3.4092940879845424,
      "objective/train/tokens_used": 513617376,
      "theoretical_loss": 3.4092940879845424,
      "tokens_seen": 2154758144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035305050505050504,
      "loss": 2.5932,
      "theoretical_loss": 3.4092940879845424,
      "tokens_seen": 2154758144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035303030303030305,
      "loss": 2.6707,
      "theoretical_loss": 3.409285568681012,
      "tokens_seen": 2154823680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353010101010101,
      "loss": 2.5527,
      "theoretical_loss": 3.409277049709127,
      "tokens_seen": 2154889216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035298989898989897,
      "loss": 2.5178,
      "theoretical_loss": 3.409268531068864,
      "tokens_seen": 2154954752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352969696969697,
      "loss": 2.6367,
      "theoretical_loss": 3.4092600127602006,
      "tokens_seen": 2155020288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352949494949495,
      "loss": 2.5147,
      "theoretical_loss": 3.409251494783113,
      "tokens_seen": 2155085824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035292929292929295,
      "loss": 2.6731,
      "theoretical_loss": 3.4092429771375787,
      "tokens_seen": 2155151360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003529090909090909,
      "loss": 2.5244,
      "theoretical_loss": 3.4092344598235753,
      "tokens_seen": 2155216896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035288888888888886,
      "loss": 2.6141,
      "theoretical_loss": 3.4092259428410783,
      "tokens_seen": 2155282432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035286868686868687,
      "loss": 2.6846,
      "theoretical_loss": 3.4092174261900663,
      "tokens_seen": 2155347968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528484848484849,
      "loss": 2.5751,
      "theoretical_loss": 3.409208909870516,
      "tokens_seen": 2155413504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035282828282828284,
      "loss": 2.625,
      "theoretical_loss": 3.4092003938824034,
      "tokens_seen": 2155479040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528080808080808,
      "loss": 2.5676,
      "theoretical_loss": 3.4091918782257062,
      "tokens_seen": 2155544576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035278787878787875,
      "loss": 2.5603,
      "theoretical_loss": 3.409183362900402,
      "tokens_seen": 2155610112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527676767676768,
      "loss": 2.6779,
      "theoretical_loss": 3.409174847906467,
      "tokens_seen": 2155675648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527474747474748,
      "loss": 2.597,
      "theoretical_loss": 3.409166333243878,
      "tokens_seen": 2155741184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035272727272727273,
      "loss": 2.6806,
      "theoretical_loss": 3.409157818912613,
      "tokens_seen": 2155806720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527070707070707,
      "loss": 2.6874,
      "theoretical_loss": 3.4091493049126487,
      "tokens_seen": 2155872256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035268686868686864,
      "loss": 2.6295,
      "theoretical_loss": 3.4091407912439617,
      "tokens_seen": 2155937792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526666666666667,
      "loss": 2.5543,
      "theoretical_loss": 3.4091322779065294,
      "tokens_seen": 2156003328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035264646464646467,
      "loss": 2.5731,
      "theoretical_loss": 3.409123764900329,
      "tokens_seen": 2156068864
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526262626262626,
      "loss": 2.6069,
      "theoretical_loss": 3.4091152522253374,
      "tokens_seen": 2156134400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526060606060606,
      "loss": 2.6663,
      "theoretical_loss": 3.409106739881531,
      "tokens_seen": 2156199936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035258585858585865,
      "loss": 2.4908,
      "theoretical_loss": 3.4090982278688875,
      "tokens_seen": 2156265472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525656565656566,
      "loss": 2.5468,
      "theoretical_loss": 3.4090897161873848,
      "tokens_seen": 2156331008
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1187470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.505795955657959,
      "objective/train/theoretical_loss": 3.409081204836998,
      "objective/train/tokens_used": 515255776,
      "theoretical_loss": 3.409081204836998,
      "tokens_seen": 2156396544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035254545454545456,
      "loss": 2.6541,
      "theoretical_loss": 3.409081204836998,
      "tokens_seen": 2156396544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525252525252525,
      "loss": 2.4961,
      "theoretical_loss": 3.4090726938177056,
      "tokens_seen": 2156462080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525050505050505,
      "loss": 2.595,
      "theoretical_loss": 3.409064183129484,
      "tokens_seen": 2156527616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035248484848484854,
      "loss": 2.6051,
      "theoretical_loss": 3.409055672772311,
      "tokens_seen": 2156593152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524646464646465,
      "loss": 2.6744,
      "theoretical_loss": 3.4090471627461625,
      "tokens_seen": 2156658688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035244444444444445,
      "loss": 2.4898,
      "theoretical_loss": 3.409038653051017,
      "tokens_seen": 2156724224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524242424242424,
      "loss": 2.6977,
      "theoretical_loss": 3.4090301436868504,
      "tokens_seen": 2156789760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524040404040404,
      "loss": 2.5083,
      "theoretical_loss": 3.40902163465364,
      "tokens_seen": 2156855296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035238383838383843,
      "loss": 2.6723,
      "theoretical_loss": 3.4090131259513634,
      "tokens_seen": 2156920832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523636363636364,
      "loss": 2.6959,
      "theoretical_loss": 3.4090046175799973,
      "tokens_seen": 2156986368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035234343434343434,
      "loss": 2.4903,
      "theoretical_loss": 3.408996109539519,
      "tokens_seen": 2157051904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523232323232323,
      "loss": 2.4713,
      "theoretical_loss": 3.408987601829905,
      "tokens_seen": 2157117440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523030303030303,
      "loss": 2.7848,
      "theoretical_loss": 3.4089790944511327,
      "tokens_seen": 2157182976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522828282828283,
      "loss": 2.8257,
      "theoretical_loss": 3.4089705874031795,
      "tokens_seen": 2157248512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522626262626263,
      "loss": 2.4483,
      "theoretical_loss": 3.408962080686022,
      "tokens_seen": 2157314048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035224242424242424,
      "loss": 2.8532,
      "theoretical_loss": 3.4089535742996375,
      "tokens_seen": 2157379584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035222222222222225,
      "loss": 2.693,
      "theoretical_loss": 3.4089450682440035,
      "tokens_seen": 2157445120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522020202020202,
      "loss": 2.6485,
      "theoretical_loss": 3.408936562519096,
      "tokens_seen": 2157510656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521818181818182,
      "loss": 2.6114,
      "theoretical_loss": 3.408928057124893,
      "tokens_seen": 2157576192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521616161616162,
      "loss": 2.567,
      "theoretical_loss": 3.4089195520613718,
      "tokens_seen": 2157641728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035214141414141413,
      "loss": 2.6924,
      "theoretical_loss": 3.408911047328509,
      "tokens_seen": 2157707264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035212121212121214,
      "loss": 2.4765,
      "theoretical_loss": 3.4089025429262816,
      "tokens_seen": 2157772800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521010101010101,
      "loss": 2.6886,
      "theoretical_loss": 3.408894038854667,
      "tokens_seen": 2157838336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003520808080808081,
      "loss": 2.5828,
      "theoretical_loss": 3.4088855351136425,
      "tokens_seen": 2157903872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035206060606060607,
      "loss": 2.507,
      "theoretical_loss": 3.4088770317031845,
      "tokens_seen": 2157969408
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1188127,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4245564937591553,
      "objective/train/theoretical_loss": 3.4088685286232705,
      "objective/train/tokens_used": 516894176,
      "theoretical_loss": 3.4088685286232705,
      "tokens_seen": 2158034944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352040404040404,
      "loss": 2.5853,
      "theoretical_loss": 3.4088685286232705,
      "tokens_seen": 2158034944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035202020202020203,
      "loss": 2.6205,
      "theoretical_loss": 3.408860025873878,
      "tokens_seen": 2158100480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352,
      "loss": 2.6775,
      "theoretical_loss": 3.4088515234549837,
      "tokens_seen": 2158166016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000351979797979798,
      "loss": 2.6753,
      "theoretical_loss": 3.4088430213665646,
      "tokens_seen": 2158231552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035195959595959596,
      "loss": 2.7198,
      "theoretical_loss": 3.408834519608598,
      "tokens_seen": 2158297088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035193939393939397,
      "loss": 2.6947,
      "theoretical_loss": 3.408826018181061,
      "tokens_seen": 2158362624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003519191919191919,
      "loss": 2.816,
      "theoretical_loss": 3.408817517083931,
      "tokens_seen": 2158428160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518989898989899,
      "loss": 2.4772,
      "theoretical_loss": 3.4088090163171847,
      "tokens_seen": 2158493696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518787878787879,
      "loss": 2.6107,
      "theoretical_loss": 3.4088005158807992,
      "tokens_seen": 2158559232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035185858585858585,
      "loss": 2.6709,
      "theoretical_loss": 3.408792015774752,
      "tokens_seen": 2158624768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035183838383838386,
      "loss": 2.3833,
      "theoretical_loss": 3.40878351599902,
      "tokens_seen": 2158690304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518181818181818,
      "loss": 2.6363,
      "theoretical_loss": 3.408775016553581,
      "tokens_seen": 2158755840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517979797979798,
      "loss": 2.389,
      "theoretical_loss": 3.408766517438411,
      "tokens_seen": 2158821376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517777777777778,
      "loss": 2.5093,
      "theoretical_loss": 3.408758018653488,
      "tokens_seen": 2158886912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517575757575758,
      "loss": 2.6983,
      "theoretical_loss": 3.4087495201987883,
      "tokens_seen": 2158952448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035173737373737376,
      "loss": 2.3098,
      "theoretical_loss": 3.40874102207429,
      "tokens_seen": 2159017984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517171717171717,
      "loss": 2.6428,
      "theoretical_loss": 3.4087325242799693,
      "tokens_seen": 2159083520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035169696969696967,
      "loss": 2.5138,
      "theoretical_loss": 3.4087240268158046,
      "tokens_seen": 2159149056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516767676767677,
      "loss": 2.6145,
      "theoretical_loss": 3.4087155296817717,
      "tokens_seen": 2159214592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516565656565657,
      "loss": 2.631,
      "theoretical_loss": 3.4087070328778486,
      "tokens_seen": 2159280128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035163636363636365,
      "loss": 2.6058,
      "theoretical_loss": 3.4086985364040125,
      "tokens_seen": 2159345664
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516161616161616,
      "loss": 2.3295,
      "theoretical_loss": 3.40869004026024,
      "tokens_seen": 2159411200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035159595959595956,
      "loss": 2.6824,
      "theoretical_loss": 3.408681544446509,
      "tokens_seen": 2159476736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515757575757576,
      "loss": 2.4669,
      "theoretical_loss": 3.4086730489627954,
      "tokens_seen": 2159542272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515555555555556,
      "loss": 2.8928,
      "theoretical_loss": 3.4086645538090776,
      "tokens_seen": 2159607808
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1189561,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.801823139190674,
      "objective/train/theoretical_loss": 3.4086560589853327,
      "objective/train/tokens_used": 518532576,
      "theoretical_loss": 3.4086560589853327,
      "tokens_seen": 2159673344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035153535353535354,
      "loss": 2.7104,
      "theoretical_loss": 3.4086560589853327,
      "tokens_seen": 2159673344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515151515151515,
      "loss": 2.534,
      "theoretical_loss": 3.4086475644915373,
      "tokens_seen": 2159738880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035149494949494945,
      "loss": 2.6664,
      "theoretical_loss": 3.4086390703276686,
      "tokens_seen": 2159804416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514747474747475,
      "loss": 2.5441,
      "theoretical_loss": 3.408630576493704,
      "tokens_seen": 2159869952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514545454545455,
      "loss": 2.495,
      "theoretical_loss": 3.4086220829896208,
      "tokens_seen": 2159935488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035143434343434343,
      "loss": 2.7479,
      "theoretical_loss": 3.408613589815396,
      "tokens_seen": 2160001024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514141414141414,
      "loss": 2.7925,
      "theoretical_loss": 3.408605096971007,
      "tokens_seen": 2160066560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513939393939394,
      "loss": 2.6846,
      "theoretical_loss": 3.4085966044564304,
      "tokens_seen": 2160132096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513737373737374,
      "loss": 2.3807,
      "theoretical_loss": 3.408588112271644,
      "tokens_seen": 2160197632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035135353535353537,
      "loss": 2.8208,
      "theoretical_loss": 3.4085796204166243,
      "tokens_seen": 2160263168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513333333333333,
      "loss": 2.5076,
      "theoretical_loss": 3.4085711288913494,
      "tokens_seen": 2160328704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513131313131313,
      "loss": 2.5498,
      "theoretical_loss": 3.4085626376957965,
      "tokens_seen": 2160394240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512929292929293,
      "loss": 2.5672,
      "theoretical_loss": 3.4085541468299416,
      "tokens_seen": 2160459776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512727272727273,
      "loss": 2.2968,
      "theoretical_loss": 3.408545656293763,
      "tokens_seen": 2160525312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035125252525252526,
      "loss": 2.6389,
      "theoretical_loss": 3.4085371660872377,
      "tokens_seen": 2160590848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512323232323232,
      "loss": 2.4206,
      "theoretical_loss": 3.4085286762103424,
      "tokens_seen": 2160656384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035121212121212123,
      "loss": 2.7318,
      "theoretical_loss": 3.4085201866630546,
      "tokens_seen": 2160721920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511919191919192,
      "loss": 2.6289,
      "theoretical_loss": 3.408511697445352,
      "tokens_seen": 2160787456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511717171717172,
      "loss": 2.6766,
      "theoretical_loss": 3.408503208557211,
      "tokens_seen": 2160852992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035115151515151515,
      "loss": 2.5316,
      "theoretical_loss": 3.4084947199986093,
      "tokens_seen": 2160918528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511313131313131,
      "loss": 2.7037,
      "theoretical_loss": 3.408486231769524,
      "tokens_seen": 2160984064
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511111111111111,
      "loss": 2.6177,
      "theoretical_loss": 3.4084777438699323,
      "tokens_seen": 2161049600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003510909090909091,
      "loss": 2.6657,
      "theoretical_loss": 3.408469256299812,
      "tokens_seen": 2161115136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003510707070707071,
      "loss": 2.5146,
      "theoretical_loss": 3.408460769059139,
      "tokens_seen": 2161180672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035105050505050505,
      "loss": 2.7131,
      "theoretical_loss": 3.4084522821478913,
      "tokens_seen": 2161246208
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 1190021,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.461451530456543,
      "objective/train/theoretical_loss": 3.408443795566046,
      "objective/train/tokens_used": 520170976,
      "theoretical_loss": 3.408443795566046,
      "tokens_seen": 2161311744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035103030303030306,
      "loss": 2.5318,
      "theoretical_loss": 3.408443795566046,
      "tokens_seen": 2161311744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000351010101010101,
      "loss": 2.7198,
      "theoretical_loss": 3.408435309313581,
      "tokens_seen": 2161377280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035098989898989897,
      "loss": 2.6733,
      "theoretical_loss": 3.4084268233904726,
      "tokens_seen": 2161442816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350969696969697,
      "loss": 2.5547,
      "theoretical_loss": 3.4084183377966983,
      "tokens_seen": 2161508352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035094949494949494,
      "loss": 2.6012,
      "theoretical_loss": 3.4084098525322357,
      "tokens_seen": 2161573888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035092929292929295,
      "loss": 2.7147,
      "theoretical_loss": 3.4084013675970617,
      "tokens_seen": 2161639424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509090909090909,
      "loss": 2.6112,
      "theoretical_loss": 3.408392882991153,
      "tokens_seen": 2161704960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035088888888888886,
      "loss": 2.3176,
      "theoretical_loss": 3.4083843987144884,
      "tokens_seen": 2161770496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508686868686869,
      "loss": 2.7915,
      "theoretical_loss": 3.408375914767044,
      "tokens_seen": 2161836032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508484848484849,
      "loss": 2.6781,
      "theoretical_loss": 3.408367431148797,
      "tokens_seen": 2161901568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035082828282828284,
      "loss": 2.7152,
      "theoretical_loss": 3.4083589478597247,
      "tokens_seen": 2161967104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508080808080808,
      "loss": 2.5664,
      "theoretical_loss": 3.4083504648998044,
      "tokens_seen": 2162032640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507878787878788,
      "loss": 2.5902,
      "theoretical_loss": 3.4083419822690137,
      "tokens_seen": 2162098176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035076767676767677,
      "loss": 2.5746,
      "theoretical_loss": 3.4083334999673296,
      "tokens_seen": 2162163712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507474747474748,
      "loss": 2.7214,
      "theoretical_loss": 3.4083250179947293,
      "tokens_seen": 2162229248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035072727272727274,
      "loss": 2.4374,
      "theoretical_loss": 3.4083165363511903,
      "tokens_seen": 2162294784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507070707070707,
      "loss": 2.5229,
      "theoretical_loss": 3.4083080550366898,
      "tokens_seen": 2162360320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506868686868687,
      "loss": 2.5961,
      "theoretical_loss": 3.4082995740512048,
      "tokens_seen": 2162425856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506666666666667,
      "loss": 2.6604,
      "theoretical_loss": 3.4082910933947126,
      "tokens_seen": 2162491392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035064646464646467,
      "loss": 2.4124,
      "theoretical_loss": 3.408282613067191,
      "tokens_seen": 2162556928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035062626262626263,
      "loss": 2.4736,
      "theoretical_loss": 3.408274133068617,
      "tokens_seen": 2162622464
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506060606060606,
      "loss": 2.6433,
      "theoretical_loss": 3.408265653398967,
      "tokens_seen": 2162688000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505858585858586,
      "loss": 2.5658,
      "theoretical_loss": 3.4082571740582197,
      "tokens_seen": 2162753536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505656565656566,
      "loss": 2.3968,
      "theoretical_loss": 3.4082486950463515,
      "tokens_seen": 2162819072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035054545454545457,
      "loss": 2.6287,
      "theoretical_loss": 3.4082402163633394,
      "tokens_seen": 2162884608
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1191353,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.469416618347168,
      "objective/train/theoretical_loss": 3.4082317380091616,
      "objective/train/tokens_used": 521809376,
      "theoretical_loss": 3.4082317380091616,
      "tokens_seen": 2162950144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505252525252525,
      "loss": 2.6831,
      "theoretical_loss": 3.4082317380091616,
      "tokens_seen": 2162950144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505050505050505,
      "loss": 2.5446,
      "theoretical_loss": 3.408223259983795,
      "tokens_seen": 2163015680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035048484848484854,
      "loss": 2.7503,
      "theoretical_loss": 3.408214782287217,
      "tokens_seen": 2163081216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504646464646465,
      "loss": 2.8492,
      "theoretical_loss": 3.4082063049194042,
      "tokens_seen": 2163146752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035044444444444446,
      "loss": 2.6935,
      "theoretical_loss": 3.4081978278803344,
      "tokens_seen": 2163212288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504242424242424,
      "loss": 2.659,
      "theoretical_loss": 3.4081893511699852,
      "tokens_seen": 2163277824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035040404040404037,
      "loss": 2.3811,
      "theoretical_loss": 3.408180874788334,
      "tokens_seen": 2163343360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035038383838383844,
      "loss": 2.7312,
      "theoretical_loss": 3.408172398735357,
      "tokens_seen": 2163408896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503636363636364,
      "loss": 2.7423,
      "theoretical_loss": 3.4081639230110325,
      "tokens_seen": 2163474432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035034343434343435,
      "loss": 2.7973,
      "theoretical_loss": 3.4081554476153375,
      "tokens_seen": 2163539968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503232323232323,
      "loss": 2.6042,
      "theoretical_loss": 3.408146972548249,
      "tokens_seen": 2163605504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035030303030303026,
      "loss": 2.4051,
      "theoretical_loss": 3.4081384978097446,
      "tokens_seen": 2163671040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035028282828282833,
      "loss": 2.5392,
      "theoretical_loss": 3.408130023399802,
      "tokens_seen": 2163736576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502626262626263,
      "loss": 2.5968,
      "theoretical_loss": 3.408121549318398,
      "tokens_seen": 2163802112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035024242424242424,
      "loss": 2.5486,
      "theoretical_loss": 3.40811307556551,
      "tokens_seen": 2163867648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502222222222222,
      "loss": 2.7453,
      "theoretical_loss": 3.4081046021411154,
      "tokens_seen": 2163933184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502020202020202,
      "loss": 2.6839,
      "theoretical_loss": 3.4080961290451914,
      "tokens_seen": 2163998720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501818181818182,
      "loss": 2.6244,
      "theoretical_loss": 3.4080876562777154,
      "tokens_seen": 2164064256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501616161616162,
      "loss": 2.4852,
      "theoretical_loss": 3.4080791838386646,
      "tokens_seen": 2164129792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035014141414141414,
      "loss": 2.5086,
      "theoretical_loss": 3.4080707117280165,
      "tokens_seen": 2164195328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501212121212121,
      "loss": 2.4746,
      "theoretical_loss": 3.4080622399457483,
      "tokens_seen": 2164260864
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501010101010101,
      "loss": 2.529,
      "theoretical_loss": 3.408053768491838,
      "tokens_seen": 2164326400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500808080808081,
      "loss": 2.2181,
      "theoretical_loss": 3.4080452973662614,
      "tokens_seen": 2164391936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035006060606060607,
      "loss": 2.7326,
      "theoretical_loss": 3.4080368265689973,
      "tokens_seen": 2164457472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035004040404040403,
      "loss": 2.4137,
      "theoretical_loss": 3.408028356100022,
      "tokens_seen": 2164523008
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1192084,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4357054233551025,
      "objective/train/theoretical_loss": 3.4080198859593143,
      "objective/train/tokens_used": 523447776,
      "theoretical_loss": 3.4080198859593143,
      "tokens_seen": 2164588544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035002020202020204,
      "loss": 2.5545,
      "theoretical_loss": 3.4080198859593143,
      "tokens_seen": 2164588544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035,
      "loss": 2.8726,
      "theoretical_loss": 3.4080114161468495,
      "tokens_seen": 2164654080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349979797979798,
      "loss": 2.5693,
      "theoretical_loss": 3.4080029466626067,
      "tokens_seen": 2164719616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034995959595959596,
      "loss": 2.4669,
      "theoretical_loss": 3.4079944775065623,
      "tokens_seen": 2164785152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499393939393939,
      "loss": 2.7631,
      "theoretical_loss": 3.407986008678694,
      "tokens_seen": 2164850688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034991919191919193,
      "loss": 2.7671,
      "theoretical_loss": 3.4079775401789787,
      "tokens_seen": 2164916224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498989898989899,
      "loss": 2.6294,
      "theoretical_loss": 3.4079690720073943,
      "tokens_seen": 2164981760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498787878787879,
      "loss": 2.4798,
      "theoretical_loss": 3.407960604163918,
      "tokens_seen": 2165047296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034985858585858586,
      "loss": 2.5692,
      "theoretical_loss": 3.4079521366485275,
      "tokens_seen": 2165112832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034983838383838387,
      "loss": 2.6114,
      "theoretical_loss": 3.4079436694611993,
      "tokens_seen": 2165178368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498181818181818,
      "loss": 2.7437,
      "theoretical_loss": 3.4079352026019114,
      "tokens_seen": 2165243904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497979797979798,
      "loss": 2.7779,
      "theoretical_loss": 3.4079267360706407,
      "tokens_seen": 2165309440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497777777777778,
      "loss": 2.4274,
      "theoretical_loss": 3.4079182698673653,
      "tokens_seen": 2165374976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034975757575757575,
      "loss": 2.5162,
      "theoretical_loss": 3.407909803992062,
      "tokens_seen": 2165440512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034973737373737376,
      "loss": 2.5543,
      "theoretical_loss": 3.407901338444708,
      "tokens_seen": 2165506048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497171717171717,
      "loss": 2.5477,
      "theoretical_loss": 3.4078928732252813,
      "tokens_seen": 2165571584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496969696969697,
      "loss": 2.5499,
      "theoretical_loss": 3.407884408333759,
      "tokens_seen": 2165637120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496767676767677,
      "loss": 2.6026,
      "theoretical_loss": 3.407875943770118,
      "tokens_seen": 2165702656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496565656565657,
      "loss": 2.6215,
      "theoretical_loss": 3.407867479534336,
      "tokens_seen": 2165768192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034963636363636365,
      "loss": 2.4358,
      "theoretical_loss": 3.407859015626391,
      "tokens_seen": 2165833728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496161616161616,
      "loss": 2.4935,
      "theoretical_loss": 3.40785055204626,
      "tokens_seen": 2165899264
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034959595959595957,
      "loss": 2.4719,
      "theoretical_loss": 3.40784208879392,
      "tokens_seen": 2165964800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495757575757576,
      "loss": 2.5277,
      "theoretical_loss": 3.4078336258693485,
      "tokens_seen": 2166030336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495555555555556,
      "loss": 2.585,
      "theoretical_loss": 3.407825163272523,
      "tokens_seen": 2166095872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034953535353535355,
      "loss": 2.7183,
      "theoretical_loss": 3.407816701003421,
      "tokens_seen": 2166161408
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1193465,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5298290252685547,
      "objective/train/theoretical_loss": 3.4078082390620197,
      "objective/train/tokens_used": 525086176,
      "theoretical_loss": 3.4078082390620197,
      "tokens_seen": 2166226944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495151515151515,
      "loss": 2.5874,
      "theoretical_loss": 3.4078082390620197,
      "tokens_seen": 2166226944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034949494949494946,
      "loss": 2.56,
      "theoretical_loss": 3.4077997774482967,
      "tokens_seen": 2166292480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494747474747475,
      "loss": 2.5426,
      "theoretical_loss": 3.4077913161622293,
      "tokens_seen": 2166358016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494545454545455,
      "loss": 2.7146,
      "theoretical_loss": 3.407782855203795,
      "tokens_seen": 2166423552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034943434343434344,
      "loss": 2.5124,
      "theoretical_loss": 3.4077743945729706,
      "tokens_seen": 2166489088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494141414141414,
      "loss": 2.6556,
      "theoretical_loss": 3.4077659342697344,
      "tokens_seen": 2166554624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034939393939393935,
      "loss": 2.5807,
      "theoretical_loss": 3.4077574742940637,
      "tokens_seen": 2166620160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493737373737374,
      "loss": 2.629,
      "theoretical_loss": 3.407749014645935,
      "tokens_seen": 2166685696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493535353535354,
      "loss": 2.538,
      "theoretical_loss": 3.4077405553253266,
      "tokens_seen": 2166751232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034933333333333333,
      "loss": 2.544,
      "theoretical_loss": 3.407732096332216,
      "tokens_seen": 2166816768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493131313131313,
      "loss": 2.6393,
      "theoretical_loss": 3.40772363766658,
      "tokens_seen": 2166882304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492929292929293,
      "loss": 2.5813,
      "theoretical_loss": 3.407715179328396,
      "tokens_seen": 2166947840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492727272727273,
      "loss": 2.7715,
      "theoretical_loss": 3.407706721317642,
      "tokens_seen": 2167013376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034925252525252527,
      "loss": 2.5057,
      "theoretical_loss": 3.4076982636342956,
      "tokens_seen": 2167078912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492323232323232,
      "loss": 2.8016,
      "theoretical_loss": 3.407689806278333,
      "tokens_seen": 2167144448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492121212121212,
      "loss": 2.8176,
      "theoretical_loss": 3.407681349249733,
      "tokens_seen": 2167209984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491919191919192,
      "loss": 2.459,
      "theoretical_loss": 3.407672892548472,
      "tokens_seen": 2167275520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491717171717172,
      "loss": 2.3919,
      "theoretical_loss": 3.4076644361745276,
      "tokens_seen": 2167341056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034915151515151516,
      "loss": 2.7155,
      "theoretical_loss": 3.4076559801278776,
      "tokens_seen": 2167406592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491313131313131,
      "loss": 2.6292,
      "theoretical_loss": 3.4076475244084996,
      "tokens_seen": 2167472128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034911111111111113,
      "loss": 2.7854,
      "theoretical_loss": 3.407639069016371,
      "tokens_seen": 2167537664
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034909090909090914,
      "loss": 2.6411,
      "theoretical_loss": 3.4076306139514685,
      "tokens_seen": 2167603200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490707070707071,
      "loss": 2.6537,
      "theoretical_loss": 3.40762215921377,
      "tokens_seen": 2167668736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034905050505050505,
      "loss": 2.8707,
      "theoretical_loss": 3.4076137048032535,
      "tokens_seen": 2167734272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349030303030303,
      "loss": 2.5394,
      "theoretical_loss": 3.4076052507198957,
      "tokens_seen": 2167799808
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1194124,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.193073272705078,
      "objective/train/theoretical_loss": 3.4075967969636745,
      "objective/train/tokens_used": 526724576,
      "theoretical_loss": 3.4075967969636745,
      "tokens_seen": 2167865344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349010101010101,
      "loss": 2.4182,
      "theoretical_loss": 3.4075967969636745,
      "tokens_seen": 2167865344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034898989898989903,
      "loss": 2.7835,
      "theoretical_loss": 3.4075883435345666,
      "tokens_seen": 2167930880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348969696969697,
      "loss": 2.639,
      "theoretical_loss": 3.40757989043255,
      "tokens_seen": 2167996416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034894949494949495,
      "loss": 2.6562,
      "theoretical_loss": 3.4075714376576025,
      "tokens_seen": 2168061952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034892929292929296,
      "loss": 2.6586,
      "theoretical_loss": 3.4075629852097014,
      "tokens_seen": 2168127488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489090909090909,
      "loss": 2.4807,
      "theoretical_loss": 3.4075545330888235,
      "tokens_seen": 2168193024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488888888888889,
      "loss": 2.6037,
      "theoretical_loss": 3.4075460812949467,
      "tokens_seen": 2168258560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488686868686869,
      "loss": 2.633,
      "theoretical_loss": 3.407537629828049,
      "tokens_seen": 2168324096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034884848484848484,
      "loss": 2.5063,
      "theoretical_loss": 3.4075291786881072,
      "tokens_seen": 2168389632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034882828282828285,
      "loss": 2.5902,
      "theoretical_loss": 3.4075207278750987,
      "tokens_seen": 2168455168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488080808080808,
      "loss": 2.7776,
      "theoretical_loss": 3.407512277389001,
      "tokens_seen": 2168520704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487878787878788,
      "loss": 2.4823,
      "theoretical_loss": 3.4075038272297924,
      "tokens_seen": 2168586240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487676767676768,
      "loss": 2.5794,
      "theoretical_loss": 3.4074953773974492,
      "tokens_seen": 2168651776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034874747474747473,
      "loss": 2.59,
      "theoretical_loss": 3.40748692789195,
      "tokens_seen": 2168717312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034872727272727274,
      "loss": 2.4795,
      "theoretical_loss": 3.4074784787132715,
      "tokens_seen": 2168782848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487070707070707,
      "loss": 2.4988,
      "theoretical_loss": 3.407470029861391,
      "tokens_seen": 2168848384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486868686868687,
      "loss": 2.6872,
      "theoretical_loss": 3.407461581336287,
      "tokens_seen": 2168913920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034866666666666667,
      "loss": 2.5071,
      "theoretical_loss": 3.407453133137936,
      "tokens_seen": 2168979456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486464646464647,
      "loss": 2.8315,
      "theoretical_loss": 3.407444685266316,
      "tokens_seen": 2169044992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034862626262626263,
      "loss": 2.5917,
      "theoretical_loss": 3.407436237721404,
      "tokens_seen": 2169110528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486060606060606,
      "loss": 2.7336,
      "theoretical_loss": 3.4074277905031782,
      "tokens_seen": 2169176064
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485858585858586,
      "loss": 2.463,
      "theoretical_loss": 3.407419343611616,
      "tokens_seen": 2169241600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034856565656565656,
      "loss": 2.4789,
      "theoretical_loss": 3.4074108970466943,
      "tokens_seen": 2169307136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034854545454545457,
      "loss": 2.6116,
      "theoretical_loss": 3.407402450808391,
      "tokens_seen": 2169372672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034852525252525253,
      "loss": 2.6967,
      "theoretical_loss": 3.4073940048966835,
      "tokens_seen": 2169438208
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1195527,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4617276191711426,
      "objective/train/theoretical_loss": 3.4073855593115496,
      "objective/train/tokens_used": 528362976,
      "theoretical_loss": 3.4073855593115496,
      "tokens_seen": 2169503744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485050505050505,
      "loss": 2.602,
      "theoretical_loss": 3.4073855593115496,
      "tokens_seen": 2169503744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484848484848485,
      "loss": 2.567,
      "theoretical_loss": 3.4073771140529665,
      "tokens_seen": 2169569280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484646464646465,
      "loss": 2.5603,
      "theoretical_loss": 3.4073686691209115,
      "tokens_seen": 2169634816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034844444444444446,
      "loss": 2.6815,
      "theoretical_loss": 3.4073602245153625,
      "tokens_seen": 2169700352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484242424242424,
      "loss": 2.6966,
      "theoretical_loss": 3.407351780236297,
      "tokens_seen": 2169765888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484040404040404,
      "loss": 2.9067,
      "theoretical_loss": 3.4073433362836925,
      "tokens_seen": 2169831424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483838383838384,
      "loss": 2.5778,
      "theoretical_loss": 3.4073348926575266,
      "tokens_seen": 2169896960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483636363636364,
      "loss": 2.5324,
      "theoretical_loss": 3.407326449357776,
      "tokens_seen": 2169962496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034834343434343436,
      "loss": 2.7399,
      "theoretical_loss": 3.4073180063844197,
      "tokens_seen": 2170028032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483232323232323,
      "loss": 2.5623,
      "theoretical_loss": 3.4073095637374338,
      "tokens_seen": 2170093568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034830303030303027,
      "loss": 2.5096,
      "theoretical_loss": 3.407301121416797,
      "tokens_seen": 2170159104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034828282828282834,
      "loss": 2.717,
      "theoretical_loss": 3.407292679422486,
      "tokens_seen": 2170224640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003482626262626263,
      "loss": 2.5496,
      "theoretical_loss": 3.407284237754478,
      "tokens_seen": 2170290176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034824242424242425,
      "loss": 2.6856,
      "theoretical_loss": 3.407275796412752,
      "tokens_seen": 2170355712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003482222222222222,
      "loss": 2.6982,
      "theoretical_loss": 3.407267355397284,
      "tokens_seen": 2170421248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034820202020202016,
      "loss": 2.6895,
      "theoretical_loss": 3.407258914708053,
      "tokens_seen": 2170486784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034818181818181823,
      "loss": 2.554,
      "theoretical_loss": 3.407250474345035,
      "tokens_seen": 2170552320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481616161616162,
      "loss": 2.62,
      "theoretical_loss": 3.407242034308209,
      "tokens_seen": 2170617856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034814141414141414,
      "loss": 2.6023,
      "theoretical_loss": 3.4072335945975514,
      "tokens_seen": 2170683392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481212121212121,
      "loss": 2.6248,
      "theoretical_loss": 3.40722515521304,
      "tokens_seen": 2170748928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481010101010101,
      "loss": 2.6655,
      "theoretical_loss": 3.407216716154653,
      "tokens_seen": 2170814464
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480808080808081,
      "loss": 2.6005,
      "theoretical_loss": 3.4072082774223675,
      "tokens_seen": 2170880000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480606060606061,
      "loss": 2.4526,
      "theoretical_loss": 3.407199839016161,
      "tokens_seen": 2170945536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034804040404040403,
      "loss": 2.634,
      "theoretical_loss": 3.4071914009360107,
      "tokens_seen": 2171011072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348020202020202,
      "loss": 2.5642,
      "theoretical_loss": 3.407182963181895,
      "tokens_seen": 2171076608
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1196718,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4015064239501953,
      "objective/train/theoretical_loss": 3.4071745257537907,
      "objective/train/tokens_used": 530001376,
      "theoretical_loss": 3.4071745257537907,
      "tokens_seen": 2171142144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348,
      "loss": 2.6376,
      "theoretical_loss": 3.4071745257537907,
      "tokens_seen": 2171142144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347979797979798,
      "loss": 2.5995,
      "theoretical_loss": 3.407166088651676,
      "tokens_seen": 2171207680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034795959595959597,
      "loss": 2.6409,
      "theoretical_loss": 3.407157651875528,
      "tokens_seen": 2171273216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479393939393939,
      "loss": 2.5,
      "theoretical_loss": 3.4071492154253242,
      "tokens_seen": 2171338752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034791919191919194,
      "loss": 2.6478,
      "theoretical_loss": 3.407140779301043,
      "tokens_seen": 2171404288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478989898989899,
      "loss": 2.527,
      "theoretical_loss": 3.407132343502661,
      "tokens_seen": 2171469824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478787878787879,
      "loss": 2.5339,
      "theoretical_loss": 3.4071239080301563,
      "tokens_seen": 2171535360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034785858585858586,
      "loss": 2.4777,
      "theoretical_loss": 3.4071154728835062,
      "tokens_seen": 2171600896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478383838383838,
      "loss": 2.3114,
      "theoretical_loss": 3.407107038062688,
      "tokens_seen": 2171666432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034781818181818183,
      "loss": 2.6652,
      "theoretical_loss": 3.4070986035676802,
      "tokens_seen": 2171731968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477979797979798,
      "loss": 2.5996,
      "theoretical_loss": 3.40709016939846,
      "tokens_seen": 2171797504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477777777777778,
      "loss": 2.6978,
      "theoretical_loss": 3.4070817355550047,
      "tokens_seen": 2171863040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034775757575757576,
      "loss": 2.4345,
      "theoretical_loss": 3.4070733020372916,
      "tokens_seen": 2171928576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034773737373737377,
      "loss": 2.6146,
      "theoretical_loss": 3.4070648688452994,
      "tokens_seen": 2171994112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477171717171717,
      "loss": 2.4398,
      "theoretical_loss": 3.4070564359790048,
      "tokens_seen": 2172059648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476969696969697,
      "loss": 2.5982,
      "theoretical_loss": 3.407048003438385,
      "tokens_seen": 2172125184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476767676767677,
      "loss": 2.7686,
      "theoretical_loss": 3.407039571223419,
      "tokens_seen": 2172190720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034765656565656565,
      "loss": 2.8074,
      "theoretical_loss": 3.4070311393340837,
      "tokens_seen": 2172256256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034763636363636366,
      "loss": 2.6708,
      "theoretical_loss": 3.407022707770356,
      "tokens_seen": 2172321792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476161616161616,
      "loss": 2.6059,
      "theoretical_loss": 3.4070142765322147,
      "tokens_seen": 2172387328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034759595959595957,
      "loss": 2.4886,
      "theoretical_loss": 3.4070058456196364,
      "tokens_seen": 2172452864
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475757575757576,
      "loss": 2.4637,
      "theoretical_loss": 3.406997415032599,
      "tokens_seen": 2172518400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475555555555556,
      "loss": 2.6471,
      "theoretical_loss": 3.406988984771081,
      "tokens_seen": 2172583936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034753535353535355,
      "loss": 2.5464,
      "theoretical_loss": 3.4069805548350587,
      "tokens_seen": 2172649472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475151515151515,
      "loss": 2.7893,
      "theoretical_loss": 3.40697212522451,
      "tokens_seen": 2172715008
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1197347,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8142454624176025,
      "objective/train/theoretical_loss": 3.4069636959394134,
      "objective/train/tokens_used": 531639776,
      "theoretical_loss": 3.4069636959394134,
      "tokens_seen": 2172780544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034749494949494947,
      "loss": 2.5133,
      "theoretical_loss": 3.4069636959394134,
      "tokens_seen": 2172780544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474747474747475,
      "loss": 2.6388,
      "theoretical_loss": 3.4069552669797454,
      "tokens_seen": 2172846080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474545454545455,
      "loss": 2.3698,
      "theoretical_loss": 3.4069468383454846,
      "tokens_seen": 2172911616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034743434343434344,
      "loss": 2.4434,
      "theoretical_loss": 3.4069384100366076,
      "tokens_seen": 2172977152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474141414141414,
      "loss": 2.4434,
      "theoretical_loss": 3.406929982053093,
      "tokens_seen": 2173042688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473939393939394,
      "loss": 2.5736,
      "theoretical_loss": 3.406921554394918,
      "tokens_seen": 2173108224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473737373737374,
      "loss": 2.6301,
      "theoretical_loss": 3.40691312706206,
      "tokens_seen": 2173173760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473535353535354,
      "loss": 2.6759,
      "theoretical_loss": 3.406904700054497,
      "tokens_seen": 2173239296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034733333333333334,
      "loss": 2.6137,
      "theoretical_loss": 3.406896273372207,
      "tokens_seen": 2173304832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473131313131313,
      "loss": 2.5298,
      "theoretical_loss": 3.4068878470151662,
      "tokens_seen": 2173370368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472929292929293,
      "loss": 2.6246,
      "theoretical_loss": 3.406879420983354,
      "tokens_seen": 2173435904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472727272727273,
      "loss": 2.5862,
      "theoretical_loss": 3.406870995276747,
      "tokens_seen": 2173501440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472525252525253,
      "loss": 2.594,
      "theoretical_loss": 3.4068625698953223,
      "tokens_seen": 2173566976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034723232323232323,
      "loss": 2.4562,
      "theoretical_loss": 3.406854144839059,
      "tokens_seen": 2173632512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472121212121212,
      "loss": 2.5802,
      "theoretical_loss": 3.4068457201079343,
      "tokens_seen": 2173698048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034719191919191925,
      "loss": 2.6286,
      "theoretical_loss": 3.406837295701925,
      "tokens_seen": 2173763584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471717171717172,
      "loss": 2.6744,
      "theoretical_loss": 3.40682887162101,
      "tokens_seen": 2173829120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034715151515151517,
      "loss": 2.4594,
      "theoretical_loss": 3.406820447865166,
      "tokens_seen": 2173894656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471313131313131,
      "loss": 2.4142,
      "theoretical_loss": 3.406812024434371,
      "tokens_seen": 2173960192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471111111111111,
      "loss": 2.5439,
      "theoretical_loss": 3.4068036013286025,
      "tokens_seen": 2174025728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034709090909090914,
      "loss": 2.4671,
      "theoretical_loss": 3.406795178547838,
      "tokens_seen": 2174091264
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470707070707071,
      "loss": 2.5607,
      "theoretical_loss": 3.406786756092056,
      "tokens_seen": 2174156800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034705050505050506,
      "loss": 2.5012,
      "theoretical_loss": 3.4067783339612334,
      "tokens_seen": 2174222336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347030303030303,
      "loss": 2.5176,
      "theoretical_loss": 3.4067699121553483,
      "tokens_seen": 2174287872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034701010101010097,
      "loss": 2.5296,
      "theoretical_loss": 3.406761490674378,
      "tokens_seen": 2174353408
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1198617,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.766582489013672,
      "objective/train/theoretical_loss": 3.4067530695183,
      "objective/train/tokens_used": 533278176,
      "theoretical_loss": 3.4067530695183,
      "tokens_seen": 2174418944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034698989898989904,
      "loss": 2.5671,
      "theoretical_loss": 3.4067530695183,
      "tokens_seen": 2174418944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346969696969697,
      "loss": 2.4026,
      "theoretical_loss": 3.4067446486870927,
      "tokens_seen": 2174484480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034694949494949495,
      "loss": 2.7165,
      "theoretical_loss": 3.4067362281807334,
      "tokens_seen": 2174550016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003469292929292929,
      "loss": 2.6798,
      "theoretical_loss": 3.4067278079991996,
      "tokens_seen": 2174615552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003469090909090909,
      "loss": 2.6543,
      "theoretical_loss": 3.4067193881424696,
      "tokens_seen": 2174681088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034688888888888893,
      "loss": 2.5014,
      "theoretical_loss": 3.40671096861052,
      "tokens_seen": 2174746624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468686868686869,
      "loss": 2.502,
      "theoretical_loss": 3.406702549403329,
      "tokens_seen": 2174812160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034684848484848484,
      "loss": 2.5578,
      "theoretical_loss": 3.4066941305208753,
      "tokens_seen": 2174877696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468282828282828,
      "loss": 2.5906,
      "theoretical_loss": 3.406685711963135,
      "tokens_seen": 2174943232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468080808080808,
      "loss": 2.6116,
      "theoretical_loss": 3.4066772937300867,
      "tokens_seen": 2175008768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467878787878788,
      "loss": 2.3412,
      "theoretical_loss": 3.406668875821708,
      "tokens_seen": 2175074304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467676767676768,
      "loss": 2.541,
      "theoretical_loss": 3.406660458237976,
      "tokens_seen": 2175139840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034674747474747474,
      "loss": 2.5753,
      "theoretical_loss": 3.406652040978869,
      "tokens_seen": 2175205376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034672727272727275,
      "loss": 2.4651,
      "theoretical_loss": 3.406643624044365,
      "tokens_seen": 2175270912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467070707070707,
      "loss": 2.4329,
      "theoretical_loss": 3.406635207434441,
      "tokens_seen": 2175336448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466868686868687,
      "loss": 2.6881,
      "theoretical_loss": 3.406626791149075,
      "tokens_seen": 2175401984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034666666666666667,
      "loss": 2.5079,
      "theoretical_loss": 3.4066183751882444,
      "tokens_seen": 2175467520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034664646464646463,
      "loss": 2.6497,
      "theoretical_loss": 3.4066099595519272,
      "tokens_seen": 2175533056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034662626262626264,
      "loss": 2.6699,
      "theoretical_loss": 3.4066015442401016,
      "tokens_seen": 2175598592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466060606060606,
      "loss": 2.6234,
      "theoretical_loss": 3.4065931292527445,
      "tokens_seen": 2175664128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465858585858586,
      "loss": 2.453,
      "theoretical_loss": 3.406584714589834,
      "tokens_seen": 2175729664
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034656565656565657,
      "loss": 2.5747,
      "theoretical_loss": 3.406576300251348,
      "tokens_seen": 2175795200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465454545454546,
      "loss": 2.7449,
      "theoretical_loss": 3.4065678862372635,
      "tokens_seen": 2175860736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034652525252525253,
      "loss": 2.6427,
      "theoretical_loss": 3.406559472547559,
      "tokens_seen": 2175926272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465050505050505,
      "loss": 2.509,
      "theoretical_loss": 3.406551059182212,
      "tokens_seen": 2175991808
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1199271,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.564194440841675,
      "objective/train/theoretical_loss": 3.4065426461412,
      "objective/train/tokens_used": 534916576,
      "theoretical_loss": 3.4065426461412,
      "tokens_seen": 2176057344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464848484848485,
      "loss": 2.619,
      "theoretical_loss": 3.4065426461412,
      "tokens_seen": 2176057344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034646464646464646,
      "loss": 2.5785,
      "theoretical_loss": 3.406534233424501,
      "tokens_seen": 2176122880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034644444444444447,
      "loss": 2.5945,
      "theoretical_loss": 3.4065258210320923,
      "tokens_seen": 2176188416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464242424242424,
      "loss": 2.3552,
      "theoretical_loss": 3.4065174089639525,
      "tokens_seen": 2176253952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464040404040404,
      "loss": 2.5352,
      "theoretical_loss": 3.406508997220058,
      "tokens_seen": 2176319488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463838383838384,
      "loss": 2.661,
      "theoretical_loss": 3.406500585800388,
      "tokens_seen": 2176385024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463636363636364,
      "loss": 2.5717,
      "theoretical_loss": 3.406492174704919,
      "tokens_seen": 2176450560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034634343434343436,
      "loss": 2.6531,
      "theoretical_loss": 3.40648376393363,
      "tokens_seen": 2176516096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463232323232323,
      "loss": 2.5537,
      "theoretical_loss": 3.4064753534864973,
      "tokens_seen": 2176581632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463030303030303,
      "loss": 2.6112,
      "theoretical_loss": 3.4064669433635,
      "tokens_seen": 2176647168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462828282828283,
      "loss": 2.5072,
      "theoretical_loss": 3.4064585335646154,
      "tokens_seen": 2176712704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462626262626263,
      "loss": 2.4548,
      "theoretical_loss": 3.4064501240898206,
      "tokens_seen": 2176778240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034624242424242425,
      "loss": 2.5108,
      "theoretical_loss": 3.4064417149390938,
      "tokens_seen": 2176843776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462222222222222,
      "loss": 2.4037,
      "theoretical_loss": 3.406433306112413,
      "tokens_seen": 2176909312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034620202020202017,
      "loss": 2.3914,
      "theoretical_loss": 3.4064248976097558,
      "tokens_seen": 2176974848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034618181818181823,
      "loss": 2.4264,
      "theoretical_loss": 3.4064164894310998,
      "tokens_seen": 2177040384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461616161616162,
      "loss": 2.6455,
      "theoretical_loss": 3.406408081576423,
      "tokens_seen": 2177105920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034614141414141415,
      "loss": 2.4222,
      "theoretical_loss": 3.406399674045703,
      "tokens_seen": 2177171456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461212121212121,
      "loss": 2.6485,
      "theoretical_loss": 3.4063912668389174,
      "tokens_seen": 2177236992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034610101010101006,
      "loss": 2.7297,
      "theoretical_loss": 3.406382859956045,
      "tokens_seen": 2177302528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460808080808081,
      "loss": 2.4382,
      "theoretical_loss": 3.4063744533970617,
      "tokens_seen": 2177368064
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460606060606061,
      "loss": 2.2843,
      "theoretical_loss": 3.406366047161947,
      "tokens_seen": 2177433600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034604040404040404,
      "loss": 2.8041,
      "theoretical_loss": 3.4063576412506777,
      "tokens_seen": 2177499136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346020202020202,
      "loss": 2.6357,
      "theoretical_loss": 3.406349235663232,
      "tokens_seen": 2177564672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346,
      "loss": 2.5708,
      "theoretical_loss": 3.4063408303995875,
      "tokens_seen": 2177630208
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 1200420,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5001168251037598,
      "objective/train/theoretical_loss": 3.4063324254597225,
      "objective/train/tokens_used": 536554976,
      "theoretical_loss": 3.4063324254597225,
      "tokens_seen": 2177695744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345979797979798,
      "loss": 2.5511,
      "theoretical_loss": 3.4063324254597225,
      "tokens_seen": 2177695744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345959595959596,
      "loss": 2.4979,
      "theoretical_loss": 3.4063240208436136,
      "tokens_seen": 2177761280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034593939393939393,
      "loss": 2.6736,
      "theoretical_loss": 3.40631561655124,
      "tokens_seen": 2177826816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003459191919191919,
      "loss": 2.628,
      "theoretical_loss": 3.406307212582578,
      "tokens_seen": 2177892352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458989898989899,
      "loss": 2.6927,
      "theoretical_loss": 3.406298808937607,
      "tokens_seen": 2177957888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458787878787879,
      "loss": 2.505,
      "theoretical_loss": 3.4062904056163035,
      "tokens_seen": 2178023424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034585858585858587,
      "loss": 2.5307,
      "theoretical_loss": 3.406282002618646,
      "tokens_seen": 2178088960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458383838383838,
      "loss": 2.6472,
      "theoretical_loss": 3.4062735999446123,
      "tokens_seen": 2178154496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034581818181818184,
      "loss": 2.4959,
      "theoretical_loss": 3.4062651975941796,
      "tokens_seen": 2178220032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457979797979798,
      "loss": 2.4972,
      "theoretical_loss": 3.4062567955673266,
      "tokens_seen": 2178285568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457777777777778,
      "loss": 2.6403,
      "theoretical_loss": 3.40624839386403,
      "tokens_seen": 2178351104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034575757575757576,
      "loss": 2.8023,
      "theoretical_loss": 3.4062399924842683,
      "tokens_seen": 2178416640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457373737373737,
      "loss": 2.6405,
      "theoretical_loss": 3.4062315914280195,
      "tokens_seen": 2178482176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034571717171717173,
      "loss": 2.4173,
      "theoretical_loss": 3.406223190695261,
      "tokens_seen": 2178547712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456969696969697,
      "loss": 2.5385,
      "theoretical_loss": 3.4062147902859707,
      "tokens_seen": 2178613248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456767676767677,
      "loss": 2.6363,
      "theoretical_loss": 3.4062063902001265,
      "tokens_seen": 2178678784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034565656565656565,
      "loss": 2.6864,
      "theoretical_loss": 3.4061979904377058,
      "tokens_seen": 2178744320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034563636363636366,
      "loss": 2.6469,
      "theoretical_loss": 3.4061895909986872,
      "tokens_seen": 2178809856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456161616161616,
      "loss": 2.4382,
      "theoretical_loss": 3.4061811918830482,
      "tokens_seen": 2178875392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034559595959595963,
      "loss": 2.5073,
      "theoretical_loss": 3.406172793090766,
      "tokens_seen": 2178940928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455757575757576,
      "loss": 2.5229,
      "theoretical_loss": 3.406164394621819,
      "tokens_seen": 2179006464
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034555555555555555,
      "loss": 2.3853,
      "theoretical_loss": 3.4061559964761856,
      "tokens_seen": 2179072000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034553535353535356,
      "loss": 2.7215,
      "theoretical_loss": 3.4061475986538428,
      "tokens_seen": 2179137536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455151515151515,
      "loss": 2.4677,
      "theoretical_loss": 3.4061392011547684,
      "tokens_seen": 2179203072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454949494949495,
      "loss": 2.7491,
      "theoretical_loss": 3.4061308039789404,
      "tokens_seen": 2179268608
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1201163,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7939035892486572,
      "objective/train/theoretical_loss": 3.406122407126337,
      "objective/train/tokens_used": 538193376,
      "theoretical_loss": 3.406122407126337,
      "tokens_seen": 2179334144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454747474747475,
      "loss": 2.5683,
      "theoretical_loss": 3.406122407126337,
      "tokens_seen": 2179334144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034545454545454544,
      "loss": 2.5964,
      "theoretical_loss": 3.406114010596936,
      "tokens_seen": 2179399680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034543434343434345,
      "loss": 2.4462,
      "theoretical_loss": 3.4061056143907145,
      "tokens_seen": 2179465216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454141414141414,
      "loss": 2.6405,
      "theoretical_loss": 3.406097218507651,
      "tokens_seen": 2179530752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453939393939394,
      "loss": 2.4423,
      "theoretical_loss": 3.406088822947723,
      "tokens_seen": 2179596288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453737373737374,
      "loss": 2.4505,
      "theoretical_loss": 3.4060804277109087,
      "tokens_seen": 2179661824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453535353535354,
      "loss": 2.5004,
      "theoretical_loss": 3.406072032797186,
      "tokens_seen": 2179727360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034533333333333334,
      "loss": 2.4309,
      "theoretical_loss": 3.406063638206532,
      "tokens_seen": 2179792896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453131313131313,
      "loss": 2.7099,
      "theoretical_loss": 3.4060552439389253,
      "tokens_seen": 2179858432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452929292929293,
      "loss": 2.3582,
      "theoretical_loss": 3.4060468499943437,
      "tokens_seen": 2179923968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034527272727272727,
      "loss": 2.4526,
      "theoretical_loss": 3.406038456372765,
      "tokens_seen": 2179989504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452525252525253,
      "loss": 2.6437,
      "theoretical_loss": 3.4060300630741667,
      "tokens_seen": 2180055040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034523232323232324,
      "loss": 2.6118,
      "theoretical_loss": 3.4060216700985273,
      "tokens_seen": 2180120576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452121212121212,
      "loss": 2.5669,
      "theoretical_loss": 3.4060132774458243,
      "tokens_seen": 2180186112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451919191919192,
      "loss": 2.625,
      "theoretical_loss": 3.406004885116035,
      "tokens_seen": 2180251648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451717171717172,
      "loss": 2.4939,
      "theoretical_loss": 3.4059964931091384,
      "tokens_seen": 2180317184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034515151515151517,
      "loss": 2.7215,
      "theoretical_loss": 3.405988101425111,
      "tokens_seen": 2180382720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034513131313131313,
      "loss": 2.525,
      "theoretical_loss": 3.4059797100639324,
      "tokens_seen": 2180448256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451111111111111,
      "loss": 2.6231,
      "theoretical_loss": 3.4059713190255794,
      "tokens_seen": 2180513792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450909090909091,
      "loss": 2.6351,
      "theoretical_loss": 3.4059629283100294,
      "tokens_seen": 2180579328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450707070707071,
      "loss": 2.4628,
      "theoretical_loss": 3.4059545379172618,
      "tokens_seen": 2180644864
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034505050505050506,
      "loss": 2.4169,
      "theoretical_loss": 3.405946147847253,
      "tokens_seen": 2180710400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345030303030303,
      "loss": 2.5559,
      "theoretical_loss": 3.4059377580999817,
      "tokens_seen": 2180775936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345010101010101,
      "loss": 2.289,
      "theoretical_loss": 3.4059293686754253,
      "tokens_seen": 2180841472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034498989898989904,
      "loss": 2.8415,
      "theoretical_loss": 3.4059209795735623,
      "tokens_seen": 2180907008
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1202317,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.683361768722534,
      "objective/train/theoretical_loss": 3.40591259079437,
      "objective/train/tokens_used": 539831776,
      "theoretical_loss": 3.40591259079437,
      "tokens_seen": 2180972544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344969696969697,
      "loss": 2.5561,
      "theoretical_loss": 3.40591259079437,
      "tokens_seen": 2180972544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034494949494949496,
      "loss": 2.8251,
      "theoretical_loss": 3.405904202337827,
      "tokens_seen": 2181038080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449292929292929,
      "loss": 2.8263,
      "theoretical_loss": 3.40589581420391,
      "tokens_seen": 2181103616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034490909090909087,
      "loss": 2.3851,
      "theoretical_loss": 3.405887426392598,
      "tokens_seen": 2181169152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034488888888888894,
      "loss": 2.6011,
      "theoretical_loss": 3.4058790389038687,
      "tokens_seen": 2181234688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448686868686869,
      "loss": 2.5665,
      "theoretical_loss": 3.4058706517376995,
      "tokens_seen": 2181300224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034484848484848485,
      "loss": 2.6355,
      "theoretical_loss": 3.4058622648940693,
      "tokens_seen": 2181365760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448282828282828,
      "loss": 2.6874,
      "theoretical_loss": 3.4058538783729544,
      "tokens_seen": 2181431296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448080808080808,
      "loss": 2.2167,
      "theoretical_loss": 3.4058454921743344,
      "tokens_seen": 2181496832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034478787878787883,
      "loss": 2.5234,
      "theoretical_loss": 3.4058371062981863,
      "tokens_seen": 2181562368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447676767676768,
      "loss": 2.4368,
      "theoretical_loss": 3.405828720744488,
      "tokens_seen": 2181627904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034474747474747474,
      "loss": 2.4132,
      "theoretical_loss": 3.4058203355132175,
      "tokens_seen": 2181693440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447272727272727,
      "loss": 2.4337,
      "theoretical_loss": 3.4058119506043534,
      "tokens_seen": 2181758976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447070707070707,
      "loss": 2.6031,
      "theoretical_loss": 3.4058035660178727,
      "tokens_seen": 2181824512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446868686868687,
      "loss": 2.5448,
      "theoretical_loss": 3.4057951817537537,
      "tokens_seen": 2181890048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446666666666667,
      "loss": 2.4287,
      "theoretical_loss": 3.405786797811974,
      "tokens_seen": 2181955584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034464646464646463,
      "loss": 2.5903,
      "theoretical_loss": 3.405778414192512,
      "tokens_seen": 2182021120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034462626262626265,
      "loss": 2.5824,
      "theoretical_loss": 3.4057700308953454,
      "tokens_seen": 2182086656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446060606060606,
      "loss": 2.5581,
      "theoretical_loss": 3.4057616479204524,
      "tokens_seen": 2182152192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445858585858586,
      "loss": 2.7394,
      "theoretical_loss": 3.4057532652678106,
      "tokens_seen": 2182217728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034456565656565657,
      "loss": 2.4358,
      "theoretical_loss": 3.405744882937398,
      "tokens_seen": 2182283264
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034454545454545453,
      "loss": 2.5903,
      "theoretical_loss": 3.4057365009291924,
      "tokens_seen": 2182348800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034452525252525254,
      "loss": 2.4804,
      "theoretical_loss": 3.405728119243172,
      "tokens_seen": 2182414336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445050505050505,
      "loss": 2.5562,
      "theoretical_loss": 3.405719737879315,
      "tokens_seen": 2182479872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444848484848485,
      "loss": 2.4465,
      "theoretical_loss": 3.4057113568375987,
      "tokens_seen": 2182545408
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1203427,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4712274074554443,
      "objective/train/theoretical_loss": 3.4057029761180013,
      "objective/train/tokens_used": 541470176,
      "theoretical_loss": 3.4057029761180013,
      "tokens_seen": 2182610944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034446464646464646,
      "loss": 2.5553,
      "theoretical_loss": 3.4057029761180013,
      "tokens_seen": 2182610944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444444444444445,
      "loss": 2.7063,
      "theoretical_loss": 3.405694595720501,
      "tokens_seen": 2182676480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034442424242424243,
      "loss": 2.6877,
      "theoretical_loss": 3.4056862156450753,
      "tokens_seen": 2182742016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444040404040404,
      "loss": 2.5239,
      "theoretical_loss": 3.4056778358917024,
      "tokens_seen": 2182807552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443838383838384,
      "loss": 2.564,
      "theoretical_loss": 3.4056694564603607,
      "tokens_seen": 2182873088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034436363636363636,
      "loss": 2.3851,
      "theoretical_loss": 3.405661077351027,
      "tokens_seen": 2182938624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034434343434343437,
      "loss": 2.463,
      "theoretical_loss": 3.4056526985636806,
      "tokens_seen": 2183004160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443232323232323,
      "loss": 2.5116,
      "theoretical_loss": 3.405644320098298,
      "tokens_seen": 2183069696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443030303030303,
      "loss": 2.7442,
      "theoretical_loss": 3.405635941954859,
      "tokens_seen": 2183135232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442828282828283,
      "loss": 2.5131,
      "theoretical_loss": 3.40562756413334,
      "tokens_seen": 2183200768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442626262626263,
      "loss": 2.6491,
      "theoretical_loss": 3.4056191866337198,
      "tokens_seen": 2183266304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034424242424242426,
      "loss": 2.4617,
      "theoretical_loss": 3.405610809455976,
      "tokens_seen": 2183331840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442222222222222,
      "loss": 2.5203,
      "theoretical_loss": 3.4056024326000864,
      "tokens_seen": 2183397376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442020202020202,
      "loss": 2.5003,
      "theoretical_loss": 3.4055940560660294,
      "tokens_seen": 2183462912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441818181818182,
      "loss": 2.528,
      "theoretical_loss": 3.405585679853783,
      "tokens_seen": 2183528448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441616161616162,
      "loss": 2.6176,
      "theoretical_loss": 3.405577303963325,
      "tokens_seen": 2183593984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034414141414141415,
      "loss": 2.6952,
      "theoretical_loss": 3.405568928394633,
      "tokens_seen": 2183659520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441212121212121,
      "loss": 2.5119,
      "theoretical_loss": 3.4055605531476854,
      "tokens_seen": 2183725056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034410101010101007,
      "loss": 2.5662,
      "theoretical_loss": 3.4055521782224605,
      "tokens_seen": 2183790592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034408080808080813,
      "loss": 2.428,
      "theoretical_loss": 3.4055438036189356,
      "tokens_seen": 2183856128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003440606060606061,
      "loss": 2.2457,
      "theoretical_loss": 3.405535429337089,
      "tokens_seen": 2183921664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034404040404040405,
      "loss": 2.6719,
      "theoretical_loss": 3.405527055376899,
      "tokens_seen": 2183987200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344020202020202,
      "loss": 2.5255,
      "theoretical_loss": 3.4055186817383434,
      "tokens_seen": 2184052736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034399999999999996,
      "loss": 2.7181,
      "theoretical_loss": 3.4055103084213996,
      "tokens_seen": 2184118272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343979797979798,
      "loss": 2.6533,
      "theoretical_loss": 3.4055019354260465,
      "tokens_seen": 2184183808
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1204022,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1142489910125732,
      "objective/train/theoretical_loss": 3.4054935627522616,
      "objective/train/tokens_used": 543108576,
      "theoretical_loss": 3.4054935627522616,
      "tokens_seen": 2184249344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343959595959596,
      "loss": 2.4261,
      "theoretical_loss": 3.4054935627522616,
      "tokens_seen": 2184249344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034393939393939394,
      "loss": 2.616,
      "theoretical_loss": 3.405485190400023,
      "tokens_seen": 2184314880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439191919191919,
      "loss": 2.3905,
      "theoretical_loss": 3.4054768183693085,
      "tokens_seen": 2184380416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034389898989898996,
      "loss": 2.3627,
      "theoretical_loss": 3.4054684466600964,
      "tokens_seen": 2184445952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438787878787879,
      "loss": 2.7071,
      "theoretical_loss": 3.405460075272365,
      "tokens_seen": 2184511488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438585858585859,
      "loss": 2.6045,
      "theoretical_loss": 3.4054517042060914,
      "tokens_seen": 2184577024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034383838383838383,
      "loss": 2.707,
      "theoretical_loss": 3.4054433334612546,
      "tokens_seen": 2184642560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438181818181818,
      "loss": 2.6189,
      "theoretical_loss": 3.4054349630378318,
      "tokens_seen": 2184708096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034379797979797985,
      "loss": 2.5901,
      "theoretical_loss": 3.4054265929358016,
      "tokens_seen": 2184773632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437777777777778,
      "loss": 2.4351,
      "theoretical_loss": 3.4054182231551415,
      "tokens_seen": 2184839168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034375757575757577,
      "loss": 2.5164,
      "theoretical_loss": 3.40540985369583,
      "tokens_seen": 2184904704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437373737373737,
      "loss": 2.519,
      "theoretical_loss": 3.4054014845578453,
      "tokens_seen": 2184970240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437171717171717,
      "loss": 2.6878,
      "theoretical_loss": 3.4053931157411648,
      "tokens_seen": 2185035776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034369696969696975,
      "loss": 2.6094,
      "theoretical_loss": 3.405384747245767,
      "tokens_seen": 2185101312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436767676767677,
      "loss": 2.4693,
      "theoretical_loss": 3.405376379071629,
      "tokens_seen": 2185166848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034365656565656566,
      "loss": 2.4004,
      "theoretical_loss": 3.40536801121873,
      "tokens_seen": 2185232384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436363636363636,
      "loss": 2.6808,
      "theoretical_loss": 3.405359643687048,
      "tokens_seen": 2185297920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034361616161616163,
      "loss": 2.4997,
      "theoretical_loss": 3.4053512764765603,
      "tokens_seen": 2185363456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034359595959595964,
      "loss": 2.448,
      "theoretical_loss": 3.405342909587245,
      "tokens_seen": 2185428992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435757575757576,
      "loss": 2.6548,
      "theoretical_loss": 3.4053345430190807,
      "tokens_seen": 2185494528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034355555555555555,
      "loss": 2.3814,
      "theoretical_loss": 3.405326176772045,
      "tokens_seen": 2185560064
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435353535353535,
      "loss": 2.5529,
      "theoretical_loss": 3.405317810846116,
      "tokens_seen": 2185625600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435151515151515,
      "loss": 2.5558,
      "theoretical_loss": 3.405309445241272,
      "tokens_seen": 2185691136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034349494949494953,
      "loss": 2.4149,
      "theoretical_loss": 3.4053010799574914,
      "tokens_seen": 2185756672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434747474747475,
      "loss": 2.6333,
      "theoretical_loss": 3.405292714994751,
      "tokens_seen": 2185822208
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1205294,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4200518131256104,
      "objective/train/theoretical_loss": 3.40528435035303,
      "objective/train/tokens_used": 544746976,
      "theoretical_loss": 3.40528435035303,
      "tokens_seen": 2185887744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034345454545454544,
      "loss": 2.4917,
      "theoretical_loss": 3.40528435035303,
      "tokens_seen": 2185887744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034343434343434346,
      "loss": 2.7631,
      "theoretical_loss": 3.405275986032306,
      "tokens_seen": 2185953280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434141414141414,
      "loss": 2.5811,
      "theoretical_loss": 3.4052676220325573,
      "tokens_seen": 2186018816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433939393939394,
      "loss": 2.6179,
      "theoretical_loss": 3.4052592583537615,
      "tokens_seen": 2186084352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433737373737374,
      "loss": 2.6252,
      "theoretical_loss": 3.405250894995897,
      "tokens_seen": 2186149888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034335353535353534,
      "loss": 2.8562,
      "theoretical_loss": 3.4052425319589417,
      "tokens_seen": 2186215424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034333333333333335,
      "loss": 2.4894,
      "theoretical_loss": 3.405234169242874,
      "tokens_seen": 2186280960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433131313131313,
      "loss": 2.5904,
      "theoretical_loss": 3.4052258068476715,
      "tokens_seen": 2186346496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432929292929293,
      "loss": 2.6932,
      "theoretical_loss": 3.405217444773313,
      "tokens_seen": 2186412032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432727272727273,
      "loss": 2.3908,
      "theoretical_loss": 3.4052090830197757,
      "tokens_seen": 2186477568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432525252525253,
      "loss": 2.6912,
      "theoretical_loss": 3.405200721587038,
      "tokens_seen": 2186543104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034323232323232324,
      "loss": 2.5519,
      "theoretical_loss": 3.405192360475078,
      "tokens_seen": 2186608640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432121212121212,
      "loss": 2.783,
      "theoretical_loss": 3.405183999683874,
      "tokens_seen": 2186674176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431919191919192,
      "loss": 2.5782,
      "theoretical_loss": 3.405175639213404,
      "tokens_seen": 2186739712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034317171717171717,
      "loss": 2.3531,
      "theoretical_loss": 3.4051672790636456,
      "tokens_seen": 2186805248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431515151515152,
      "loss": 2.6121,
      "theoretical_loss": 3.4051589192345775,
      "tokens_seen": 2186870784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034313131313131313,
      "loss": 2.5944,
      "theoretical_loss": 3.4051505597261778,
      "tokens_seen": 2186936320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431111111111111,
      "loss": 2.507,
      "theoretical_loss": 3.405142200538424,
      "tokens_seen": 2187001856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430909090909091,
      "loss": 2.7761,
      "theoretical_loss": 3.405133841671295,
      "tokens_seen": 2187067392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430707070707071,
      "loss": 2.6267,
      "theoretical_loss": 3.405125483124768,
      "tokens_seen": 2187132928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034305050505050507,
      "loss": 2.5586,
      "theoretical_loss": 3.4051171248988217,
      "tokens_seen": 2187198464
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343030303030303,
      "loss": 2.4967,
      "theoretical_loss": 3.4051087669934343,
      "tokens_seen": 2187264000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343010101010101,
      "loss": 2.5115,
      "theoretical_loss": 3.405100409408583,
      "tokens_seen": 2187329536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342989898989899,
      "loss": 2.5764,
      "theoretical_loss": 3.405092052144247,
      "tokens_seen": 2187395072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342969696969697,
      "loss": 2.6482,
      "theoretical_loss": 3.405083695200404,
      "tokens_seen": 2187460608
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1205951,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2435500621795654,
      "objective/train/theoretical_loss": 3.405075338577032,
      "objective/train/tokens_used": 546385376,
      "theoretical_loss": 3.405075338577032,
      "tokens_seen": 2187526144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034294949494949496,
      "loss": 2.0967,
      "theoretical_loss": 3.405075338577032,
      "tokens_seen": 2187526144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429292929292929,
      "loss": 2.4133,
      "theoretical_loss": 3.405066982274109,
      "tokens_seen": 2187591680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429090909090909,
      "loss": 2.478,
      "theoretical_loss": 3.4050586262916136,
      "tokens_seen": 2187657216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034288888888888894,
      "loss": 2.5769,
      "theoretical_loss": 3.405050270629523,
      "tokens_seen": 2187722752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428686868686869,
      "loss": 2.441,
      "theoretical_loss": 3.4050419152878164,
      "tokens_seen": 2187788288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034284848484848486,
      "loss": 2.5369,
      "theoretical_loss": 3.4050335602664714,
      "tokens_seen": 2187853824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428282828282828,
      "loss": 2.6225,
      "theoretical_loss": 3.405025205565466,
      "tokens_seen": 2187919360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034280808080808077,
      "loss": 2.4762,
      "theoretical_loss": 3.4050168511847785,
      "tokens_seen": 2187984896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034278787878787883,
      "loss": 2.482,
      "theoretical_loss": 3.405008497124387,
      "tokens_seen": 2188050432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427676767676768,
      "loss": 2.6874,
      "theoretical_loss": 3.4050001433842696,
      "tokens_seen": 2188115968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034274747474747475,
      "loss": 2.7971,
      "theoretical_loss": 3.404991789964405,
      "tokens_seen": 2188181504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427272727272727,
      "loss": 2.7184,
      "theoretical_loss": 3.4049834368647702,
      "tokens_seen": 2188247040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427070707070707,
      "loss": 2.633,
      "theoretical_loss": 3.404975084085344,
      "tokens_seen": 2188312576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426868686868687,
      "loss": 2.6838,
      "theoretical_loss": 3.4049667316261045,
      "tokens_seen": 2188378112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426666666666667,
      "loss": 2.5053,
      "theoretical_loss": 3.40495837948703,
      "tokens_seen": 2188443648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034264646464646464,
      "loss": 2.4909,
      "theoretical_loss": 3.404950027668098,
      "tokens_seen": 2188509184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426262626262626,
      "loss": 2.6741,
      "theoretical_loss": 3.4049416761692872,
      "tokens_seen": 2188574720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426060606060606,
      "loss": 2.6215,
      "theoretical_loss": 3.404933324990576,
      "tokens_seen": 2188640256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425858585858586,
      "loss": 2.4817,
      "theoretical_loss": 3.4049249741319416,
      "tokens_seen": 2188705792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425656565656566,
      "loss": 2.6325,
      "theoretical_loss": 3.4049166235933632,
      "tokens_seen": 2188771328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034254545454545453,
      "loss": 2.4636,
      "theoretical_loss": 3.4049082733748186,
      "tokens_seen": 2188836864
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034252525252525254,
      "loss": 2.5871,
      "theoretical_loss": 3.404899923476285,
      "tokens_seen": 2188902400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425050505050505,
      "loss": 2.6458,
      "theoretical_loss": 3.4048915738977423,
      "tokens_seen": 2188967936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424848484848485,
      "loss": 2.6708,
      "theoretical_loss": 3.4048832246391676,
      "tokens_seen": 2189033472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034246464646464647,
      "loss": 2.3265,
      "theoretical_loss": 3.4048748757005387,
      "tokens_seen": 2189099008
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1207093,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.545804262161255,
      "objective/train/theoretical_loss": 3.4048665270818343,
      "objective/train/tokens_used": 548023776,
      "theoretical_loss": 3.4048665270818343,
      "tokens_seen": 2189164544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424444444444444,
      "loss": 2.6491,
      "theoretical_loss": 3.4048665270818343,
      "tokens_seen": 2189164544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034242424242424244,
      "loss": 2.376,
      "theoretical_loss": 3.4048581787830328,
      "tokens_seen": 2189230080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424040404040404,
      "loss": 2.4601,
      "theoretical_loss": 3.404849830804112,
      "tokens_seen": 2189295616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423838383838384,
      "loss": 2.5238,
      "theoretical_loss": 3.4048414831450504,
      "tokens_seen": 2189361152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034236363636363636,
      "loss": 2.7257,
      "theoretical_loss": 3.4048331358058257,
      "tokens_seen": 2189426688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423434343434343,
      "loss": 2.5514,
      "theoretical_loss": 3.4048247887864163,
      "tokens_seen": 2189492224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034232323232323233,
      "loss": 2.5212,
      "theoretical_loss": 3.4048164420868003,
      "tokens_seen": 2189557760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423030303030303,
      "loss": 2.7112,
      "theoretical_loss": 3.404808095706956,
      "tokens_seen": 2189623296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422828282828283,
      "loss": 2.4652,
      "theoretical_loss": 3.4047997496468616,
      "tokens_seen": 2189688832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034226262626262625,
      "loss": 2.3664,
      "theoretical_loss": 3.4047914039064953,
      "tokens_seen": 2189754368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034224242424242427,
      "loss": 2.5596,
      "theoretical_loss": 3.404783058485835,
      "tokens_seen": 2189819904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422222222222222,
      "loss": 2.6098,
      "theoretical_loss": 3.404774713384859,
      "tokens_seen": 2189885440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034220202020202023,
      "loss": 2.4988,
      "theoretical_loss": 3.404766368603546,
      "tokens_seen": 2189950976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421818181818182,
      "loss": 2.5537,
      "theoretical_loss": 3.404758024141873,
      "tokens_seen": 2190016512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034216161616161615,
      "loss": 2.584,
      "theoretical_loss": 3.40474967999982,
      "tokens_seen": 2190082048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034214141414141416,
      "loss": 2.5137,
      "theoretical_loss": 3.4047413361773633,
      "tokens_seen": 2190147584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421212121212121,
      "loss": 2.7074,
      "theoretical_loss": 3.4047329926744823,
      "tokens_seen": 2190213120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421010101010101,
      "loss": 2.6088,
      "theoretical_loss": 3.4047246494911545,
      "tokens_seen": 2190278656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420808080808081,
      "loss": 2.4021,
      "theoretical_loss": 3.4047163066273587,
      "tokens_seen": 2190344192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420606060606061,
      "loss": 2.5995,
      "theoretical_loss": 3.404707964083073,
      "tokens_seen": 2190409728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034204040404040405,
      "loss": 2.5752,
      "theoretical_loss": 3.404699621858275,
      "tokens_seen": 2190475264
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342020202020202,
      "loss": 2.8434,
      "theoretical_loss": 3.4046912799529436,
      "tokens_seen": 2190540800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342,
      "loss": 2.4992,
      "theoretical_loss": 3.404682938367057,
      "tokens_seen": 2190606336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341979797979798,
      "loss": 2.5464,
      "theoretical_loss": 3.404674597100593,
      "tokens_seen": 2190671872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341959595959596,
      "loss": 2.6982,
      "theoretical_loss": 3.4046662561535297,
      "tokens_seen": 2190737408
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1208190,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.561566114425659,
      "objective/train/theoretical_loss": 3.4046579155258456,
      "objective/train/tokens_used": 549662176,
      "theoretical_loss": 3.4046579155258456,
      "tokens_seen": 2190802944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034193939393939394,
      "loss": 2.5418,
      "theoretical_loss": 3.4046579155258456,
      "tokens_seen": 2190802944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419191919191919,
      "loss": 2.2125,
      "theoretical_loss": 3.4046495752175194,
      "tokens_seen": 2190868480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418989898989899,
      "loss": 2.6678,
      "theoretical_loss": 3.4046412352285285,
      "tokens_seen": 2190934016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418787878787879,
      "loss": 2.4224,
      "theoretical_loss": 3.4046328955588514,
      "tokens_seen": 2190999552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418585858585859,
      "loss": 2.5616,
      "theoretical_loss": 3.4046245562084665,
      "tokens_seen": 2191065088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034183838383838384,
      "loss": 2.5746,
      "theoretical_loss": 3.404616217177352,
      "tokens_seen": 2191130624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418181818181818,
      "loss": 2.6096,
      "theoretical_loss": 3.404607878465486,
      "tokens_seen": 2191196160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417979797979798,
      "loss": 2.3851,
      "theoretical_loss": 3.4045995400728466,
      "tokens_seen": 2191261696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417777777777778,
      "loss": 2.5008,
      "theoretical_loss": 3.4045912019994122,
      "tokens_seen": 2191327232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034175757575757577,
      "loss": 2.4656,
      "theoretical_loss": 3.404582864245161,
      "tokens_seen": 2191392768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034173737373737373,
      "loss": 2.4672,
      "theoretical_loss": 3.4045745268100718,
      "tokens_seen": 2191458304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417171717171717,
      "loss": 2.6212,
      "theoretical_loss": 3.4045661896941217,
      "tokens_seen": 2191523840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034169696969696975,
      "loss": 2.4015,
      "theoretical_loss": 3.40455785289729,
      "tokens_seen": 2191589376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416767676767677,
      "loss": 2.591,
      "theoretical_loss": 3.404549516419554,
      "tokens_seen": 2191654912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034165656565656566,
      "loss": 2.7445,
      "theoretical_loss": 3.404541180260893,
      "tokens_seen": 2191720448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416363636363636,
      "loss": 2.4693,
      "theoretical_loss": 3.404532844421284,
      "tokens_seen": 2191785984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416161616161616,
      "loss": 2.2786,
      "theoretical_loss": 3.404524508900707,
      "tokens_seen": 2191851520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034159595959595964,
      "loss": 2.6081,
      "theoretical_loss": 3.404516173699138,
      "tokens_seen": 2191917056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415757575757576,
      "loss": 2.5328,
      "theoretical_loss": 3.404507838816557,
      "tokens_seen": 2191982592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034155555555555556,
      "loss": 2.6958,
      "theoretical_loss": 3.4044995042529416,
      "tokens_seen": 2192048128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415353535353535,
      "loss": 2.8967,
      "theoretical_loss": 3.40449117000827,
      "tokens_seen": 2192113664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415151515151515,
      "loss": 2.4535,
      "theoretical_loss": 3.404482836082521,
      "tokens_seen": 2192179200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034149494949494954,
      "loss": 2.4954,
      "theoretical_loss": 3.404474502475672,
      "tokens_seen": 2192244736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414747474747475,
      "loss": 2.5461,
      "theoretical_loss": 3.4044661691877023,
      "tokens_seen": 2192310272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034145454545454545,
      "loss": 2.8149,
      "theoretical_loss": 3.4044578362185893,
      "tokens_seen": 2192375808
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1208841,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7217459678649902,
      "objective/train/theoretical_loss": 3.4044495035683116,
      "objective/train/tokens_used": 551300576,
      "theoretical_loss": 3.4044495035683116,
      "tokens_seen": 2192441344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414343434343434,
      "loss": 2.5229,
      "theoretical_loss": 3.4044495035683116,
      "tokens_seen": 2192441344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414141414141414,
      "loss": 2.7207,
      "theoretical_loss": 3.404441171236847,
      "tokens_seen": 2192506880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034139393939393943,
      "loss": 2.8168,
      "theoretical_loss": 3.404432839224175,
      "tokens_seen": 2192572416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413737373737374,
      "loss": 2.67,
      "theoretical_loss": 3.4044245075302726,
      "tokens_seen": 2192637952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034135353535353534,
      "loss": 2.4568,
      "theoretical_loss": 3.404416176155119,
      "tokens_seen": 2192703488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034133333333333335,
      "loss": 2.649,
      "theoretical_loss": 3.4044078450986914,
      "tokens_seen": 2192769024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413131313131313,
      "loss": 2.4792,
      "theoretical_loss": 3.404399514360969,
      "tokens_seen": 2192834560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412929292929293,
      "loss": 2.4585,
      "theoretical_loss": 3.40439118394193,
      "tokens_seen": 2192900096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412727272727273,
      "loss": 2.6644,
      "theoretical_loss": 3.4043828538415526,
      "tokens_seen": 2192965632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034125252525252524,
      "loss": 2.7125,
      "theoretical_loss": 3.4043745240598144,
      "tokens_seen": 2193031168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034123232323232325,
      "loss": 2.8195,
      "theoretical_loss": 3.4043661945966948,
      "tokens_seen": 2193096704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412121212121212,
      "loss": 2.6109,
      "theoretical_loss": 3.4043578654521713,
      "tokens_seen": 2193162240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411919191919192,
      "loss": 2.5872,
      "theoretical_loss": 3.404349536626223,
      "tokens_seen": 2193227776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034117171717171717,
      "loss": 2.6027,
      "theoretical_loss": 3.404341208118827,
      "tokens_seen": 2193293312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411515151515152,
      "loss": 2.457,
      "theoretical_loss": 3.4043328799299624,
      "tokens_seen": 2193358848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034113131313131314,
      "loss": 2.6272,
      "theoretical_loss": 3.4043245520596077,
      "tokens_seen": 2193424384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411111111111111,
      "loss": 2.5235,
      "theoretical_loss": 3.4043162245077405,
      "tokens_seen": 2193489920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410909090909091,
      "loss": 2.6016,
      "theoretical_loss": 3.4043078972743395,
      "tokens_seen": 2193555456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034107070707070706,
      "loss": 2.6477,
      "theoretical_loss": 3.4042995703593832,
      "tokens_seen": 2193620992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410505050505051,
      "loss": 2.4979,
      "theoretical_loss": 3.4042912437628496,
      "tokens_seen": 2193686528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034103030303030303,
      "loss": 2.5399,
      "theoretical_loss": 3.404282917484717,
      "tokens_seen": 2193752064
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341010101010101,
      "loss": 2.6218,
      "theoretical_loss": 3.4042745915249637,
      "tokens_seen": 2193817600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000340989898989899,
      "loss": 2.5061,
      "theoretical_loss": 3.4042662658835683,
      "tokens_seen": 2193883136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000340969696969697,
      "loss": 2.6345,
      "theoretical_loss": 3.404257940560509,
      "tokens_seen": 2193948672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034094949494949497,
      "loss": 2.7469,
      "theoretical_loss": 3.4042496155557638,
      "tokens_seen": 2194014208
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 1210098,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.358370542526245,
      "objective/train/theoretical_loss": 3.4042412908693116,
      "objective/train/tokens_used": 552938976,
      "theoretical_loss": 3.4042412908693116,
      "tokens_seen": 2194079744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003409292929292929,
      "loss": 2.4407,
      "theoretical_loss": 3.4042412908693116,
      "tokens_seen": 2194079744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003409090909090909,
      "loss": 2.5013,
      "theoretical_loss": 3.40423296650113,
      "tokens_seen": 2194145280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408888888888889,
      "loss": 2.3507,
      "theoretical_loss": 3.404224642451198,
      "tokens_seen": 2194210816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408686868686869,
      "loss": 2.6394,
      "theoretical_loss": 3.404216318719494,
      "tokens_seen": 2194276352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034084848484848486,
      "loss": 2.6258,
      "theoretical_loss": 3.4042079953059954,
      "tokens_seen": 2194341888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408282828282828,
      "loss": 2.5948,
      "theoretical_loss": 3.4041996722106815,
      "tokens_seen": 2194407424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408080808080808,
      "loss": 2.5158,
      "theoretical_loss": 3.40419134943353,
      "tokens_seen": 2194472960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034078787878787884,
      "loss": 2.4552,
      "theoretical_loss": 3.4041830269745197,
      "tokens_seen": 2194538496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003407676767676768,
      "loss": 2.6235,
      "theoretical_loss": 3.4041747048336286,
      "tokens_seen": 2194604032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034074747474747475,
      "loss": 2.5909,
      "theoretical_loss": 3.404166383010835,
      "tokens_seen": 2194669568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003407272727272727,
      "loss": 2.7844,
      "theoretical_loss": 3.4041580615061178,
      "tokens_seen": 2194735104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034070707070707067,
      "loss": 2.4736,
      "theoretical_loss": 3.404149740319455,
      "tokens_seen": 2194800640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034068686868686873,
      "loss": 2.7426,
      "theoretical_loss": 3.4041414194508244,
      "tokens_seen": 2194866176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406666666666667,
      "loss": 2.5016,
      "theoretical_loss": 3.404133098900205,
      "tokens_seen": 2194931712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034064646464646465,
      "loss": 2.4223,
      "theoretical_loss": 3.4041247786675752,
      "tokens_seen": 2194997248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406262626262626,
      "loss": 2.6775,
      "theoretical_loss": 3.404116458752913,
      "tokens_seen": 2195062784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034060606060606056,
      "loss": 2.7619,
      "theoretical_loss": 3.404108139156197,
      "tokens_seen": 2195128320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405858585858586,
      "loss": 2.5595,
      "theoretical_loss": 3.4040998198774055,
      "tokens_seen": 2195193856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405656565656566,
      "loss": 2.4847,
      "theoretical_loss": 3.4040915009165174,
      "tokens_seen": 2195259392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034054545454545454,
      "loss": 2.5576,
      "theoretical_loss": 3.4040831822735096,
      "tokens_seen": 2195324928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405252525252525,
      "loss": 2.6036,
      "theoretical_loss": 3.404074863948362,
      "tokens_seen": 2195390464
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034050505050505056,
      "loss": 2.7201,
      "theoretical_loss": 3.4040665459410517,
      "tokens_seen": 2195456000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404848484848485,
      "loss": 2.6916,
      "theoretical_loss": 3.404058228251558,
      "tokens_seen": 2195521536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404646464646465,
      "loss": 2.5531,
      "theoretical_loss": 3.404049910879859,
      "tokens_seen": 2195587072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034044444444444443,
      "loss": 2.4956,
      "theoretical_loss": 3.404041593825933,
      "tokens_seen": 2195652608
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1211496,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.74421763420105,
      "objective/train/theoretical_loss": 3.404033277089759,
      "objective/train/tokens_used": 554577376,
      "theoretical_loss": 3.404033277089759,
      "tokens_seen": 2195718144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404242424242424,
      "loss": 2.6424,
      "theoretical_loss": 3.404033277089759,
      "tokens_seen": 2195718144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034040404040404045,
      "loss": 2.6356,
      "theoretical_loss": 3.4040249606713138,
      "tokens_seen": 2195783680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403838383838384,
      "loss": 2.5341,
      "theoretical_loss": 3.4040166445705773,
      "tokens_seen": 2195849216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034036363636363637,
      "loss": 2.7132,
      "theoretical_loss": 3.404008328787527,
      "tokens_seen": 2195914752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403434343434343,
      "loss": 2.6677,
      "theoretical_loss": 3.4040000133221424,
      "tokens_seen": 2195980288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034032323232323234,
      "loss": 2.4513,
      "theoretical_loss": 3.4039916981744005,
      "tokens_seen": 2196045824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034030303030303035,
      "loss": 2.4427,
      "theoretical_loss": 3.4039833833442805,
      "tokens_seen": 2196111360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402828282828283,
      "loss": 2.5379,
      "theoretical_loss": 3.4039750688317603,
      "tokens_seen": 2196176896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034026262626262626,
      "loss": 2.4633,
      "theoretical_loss": 3.403966754636819,
      "tokens_seen": 2196242432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402424242424242,
      "loss": 2.3333,
      "theoretical_loss": 3.4039584407594345,
      "tokens_seen": 2196307968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034022222222222223,
      "loss": 2.4106,
      "theoretical_loss": 3.403950127199585,
      "tokens_seen": 2196373504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034020202020202024,
      "loss": 2.76,
      "theoretical_loss": 3.4039418139572493,
      "tokens_seen": 2196439040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401818181818182,
      "loss": 2.6414,
      "theoretical_loss": 3.403933501032406,
      "tokens_seen": 2196504576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034016161616161615,
      "loss": 2.4559,
      "theoretical_loss": 3.403925188425033,
      "tokens_seen": 2196570112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034014141414141416,
      "loss": 2.3674,
      "theoretical_loss": 3.4039168761351086,
      "tokens_seen": 2196635648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401212121212121,
      "loss": 2.4677,
      "theoretical_loss": 3.4039085641626117,
      "tokens_seen": 2196701184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034010101010101013,
      "loss": 2.3859,
      "theoretical_loss": 3.4039002525075204,
      "tokens_seen": 2196766720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400808080808081,
      "loss": 2.5589,
      "theoretical_loss": 3.4038919411698134,
      "tokens_seen": 2196832256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034006060606060605,
      "loss": 2.6093,
      "theoretical_loss": 3.403883630149469,
      "tokens_seen": 2196897792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034004040404040406,
      "loss": 2.5632,
      "theoretical_loss": 3.403875319446465,
      "tokens_seen": 2196963328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340020202020202,
      "loss": 2.4322,
      "theoretical_loss": 3.403867009060781,
      "tokens_seen": 2197028864
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034,
      "loss": 2.8422,
      "theoretical_loss": 3.4038586989923942,
      "tokens_seen": 2197094400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339979797979798,
      "loss": 2.6952,
      "theoretical_loss": 3.4038503892412844,
      "tokens_seen": 2197159936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339959595959596,
      "loss": 2.4847,
      "theoretical_loss": 3.4038420798074283,
      "tokens_seen": 2197225472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033993939393939395,
      "loss": 2.551,
      "theoretical_loss": 3.403833770690806,
      "tokens_seen": 2197291008
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1212273,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0346760749816895,
      "objective/train/theoretical_loss": 3.403825461891395,
      "objective/train/tokens_used": 556215776,
      "theoretical_loss": 3.403825461891395,
      "tokens_seen": 2197356544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399191919191919,
      "loss": 2.7611,
      "theoretical_loss": 3.403825461891395,
      "tokens_seen": 2197356544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398989898989899,
      "loss": 2.4756,
      "theoretical_loss": 3.4038171534091735,
      "tokens_seen": 2197422080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398787878787879,
      "loss": 2.6328,
      "theoretical_loss": 3.4038088452441206,
      "tokens_seen": 2197487616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398585858585859,
      "loss": 2.7409,
      "theoretical_loss": 3.403800537396214,
      "tokens_seen": 2197553152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033983838383838384,
      "loss": 2.5518,
      "theoretical_loss": 3.4037922298654335,
      "tokens_seen": 2197618688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398181818181818,
      "loss": 2.6347,
      "theoretical_loss": 3.403783922651756,
      "tokens_seen": 2197684224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397979797979798,
      "loss": 2.4669,
      "theoretical_loss": 3.403775615755161,
      "tokens_seen": 2197749760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397777777777778,
      "loss": 2.5773,
      "theoretical_loss": 3.403767309175626,
      "tokens_seen": 2197815296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397575757575758,
      "loss": 2.6887,
      "theoretical_loss": 3.4037590029131306,
      "tokens_seen": 2197880832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033973737373737373,
      "loss": 2.4716,
      "theoretical_loss": 3.403750696967652,
      "tokens_seen": 2197946368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397171717171717,
      "loss": 2.5483,
      "theoretical_loss": 3.40374239133917,
      "tokens_seen": 2198011904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396969696969697,
      "loss": 2.4754,
      "theoretical_loss": 3.4037340860276615,
      "tokens_seen": 2198077440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396767676767677,
      "loss": 2.6292,
      "theoretical_loss": 3.4037257810331063,
      "tokens_seen": 2198142976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033965656565656567,
      "loss": 2.6845,
      "theoretical_loss": 3.403717476355482,
      "tokens_seen": 2198208512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033963636363636363,
      "loss": 2.6357,
      "theoretical_loss": 3.403709171994768,
      "tokens_seen": 2198274048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396161616161616,
      "loss": 2.3748,
      "theoretical_loss": 3.4037008679509415,
      "tokens_seen": 2198339584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033959595959595965,
      "loss": 2.4924,
      "theoretical_loss": 3.4036925642239817,
      "tokens_seen": 2198405120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395757575757576,
      "loss": 2.7771,
      "theoretical_loss": 3.4036842608138667,
      "tokens_seen": 2198470656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033955555555555556,
      "loss": 2.506,
      "theoretical_loss": 3.403675957720576,
      "tokens_seen": 2198536192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395353535353535,
      "loss": 2.4778,
      "theoretical_loss": 3.4036676549440865,
      "tokens_seen": 2198601728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395151515151515,
      "loss": 2.5482,
      "theoretical_loss": 3.403659352484378,
      "tokens_seen": 2198667264
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033949494949494954,
      "loss": 2.5968,
      "theoretical_loss": 3.4036510503414283,
      "tokens_seen": 2198732800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394747474747475,
      "loss": 2.6466,
      "theoretical_loss": 3.403642748515216,
      "tokens_seen": 2198798336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033945454545454546,
      "loss": 2.4165,
      "theoretical_loss": 3.40363444700572,
      "tokens_seen": 2198863872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394343434343434,
      "loss": 2.4413,
      "theoretical_loss": 3.403626145812918,
      "tokens_seen": 2198929408
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1213496,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7383503913879395,
      "objective/train/theoretical_loss": 3.4036178449367887,
      "objective/train/tokens_used": 557854176,
      "theoretical_loss": 3.4036178449367887,
      "tokens_seen": 2198994944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394141414141414,
      "loss": 2.5066,
      "theoretical_loss": 3.4036178449367887,
      "tokens_seen": 2198994944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033939393939393943,
      "loss": 2.4764,
      "theoretical_loss": 3.4036095443773107,
      "tokens_seen": 2199060480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393737373737374,
      "loss": 2.5683,
      "theoretical_loss": 3.4036012441344625,
      "tokens_seen": 2199126016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033935353535353535,
      "loss": 2.591,
      "theoretical_loss": 3.403592944208223,
      "tokens_seen": 2199191552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393333333333333,
      "loss": 2.5321,
      "theoretical_loss": 3.40358464459857,
      "tokens_seen": 2199257088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393131313131313,
      "loss": 2.2204,
      "theoretical_loss": 3.403576345305482,
      "tokens_seen": 2199322624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033929292929292933,
      "loss": 2.6112,
      "theoretical_loss": 3.4035680463289384,
      "tokens_seen": 2199388160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392727272727273,
      "loss": 2.5665,
      "theoretical_loss": 3.4035597476689166,
      "tokens_seen": 2199453696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033925252525252524,
      "loss": 2.5351,
      "theoretical_loss": 3.4035514493253958,
      "tokens_seen": 2199519232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033923232323232325,
      "loss": 2.7012,
      "theoretical_loss": 3.4035431512983543,
      "tokens_seen": 2199584768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392121212121212,
      "loss": 2.5129,
      "theoretical_loss": 3.4035348535877707,
      "tokens_seen": 2199650304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391919191919192,
      "loss": 2.7113,
      "theoretical_loss": 3.403526556193623,
      "tokens_seen": 2199715840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391717171717172,
      "loss": 2.4473,
      "theoretical_loss": 3.40351825911589,
      "tokens_seen": 2199781376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033915151515151513,
      "loss": 2.5923,
      "theoretical_loss": 3.403509962354551,
      "tokens_seen": 2199846912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033913131313131315,
      "loss": 2.6168,
      "theoretical_loss": 3.403501665909583,
      "tokens_seen": 2199912448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391111111111111,
      "loss": 2.5153,
      "theoretical_loss": 3.403493369780966,
      "tokens_seen": 2199977984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003390909090909091,
      "loss": 2.5009,
      "theoretical_loss": 3.4034850739686773,
      "tokens_seen": 2200043520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033907070707070707,
      "loss": 2.6043,
      "theoretical_loss": 3.403476778472696,
      "tokens_seen": 2200109056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339050505050505,
      "loss": 2.2239,
      "theoretical_loss": 3.4034684832930004,
      "tokens_seen": 2200174592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033903030303030304,
      "loss": 2.4125,
      "theoretical_loss": 3.4034601884295697,
      "tokens_seen": 2200240128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339010101010101,
      "loss": 2.661,
      "theoretical_loss": 3.4034518938823815,
      "tokens_seen": 2200305664
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338989898989899,
      "loss": 2.6988,
      "theoretical_loss": 3.403443599651415,
      "tokens_seen": 2200371200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033896969696969696,
      "loss": 2.4141,
      "theoretical_loss": 3.4034353057366484,
      "tokens_seen": 2200436736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338949494949495,
      "loss": 2.3495,
      "theoretical_loss": 3.40342701213806,
      "tokens_seen": 2200502272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033892929292929293,
      "loss": 2.6577,
      "theoretical_loss": 3.4034187188556286,
      "tokens_seen": 2200567808
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1214048,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5324034690856934,
      "objective/train/theoretical_loss": 3.403410425889333,
      "objective/train/tokens_used": 559492576,
      "theoretical_loss": 3.403410425889333,
      "tokens_seen": 2200633344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003389090909090909,
      "loss": 2.6052,
      "theoretical_loss": 3.403410425889333,
      "tokens_seen": 2200633344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388888888888889,
      "loss": 2.8732,
      "theoretical_loss": 3.4034021332391515,
      "tokens_seen": 2200698880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033886868686868686,
      "loss": 2.3591,
      "theoretical_loss": 3.4033938409050624,
      "tokens_seen": 2200764416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033884848484848487,
      "loss": 2.3755,
      "theoretical_loss": 3.4033855488870444,
      "tokens_seen": 2200829952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388282828282828,
      "loss": 2.5591,
      "theoretical_loss": 3.403377257185076,
      "tokens_seen": 2200895488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388080808080808,
      "loss": 2.4537,
      "theoretical_loss": 3.403368965799136,
      "tokens_seen": 2200961024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387878787878788,
      "loss": 2.6963,
      "theoretical_loss": 3.4033606747292033,
      "tokens_seen": 2201026560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387676767676768,
      "loss": 2.5424,
      "theoretical_loss": 3.4033523839752555,
      "tokens_seen": 2201092096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033874747474747476,
      "loss": 2.6247,
      "theoretical_loss": 3.403344093537272,
      "tokens_seen": 2201157632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387272727272727,
      "loss": 2.7775,
      "theoretical_loss": 3.40333580341523,
      "tokens_seen": 2201223168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387070707070707,
      "loss": 2.404,
      "theoretical_loss": 3.4033275136091095,
      "tokens_seen": 2201288704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386868686868687,
      "loss": 2.528,
      "theoretical_loss": 3.4033192241188885,
      "tokens_seen": 2201354240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386666666666667,
      "loss": 2.6364,
      "theoretical_loss": 3.4033109349445456,
      "tokens_seen": 2201419776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033864646464646465,
      "loss": 2.7123,
      "theoretical_loss": 3.4033026460860594,
      "tokens_seen": 2201485312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386262626262626,
      "loss": 2.6844,
      "theoretical_loss": 3.4032943575434085,
      "tokens_seen": 2201550848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386060606060606,
      "loss": 2.433,
      "theoretical_loss": 3.403286069316571,
      "tokens_seen": 2201616384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033858585858585863,
      "loss": 2.8952,
      "theoretical_loss": 3.4032777814055266,
      "tokens_seen": 2201681920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385656565656566,
      "loss": 2.6094,
      "theoretical_loss": 3.4032694938102526,
      "tokens_seen": 2201747456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033854545454545454,
      "loss": 2.6436,
      "theoretical_loss": 3.403261206530728,
      "tokens_seen": 2201812992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385252525252525,
      "loss": 2.4752,
      "theoretical_loss": 3.4032529195669317,
      "tokens_seen": 2201878528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385050505050505,
      "loss": 2.4291,
      "theoretical_loss": 3.4032446329188417,
      "tokens_seen": 2201944064
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384848484848485,
      "loss": 2.9939,
      "theoretical_loss": 3.4032363465864375,
      "tokens_seen": 2202009600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384646464646465,
      "loss": 2.618,
      "theoretical_loss": 3.4032280605696967,
      "tokens_seen": 2202075136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033844444444444444,
      "loss": 2.5891,
      "theoretical_loss": 3.403219774868598,
      "tokens_seen": 2202140672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384242424242424,
      "loss": 2.612,
      "theoretical_loss": 3.403211489483121,
      "tokens_seen": 2202206208
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1214686,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2311205863952637,
      "objective/train/theoretical_loss": 3.403203204413243,
      "objective/train/tokens_used": 561130976,
      "theoretical_loss": 3.403203204413243,
      "tokens_seen": 2202271744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033840404040404046,
      "loss": 2.316,
      "theoretical_loss": 3.403203204413243,
      "tokens_seen": 2202271744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383838383838384,
      "loss": 2.7264,
      "theoretical_loss": 3.4031949196589433,
      "tokens_seen": 2202337280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383636363636364,
      "loss": 2.5682,
      "theoretical_loss": 3.4031866352202003,
      "tokens_seen": 2202402816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033834343434343433,
      "loss": 2.6374,
      "theoretical_loss": 3.4031783510969924,
      "tokens_seen": 2202468352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383232323232323,
      "loss": 2.6837,
      "theoretical_loss": 3.403170067289299,
      "tokens_seen": 2202533888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033830303030303035,
      "loss": 2.4823,
      "theoretical_loss": 3.4031617837970973,
      "tokens_seen": 2202599424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382828282828283,
      "loss": 2.4551,
      "theoretical_loss": 3.403153500620367,
      "tokens_seen": 2202664960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033826262626262627,
      "loss": 2.7068,
      "theoretical_loss": 3.4031452177590866,
      "tokens_seen": 2202730496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382424242424242,
      "loss": 2.3841,
      "theoretical_loss": 3.4031369352132343,
      "tokens_seen": 2202796032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033822222222222223,
      "loss": 2.6533,
      "theoretical_loss": 3.4031286529827893,
      "tokens_seen": 2202861568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033820202020202024,
      "loss": 2.5688,
      "theoretical_loss": 3.4031203710677294,
      "tokens_seen": 2202927104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381818181818182,
      "loss": 2.5709,
      "theoretical_loss": 3.4031120894680336,
      "tokens_seen": 2202992640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033816161616161616,
      "loss": 2.6249,
      "theoretical_loss": 3.4031038081836806,
      "tokens_seen": 2203058176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381414141414141,
      "loss": 2.5922,
      "theoretical_loss": 3.4030955272146493,
      "tokens_seen": 2203123712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381212121212121,
      "loss": 2.4562,
      "theoretical_loss": 3.403087246560917,
      "tokens_seen": 2203189248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033810101010101014,
      "loss": 2.4766,
      "theoretical_loss": 3.4030789662224645,
      "tokens_seen": 2203254784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003380808080808081,
      "loss": 2.4682,
      "theoretical_loss": 3.4030706861992686,
      "tokens_seen": 2203320320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033806060606060605,
      "loss": 2.7268,
      "theoretical_loss": 3.4030624064913084,
      "tokens_seen": 2203385856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033804040404040406,
      "loss": 2.3349,
      "theoretical_loss": 3.403054127098563,
      "tokens_seen": 2203451392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338020202020202,
      "loss": 2.449,
      "theoretical_loss": 3.40304584802101,
      "tokens_seen": 2203516928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033800000000000003,
      "loss": 2.6137,
      "theoretical_loss": 3.403037569258629,
      "tokens_seen": 2203582464
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337979797979798,
      "loss": 2.3822,
      "theoretical_loss": 3.4030292908113986,
      "tokens_seen": 2203648000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033795959595959594,
      "loss": 2.521,
      "theoretical_loss": 3.403021012679297,
      "tokens_seen": 2203713536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033793939393939395,
      "loss": 2.4525,
      "theoretical_loss": 3.403012734862303,
      "tokens_seen": 2203779072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379191919191919,
      "loss": 2.4292,
      "theoretical_loss": 3.403004457360395,
      "tokens_seen": 2203844608
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1215685,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3675143718719482,
      "objective/train/theoretical_loss": 3.402996180173552,
      "objective/train/tokens_used": 562769376,
      "theoretical_loss": 3.402996180173552,
      "tokens_seen": 2203910144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378989898989899,
      "loss": 2.612,
      "theoretical_loss": 3.402996180173552,
      "tokens_seen": 2203910144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378787878787879,
      "loss": 2.6328,
      "theoretical_loss": 3.4029879033017525,
      "tokens_seen": 2203975680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378585858585859,
      "loss": 2.5141,
      "theoretical_loss": 3.4029796267449752,
      "tokens_seen": 2204041216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033783838383838385,
      "loss": 2.5341,
      "theoretical_loss": 3.4029713505031984,
      "tokens_seen": 2204106752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378181818181818,
      "loss": 2.698,
      "theoretical_loss": 3.4029630745764012,
      "tokens_seen": 2204172288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377979797979798,
      "loss": 2.671,
      "theoretical_loss": 3.4029547989645623,
      "tokens_seen": 2204237824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033777777777777777,
      "loss": 2.6546,
      "theoretical_loss": 3.4029465236676595,
      "tokens_seen": 2204303360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377575757575758,
      "loss": 2.5107,
      "theoretical_loss": 3.4029382486856727,
      "tokens_seen": 2204368896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033773737373737374,
      "loss": 2.5914,
      "theoretical_loss": 3.40292997401858,
      "tokens_seen": 2204434432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377171717171717,
      "loss": 2.4237,
      "theoretical_loss": 3.4029216996663596,
      "tokens_seen": 2204499968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376969696969697,
      "loss": 2.5588,
      "theoretical_loss": 3.4029134256289906,
      "tokens_seen": 2204565504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376767676767677,
      "loss": 2.5284,
      "theoretical_loss": 3.4029051519064515,
      "tokens_seen": 2204631040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376565656565657,
      "loss": 2.3153,
      "theoretical_loss": 3.402896878498721,
      "tokens_seen": 2204696576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033763636363636363,
      "loss": 2.8618,
      "theoretical_loss": 3.402888605405778,
      "tokens_seen": 2204762112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376161616161616,
      "loss": 2.607,
      "theoretical_loss": 3.402880332627601,
      "tokens_seen": 2204827648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375959595959596,
      "loss": 2.4707,
      "theoretical_loss": 3.4028720601641687,
      "tokens_seen": 2204893184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375757575757576,
      "loss": 2.5004,
      "theoretical_loss": 3.4028637880154595,
      "tokens_seen": 2204958720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033755555555555557,
      "loss": 2.6331,
      "theoretical_loss": 3.4028555161814524,
      "tokens_seen": 2205024256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375353535353535,
      "loss": 2.4841,
      "theoretical_loss": 3.4028472446621265,
      "tokens_seen": 2205089792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375151515151515,
      "loss": 2.4584,
      "theoretical_loss": 3.4028389734574596,
      "tokens_seen": 2205155328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033749494949494955,
      "loss": 2.7482,
      "theoretical_loss": 3.4028307025674303,
      "tokens_seen": 2205220864
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374747474747475,
      "loss": 2.3796,
      "theoretical_loss": 3.402822431992018,
      "tokens_seen": 2205286400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033745454545454546,
      "loss": 2.6053,
      "theoretical_loss": 3.402814161731201,
      "tokens_seen": 2205351936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374343434343434,
      "loss": 2.5552,
      "theoretical_loss": 3.4028058917849586,
      "tokens_seen": 2205417472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374141414141414,
      "loss": 2.416,
      "theoretical_loss": 3.4027976221532685,
      "tokens_seen": 2205483008
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1216619,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1585769653320312,
      "objective/train/theoretical_loss": 3.40278935283611,
      "objective/train/tokens_used": 564407776,
      "theoretical_loss": 3.40278935283611,
      "tokens_seen": 2205548544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033739393939393944,
      "loss": 2.3125,
      "theoretical_loss": 3.40278935283611,
      "tokens_seen": 2205548544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373737373737374,
      "loss": 2.8216,
      "theoretical_loss": 3.402781083833462,
      "tokens_seen": 2205614080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033735353535353535,
      "loss": 2.5437,
      "theoretical_loss": 3.402772815145302,
      "tokens_seen": 2205679616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373333333333333,
      "loss": 2.3755,
      "theoretical_loss": 3.4027645467716106,
      "tokens_seen": 2205745152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033731313131313127,
      "loss": 2.5188,
      "theoretical_loss": 3.4027562787123644,
      "tokens_seen": 2205810688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033729292929292933,
      "loss": 2.4873,
      "theoretical_loss": 3.4027480109675436,
      "tokens_seen": 2205876224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372727272727273,
      "loss": 2.5146,
      "theoretical_loss": 3.4027397435371265,
      "tokens_seen": 2205941760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033725252525252525,
      "loss": 2.6188,
      "theoretical_loss": 3.402731476421092,
      "tokens_seen": 2206007296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372323232323232,
      "loss": 2.4189,
      "theoretical_loss": 3.4027232096194178,
      "tokens_seen": 2206072832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372121212121212,
      "loss": 2.684,
      "theoretical_loss": 3.402714943132084,
      "tokens_seen": 2206138368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371919191919192,
      "loss": 2.6022,
      "theoretical_loss": 3.4027066769590686,
      "tokens_seen": 2206203904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371717171717172,
      "loss": 2.6802,
      "theoretical_loss": 3.40269841110035,
      "tokens_seen": 2206269440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033715151515151514,
      "loss": 2.6851,
      "theoretical_loss": 3.4026901455559075,
      "tokens_seen": 2206334976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371313131313131,
      "loss": 2.46,
      "theoretical_loss": 3.40268188032572,
      "tokens_seen": 2206400512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371111111111111,
      "loss": 2.5575,
      "theoretical_loss": 3.4026736154097654,
      "tokens_seen": 2206466048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370909090909091,
      "loss": 2.4875,
      "theoretical_loss": 3.402665350808023,
      "tokens_seen": 2206531584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370707070707071,
      "loss": 2.2709,
      "theoretical_loss": 3.4026570865204713,
      "tokens_seen": 2206597120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033705050505050503,
      "loss": 2.6232,
      "theoretical_loss": 3.4026488225470892,
      "tokens_seen": 2206662656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033703030303030304,
      "loss": 2.7487,
      "theoretical_loss": 3.402640558887855,
      "tokens_seen": 2206728192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033701010101010105,
      "loss": 2.3714,
      "theoretical_loss": 3.402632295542748,
      "tokens_seen": 2206793728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336989898989899,
      "loss": 2.7876,
      "theoretical_loss": 3.402624032511747,
      "tokens_seen": 2206859264
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033696969696969697,
      "loss": 2.3925,
      "theoretical_loss": 3.40261576979483,
      "tokens_seen": 2206924800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369494949494949,
      "loss": 2.511,
      "theoretical_loss": 3.402607507391976,
      "tokens_seen": 2206990336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033692929292929294,
      "loss": 2.6294,
      "theoretical_loss": 3.4025992453031644,
      "tokens_seen": 2207055872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033690909090909095,
      "loss": 2.6245,
      "theoretical_loss": 3.402590983528373,
      "tokens_seen": 2207121408
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1217122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3241126537323,
      "objective/train/theoretical_loss": 3.402582722067581,
      "objective/train/tokens_used": 566046176,
      "theoretical_loss": 3.402582722067581,
      "tokens_seen": 2207186944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368888888888889,
      "loss": 2.5174,
      "theoretical_loss": 3.402582722067581,
      "tokens_seen": 2207186944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033686868686868686,
      "loss": 2.6031,
      "theoretical_loss": 3.4025744609207673,
      "tokens_seen": 2207252480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033684848484848487,
      "loss": 2.6602,
      "theoretical_loss": 3.4025662000879104,
      "tokens_seen": 2207318016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033682828282828283,
      "loss": 2.4869,
      "theoretical_loss": 3.402557939568989,
      "tokens_seen": 2207383552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033680808080808084,
      "loss": 2.4876,
      "theoretical_loss": 3.4025496793639824,
      "tokens_seen": 2207449088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367878787878788,
      "loss": 2.7501,
      "theoretical_loss": 3.4025414194728683,
      "tokens_seen": 2207514624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033676767676767675,
      "loss": 2.3752,
      "theoretical_loss": 3.4025331598956265,
      "tokens_seen": 2207580160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033674747474747476,
      "loss": 2.7045,
      "theoretical_loss": 3.402524900632235,
      "tokens_seen": 2207645696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367272727272727,
      "loss": 2.6499,
      "theoretical_loss": 3.402516641682673,
      "tokens_seen": 2207711232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033670707070707073,
      "loss": 2.6278,
      "theoretical_loss": 3.4025083830469196,
      "tokens_seen": 2207776768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366868686868687,
      "loss": 2.5024,
      "theoretical_loss": 3.4025001247249524,
      "tokens_seen": 2207842304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366666666666667,
      "loss": 2.5961,
      "theoretical_loss": 3.402491866716751,
      "tokens_seen": 2207907840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033664646464646466,
      "loss": 2.3533,
      "theoretical_loss": 3.402483609022294,
      "tokens_seen": 2207973376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366262626262626,
      "loss": 2.5906,
      "theoretical_loss": 3.4024753516415607,
      "tokens_seen": 2208038912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366060606060606,
      "loss": 2.6582,
      "theoretical_loss": 3.402467094574529,
      "tokens_seen": 2208104448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365858585858586,
      "loss": 2.6773,
      "theoretical_loss": 3.402458837821178,
      "tokens_seen": 2208169984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365656565656566,
      "loss": 2.4831,
      "theoretical_loss": 3.4024505813814865,
      "tokens_seen": 2208235520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033654545454545455,
      "loss": 2.2427,
      "theoretical_loss": 3.402442325255433,
      "tokens_seen": 2208301056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365252525252525,
      "loss": 2.6192,
      "theoretical_loss": 3.402434069442997,
      "tokens_seen": 2208366592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365050505050505,
      "loss": 2.6418,
      "theoretical_loss": 3.402425813944157,
      "tokens_seen": 2208432128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033648484848484853,
      "loss": 2.5576,
      "theoretical_loss": 3.4024175587588914,
      "tokens_seen": 2208497664
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364646464646465,
      "loss": 2.6503,
      "theoretical_loss": 3.402409303887179,
      "tokens_seen": 2208563200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033644444444444444,
      "loss": 2.4943,
      "theoretical_loss": 3.402401049328999,
      "tokens_seen": 2208628736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364242424242424,
      "loss": 2.4334,
      "theoretical_loss": 3.40239279508433,
      "tokens_seen": 2208694272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364040404040404,
      "loss": 2.694,
      "theoretical_loss": 3.402384541153151,
      "tokens_seen": 2208759808
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1218187,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.472285509109497,
      "objective/train/theoretical_loss": 3.40237628753544,
      "objective/train/tokens_used": 567684576,
      "theoretical_loss": 3.40237628753544,
      "tokens_seen": 2208825344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363838383838384,
      "loss": 2.5674,
      "theoretical_loss": 3.40237628753544,
      "tokens_seen": 2208825344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363636363636364,
      "loss": 2.4357,
      "theoretical_loss": 3.402368034231177,
      "tokens_seen": 2208890880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033634343434343434,
      "loss": 2.5027,
      "theoretical_loss": 3.4023597812403397,
      "tokens_seen": 2208956416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363232323232323,
      "loss": 2.57,
      "theoretical_loss": 3.4023515285629076,
      "tokens_seen": 2209021952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033630303030303036,
      "loss": 2.6849,
      "theoretical_loss": 3.402343276198859,
      "tokens_seen": 2209087488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362828282828283,
      "loss": 2.6881,
      "theoretical_loss": 3.402335024148173,
      "tokens_seen": 2209153024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033626262626262627,
      "loss": 2.4768,
      "theoretical_loss": 3.4023267724108286,
      "tokens_seen": 2209218560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033624242424242423,
      "loss": 2.4945,
      "theoretical_loss": 3.4023185209868045,
      "tokens_seen": 2209284096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362222222222222,
      "loss": 2.6198,
      "theoretical_loss": 3.402310269876079,
      "tokens_seen": 2209349632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033620202020202025,
      "loss": 2.716,
      "theoretical_loss": 3.4023020190786317,
      "tokens_seen": 2209415168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361818181818182,
      "loss": 2.6395,
      "theoretical_loss": 3.4022937685944408,
      "tokens_seen": 2209480704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033616161616161616,
      "loss": 2.6683,
      "theoretical_loss": 3.4022855184234855,
      "tokens_seen": 2209546240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361414141414141,
      "loss": 2.4532,
      "theoretical_loss": 3.402277268565744,
      "tokens_seen": 2209611776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033612121212121213,
      "loss": 2.7676,
      "theoretical_loss": 3.4022690190211957,
      "tokens_seen": 2209677312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033610101010101014,
      "loss": 2.6974,
      "theoretical_loss": 3.4022607697898195,
      "tokens_seen": 2209742848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360808080808081,
      "loss": 2.4207,
      "theoretical_loss": 3.402252520871594,
      "tokens_seen": 2209808384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033606060606060606,
      "loss": 2.4476,
      "theoretical_loss": 3.4022442722664983,
      "tokens_seen": 2209873920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336040404040404,
      "loss": 2.5548,
      "theoretical_loss": 3.4022360239745106,
      "tokens_seen": 2209939456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336020202020202,
      "loss": 2.4813,
      "theoretical_loss": 3.40222777599561,
      "tokens_seen": 2210004992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033600000000000004,
      "loss": 2.5984,
      "theoretical_loss": 3.4022195283297756,
      "tokens_seen": 2210070528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000335979797979798,
      "loss": 2.7201,
      "theoretical_loss": 3.402211280976986,
      "tokens_seen": 2210136064
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033595959595959595,
      "loss": 2.4198,
      "theoretical_loss": 3.4022030339372202,
      "tokens_seen": 2210201600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033593939393939396,
      "loss": 2.6951,
      "theoretical_loss": 3.402194787210457,
      "tokens_seen": 2210267136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003359191919191919,
      "loss": 2.6134,
      "theoretical_loss": 3.402186540796675,
      "tokens_seen": 2210332672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033589898989898993,
      "loss": 2.4243,
      "theoretical_loss": 3.402178294695853,
      "tokens_seen": 2210398208
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 1219601,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.494659900665283,
      "objective/train/theoretical_loss": 3.4021700489079705,
      "objective/train/tokens_used": 569322976,
      "theoretical_loss": 3.4021700489079705,
      "tokens_seen": 2210463744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358787878787879,
      "loss": 2.4444,
      "theoretical_loss": 3.4021700489079705,
      "tokens_seen": 2210463744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033585858585858584,
      "loss": 2.5409,
      "theoretical_loss": 3.402161803433006,
      "tokens_seen": 2210529280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033583838383838385,
      "loss": 2.5478,
      "theoretical_loss": 3.402153558270938,
      "tokens_seen": 2210594816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358181818181818,
      "loss": 2.6713,
      "theoretical_loss": 3.4021453134217454,
      "tokens_seen": 2210660352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357979797979798,
      "loss": 2.5541,
      "theoretical_loss": 3.4021370688854073,
      "tokens_seen": 2210725888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357777777777778,
      "loss": 2.7063,
      "theoretical_loss": 3.402128824661903,
      "tokens_seen": 2210791424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033575757575757573,
      "loss": 2.4704,
      "theoretical_loss": 3.4021205807512107,
      "tokens_seen": 2210856960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033573737373737375,
      "loss": 2.583,
      "theoretical_loss": 3.402112337153309,
      "tokens_seen": 2210922496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357171717171717,
      "loss": 2.8334,
      "theoretical_loss": 3.4021040938681777,
      "tokens_seen": 2210988032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356969696969697,
      "loss": 2.5484,
      "theoretical_loss": 3.402095850895795,
      "tokens_seen": 2211053568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033567676767676767,
      "loss": 2.6949,
      "theoretical_loss": 3.40208760823614,
      "tokens_seen": 2211119104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356565656565657,
      "loss": 2.58,
      "theoretical_loss": 3.4020793658891915,
      "tokens_seen": 2211184640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033563636363636364,
      "loss": 2.5433,
      "theoretical_loss": 3.402071123854928,
      "tokens_seen": 2211250176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356161616161616,
      "loss": 2.7101,
      "theoretical_loss": 3.402062882133329,
      "tokens_seen": 2211315712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355959595959596,
      "loss": 2.5147,
      "theoretical_loss": 3.4020546407243732,
      "tokens_seen": 2211381248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033557575757575756,
      "loss": 2.4719,
      "theoretical_loss": 3.402046399628039,
      "tokens_seen": 2211446784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355555555555556,
      "loss": 2.6339,
      "theoretical_loss": 3.4020381588443063,
      "tokens_seen": 2211512320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033553535353535353,
      "loss": 2.4807,
      "theoretical_loss": 3.402029918373153,
      "tokens_seen": 2211577856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355151515151515,
      "loss": 2.4844,
      "theoretical_loss": 3.402021678214558,
      "tokens_seen": 2211643392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354949494949495,
      "loss": 2.6246,
      "theoretical_loss": 3.402013438368501,
      "tokens_seen": 2211708928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354747474747475,
      "loss": 2.3825,
      "theoretical_loss": 3.40200519883496,
      "tokens_seen": 2211774464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033545454545454547,
      "loss": 2.3494,
      "theoretical_loss": 3.401996959613914,
      "tokens_seen": 2211840000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354343434343434,
      "loss": 2.5031,
      "theoretical_loss": 3.4019887207053428,
      "tokens_seen": 2211905536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354141414141414,
      "loss": 2.484,
      "theoretical_loss": 3.4019804821092245,
      "tokens_seen": 2211971072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353939393939394,
      "loss": 2.884,
      "theoretical_loss": 3.401972243825538,
      "tokens_seen": 2212036608
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1220207,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6348462104797363,
      "objective/train/theoretical_loss": 3.4019640058542624,
      "objective/train/tokens_used": 570961376,
      "theoretical_loss": 3.4019640058542624,
      "tokens_seen": 2212102144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353737373737374,
      "loss": 2.3554,
      "theoretical_loss": 3.4019640058542624,
      "tokens_seen": 2212102144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033535353535353536,
      "loss": 2.5459,
      "theoretical_loss": 3.4019557681953767,
      "tokens_seen": 2212167680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353333333333333,
      "loss": 2.4922,
      "theoretical_loss": 3.4019475308488594,
      "tokens_seen": 2212233216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033531313131313133,
      "loss": 2.7239,
      "theoretical_loss": 3.4019392938146895,
      "tokens_seen": 2212298752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033529292929292934,
      "loss": 2.414,
      "theoretical_loss": 3.4019310570928463,
      "tokens_seen": 2212364288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352727272727273,
      "loss": 2.8008,
      "theoretical_loss": 3.4019228206833088,
      "tokens_seen": 2212429824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033525252525252525,
      "loss": 2.6254,
      "theoretical_loss": 3.4019145845860552,
      "tokens_seen": 2212495360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352323232323232,
      "loss": 2.5691,
      "theoretical_loss": 3.401906348801065,
      "tokens_seen": 2212560896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352121212121212,
      "loss": 2.8517,
      "theoretical_loss": 3.401898113328316,
      "tokens_seen": 2212626432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033519191919191923,
      "loss": 2.6577,
      "theoretical_loss": 3.401889878167789,
      "tokens_seen": 2212691968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351717171717172,
      "loss": 2.5093,
      "theoretical_loss": 3.4018816433194616,
      "tokens_seen": 2212757504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033515151515151515,
      "loss": 2.5333,
      "theoretical_loss": 3.401873408783313,
      "tokens_seen": 2212823040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351313131313131,
      "loss": 2.7045,
      "theoretical_loss": 3.401865174559322,
      "tokens_seen": 2212888576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033511111111111117,
      "loss": 2.5791,
      "theoretical_loss": 3.401856940647468,
      "tokens_seen": 2212954112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350909090909091,
      "loss": 2.4296,
      "theoretical_loss": 3.401848707047729,
      "tokens_seen": 2213019648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350707070707071,
      "loss": 2.6218,
      "theoretical_loss": 3.4018404737600854,
      "tokens_seen": 2213085184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033505050505050504,
      "loss": 2.6225,
      "theoretical_loss": 3.4018322407845147,
      "tokens_seen": 2213150720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000335030303030303,
      "loss": 2.6526,
      "theoretical_loss": 3.4018240081209963,
      "tokens_seen": 2213216256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033501010101010106,
      "loss": 2.6658,
      "theoretical_loss": 3.4018157757695096,
      "tokens_seen": 2213281792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334989898989899,
      "loss": 2.3678,
      "theoretical_loss": 3.401807543730033,
      "tokens_seen": 2213347328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334969696969697,
      "loss": 2.5685,
      "theoretical_loss": 3.401799312002545,
      "tokens_seen": 2213412864
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033494949494949493,
      "loss": 2.6017,
      "theoretical_loss": 3.401791080587026,
      "tokens_seen": 2213478400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033492929292929294,
      "loss": 2.6115,
      "theoretical_loss": 3.4017828494834537,
      "tokens_seen": 2213543936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033490909090909095,
      "loss": 2.5889,
      "theoretical_loss": 3.4017746186918076,
      "tokens_seen": 2213609472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348888888888889,
      "loss": 2.6011,
      "theoretical_loss": 3.4017663882120663,
      "tokens_seen": 2213675008
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1221489,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4770944118499756,
      "objective/train/theoretical_loss": 3.401758158044209,
      "objective/train/tokens_used": 572599776,
      "theoretical_loss": 3.401758158044209,
      "tokens_seen": 2213740544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033486868686868687,
      "loss": 2.4942,
      "theoretical_loss": 3.401758158044209,
      "tokens_seen": 2213740544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348484848484848,
      "loss": 2.5457,
      "theoretical_loss": 3.4017499281882144,
      "tokens_seen": 2213806080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033482828282828283,
      "loss": 2.5708,
      "theoretical_loss": 3.4017416986440616,
      "tokens_seen": 2213871616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033480808080808085,
      "loss": 2.5185,
      "theoretical_loss": 3.4017334694117296,
      "tokens_seen": 2213937152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347878787878788,
      "loss": 2.5026,
      "theoretical_loss": 3.4017252404911975,
      "tokens_seen": 2214002688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033476767676767676,
      "loss": 2.5462,
      "theoretical_loss": 3.4017170118824436,
      "tokens_seen": 2214068224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033474747474747477,
      "loss": 2.5386,
      "theoretical_loss": 3.4017087835854474,
      "tokens_seen": 2214133760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347272727272727,
      "loss": 2.5107,
      "theoretical_loss": 3.401700555600188,
      "tokens_seen": 2214199296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033470707070707074,
      "loss": 2.3553,
      "theoretical_loss": 3.401692327926644,
      "tokens_seen": 2214264832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346868686868687,
      "loss": 2.5025,
      "theoretical_loss": 3.401684100564795,
      "tokens_seen": 2214330368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033466666666666665,
      "loss": 2.4847,
      "theoretical_loss": 3.4016758735146193,
      "tokens_seen": 2214395904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033464646464646466,
      "loss": 2.5301,
      "theoretical_loss": 3.4016676467760956,
      "tokens_seen": 2214461440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346262626262626,
      "loss": 2.5679,
      "theoretical_loss": 3.401659420349204,
      "tokens_seen": 2214526976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033460606060606063,
      "loss": 2.5658,
      "theoretical_loss": 3.4016511942339225,
      "tokens_seen": 2214592512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345858585858586,
      "loss": 2.453,
      "theoretical_loss": 3.40164296843023,
      "tokens_seen": 2214658048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345656565656566,
      "loss": 2.5215,
      "theoretical_loss": 3.401634742938106,
      "tokens_seen": 2214723584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033454545454545456,
      "loss": 2.5998,
      "theoretical_loss": 3.40162651775753,
      "tokens_seen": 2214789120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345252525252525,
      "loss": 2.7373,
      "theoretical_loss": 3.4016182928884797,
      "tokens_seen": 2214854656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345050505050505,
      "loss": 2.2561,
      "theoretical_loss": 3.4016100683309345,
      "tokens_seen": 2214920192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344848484848485,
      "loss": 2.4015,
      "theoretical_loss": 3.4016018440848743,
      "tokens_seen": 2214985728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344646464646465,
      "loss": 2.746,
      "theoretical_loss": 3.401593620150277,
      "tokens_seen": 2215051264
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033444444444444445,
      "loss": 2.3898,
      "theoretical_loss": 3.4015853965271217,
      "tokens_seen": 2215116800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344242424242424,
      "loss": 2.5943,
      "theoretical_loss": 3.401577173215388,
      "tokens_seen": 2215182336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344040404040404,
      "loss": 2.4772,
      "theoretical_loss": 3.4015689502150543,
      "tokens_seen": 2215247872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033438383838383843,
      "loss": 2.5286,
      "theoretical_loss": 3.4015607275260997,
      "tokens_seen": 2215313408
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1222176,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.791827440261841,
      "objective/train/theoretical_loss": 3.401552505148504,
      "objective/train/tokens_used": 574238176,
      "theoretical_loss": 3.401552505148504,
      "tokens_seen": 2215378944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003343636363636364,
      "loss": 2.6742,
      "theoretical_loss": 3.401552505148504,
      "tokens_seen": 2215378944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033434343434343434,
      "loss": 2.4913,
      "theoretical_loss": 3.401544283082245,
      "tokens_seen": 2215444480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003343232323232323,
      "loss": 2.6004,
      "theoretical_loss": 3.4015360613273025,
      "tokens_seen": 2215510016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003343030303030303,
      "loss": 2.3775,
      "theoretical_loss": 3.401527839883655,
      "tokens_seen": 2215575552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342828282828283,
      "loss": 2.4678,
      "theoretical_loss": 3.4015196187512817,
      "tokens_seen": 2215641088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342626262626263,
      "loss": 2.4456,
      "theoretical_loss": 3.401511397930162,
      "tokens_seen": 2215706624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033424242424242423,
      "loss": 2.5144,
      "theoretical_loss": 3.4015031774202744,
      "tokens_seen": 2215772160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342222222222222,
      "loss": 2.2735,
      "theoretical_loss": 3.401494957221598,
      "tokens_seen": 2215837696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033420202020202026,
      "loss": 2.4292,
      "theoretical_loss": 3.401486737334112,
      "tokens_seen": 2215903232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341818181818182,
      "loss": 2.4665,
      "theoretical_loss": 3.4014785177577953,
      "tokens_seen": 2215968768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033416161616161617,
      "loss": 2.4636,
      "theoretical_loss": 3.401470298492627,
      "tokens_seen": 2216034304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341414141414141,
      "loss": 2.5201,
      "theoretical_loss": 3.401462079538586,
      "tokens_seen": 2216099840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341212121212121,
      "loss": 2.6508,
      "theoretical_loss": 3.4014538608956513,
      "tokens_seen": 2216165376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033410101010101015,
      "loss": 2.7864,
      "theoretical_loss": 3.401445642563802,
      "tokens_seen": 2216230912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340808080808081,
      "loss": 2.6858,
      "theoretical_loss": 3.4014374245430172,
      "tokens_seen": 2216296448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033406060606060606,
      "loss": 2.4874,
      "theoretical_loss": 3.401429206833276,
      "tokens_seen": 2216361984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334040404040404,
      "loss": 2.3966,
      "theoretical_loss": 3.4014209894345573,
      "tokens_seen": 2216427520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334020202020202,
      "loss": 2.562,
      "theoretical_loss": 3.40141277234684,
      "tokens_seen": 2216493056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033400000000000004,
      "loss": 2.7305,
      "theoretical_loss": 3.401404555570103,
      "tokens_seen": 2216558592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333979797979798,
      "loss": 2.4205,
      "theoretical_loss": 3.401396339104326,
      "tokens_seen": 2216624128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033395959595959595,
      "loss": 2.6097,
      "theoretical_loss": 3.4013881229494873,
      "tokens_seen": 2216689664
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339393939393939,
      "loss": 2.5154,
      "theoretical_loss": 3.401379907105566,
      "tokens_seen": 2216755200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339191919191919,
      "loss": 2.604,
      "theoretical_loss": 3.401371691572542,
      "tokens_seen": 2216820736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033389898989898993,
      "loss": 2.5253,
      "theoretical_loss": 3.4013634763503937,
      "tokens_seen": 2216886272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338787878787879,
      "loss": 2.4062,
      "theoretical_loss": 3.4013552614391003,
      "tokens_seen": 2216951808
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1223215,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9110116958618164,
      "objective/train/theoretical_loss": 3.40134704683864,
      "objective/train/tokens_used": 575876576,
      "theoretical_loss": 3.40134704683864,
      "tokens_seen": 2217017344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033385858585858585,
      "loss": 2.5671,
      "theoretical_loss": 3.40134704683864,
      "tokens_seen": 2217017344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338383838383838,
      "loss": 2.771,
      "theoretical_loss": 3.4013388325489933,
      "tokens_seen": 2217082880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338181818181818,
      "loss": 2.5389,
      "theoretical_loss": 3.4013306185701384,
      "tokens_seen": 2217148416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337979797979798,
      "loss": 2.7053,
      "theoretical_loss": 3.4013224049020545,
      "tokens_seen": 2217213952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337777777777778,
      "loss": 2.5321,
      "theoretical_loss": 3.4013141915447207,
      "tokens_seen": 2217279488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033375757575757574,
      "loss": 2.5079,
      "theoretical_loss": 3.401305978498116,
      "tokens_seen": 2217345024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033373737373737375,
      "loss": 2.5116,
      "theoretical_loss": 3.4012977657622194,
      "tokens_seen": 2217410560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337171717171717,
      "loss": 2.6053,
      "theoretical_loss": 3.40128955333701,
      "tokens_seen": 2217476096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336969696969697,
      "loss": 2.498,
      "theoretical_loss": 3.401281341222467,
      "tokens_seen": 2217541632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336767676767677,
      "loss": 2.6724,
      "theoretical_loss": 3.4012731294185694,
      "tokens_seen": 2217607168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033365656565656563,
      "loss": 2.5151,
      "theoretical_loss": 3.4012649179252965,
      "tokens_seen": 2217672704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033363636363636364,
      "loss": 2.6096,
      "theoretical_loss": 3.4012567067426267,
      "tokens_seen": 2217738240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033361616161616166,
      "loss": 2.7375,
      "theoretical_loss": 3.4012484958705396,
      "tokens_seen": 2217803776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335959595959596,
      "loss": 2.3421,
      "theoretical_loss": 3.4012402853090142,
      "tokens_seen": 2217869312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033357575757575757,
      "loss": 2.5651,
      "theoretical_loss": 3.40123207505803,
      "tokens_seen": 2217934848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335555555555556,
      "loss": 2.5675,
      "theoretical_loss": 3.401223865117565,
      "tokens_seen": 2218000384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033353535353535354,
      "loss": 2.5096,
      "theoretical_loss": 3.401215655487599,
      "tokens_seen": 2218065920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033351515151515155,
      "loss": 2.7468,
      "theoretical_loss": 3.401207446168111,
      "tokens_seen": 2218131456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334949494949495,
      "loss": 2.5889,
      "theoretical_loss": 3.4011992371590805,
      "tokens_seen": 2218196992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033347474747474746,
      "loss": 2.639,
      "theoretical_loss": 3.401191028460486,
      "tokens_seen": 2218262528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033345454545454547,
      "loss": 2.392,
      "theoretical_loss": 3.4011828200723064,
      "tokens_seen": 2218328064
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033343434343434343,
      "loss": 2.6493,
      "theoretical_loss": 3.4011746119945214,
      "tokens_seen": 2218393600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033341414141414144,
      "loss": 2.5369,
      "theoretical_loss": 3.40116640422711,
      "tokens_seen": 2218459136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333939393939394,
      "loss": 2.5511,
      "theoretical_loss": 3.4011581967700506,
      "tokens_seen": 2218524672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333737373737374,
      "loss": 2.7442,
      "theoretical_loss": 3.401149989623323,
      "tokens_seen": 2218590208
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1224236,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.486609935760498,
      "objective/train/theoretical_loss": 3.4011417827869064,
      "objective/train/tokens_used": 577514976,
      "theoretical_loss": 3.4011417827869064,
      "tokens_seen": 2218655744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033335353535353537,
      "loss": 2.6855,
      "theoretical_loss": 3.4011417827869064,
      "tokens_seen": 2218655744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333333333333333,
      "loss": 2.6871,
      "theoretical_loss": 3.4011335762607797,
      "tokens_seen": 2218721280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033331313131313133,
      "loss": 2.5971,
      "theoretical_loss": 3.4011253700449218,
      "tokens_seen": 2218786816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332929292929293,
      "loss": 2.4757,
      "theoretical_loss": 3.401117164139312,
      "tokens_seen": 2218852352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332727272727273,
      "loss": 2.6521,
      "theoretical_loss": 3.4011089585439294,
      "tokens_seen": 2218917888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033325252525252526,
      "loss": 2.6123,
      "theoretical_loss": 3.4011007532587527,
      "tokens_seen": 2218983424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332323232323232,
      "loss": 2.6107,
      "theoretical_loss": 3.4010925482837617,
      "tokens_seen": 2219048960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332121212121212,
      "loss": 2.6139,
      "theoretical_loss": 3.401084343618935,
      "tokens_seen": 2219114496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033319191919191924,
      "loss": 2.6066,
      "theoretical_loss": 3.401076139264252,
      "tokens_seen": 2219180032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331717171717172,
      "loss": 2.475,
      "theoretical_loss": 3.4010679352196918,
      "tokens_seen": 2219245568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033315151515151515,
      "loss": 2.5528,
      "theoretical_loss": 3.4010597314852333,
      "tokens_seen": 2219311104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331313131313131,
      "loss": 2.704,
      "theoretical_loss": 3.4010515280608558,
      "tokens_seen": 2219376640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331111111111111,
      "loss": 2.4247,
      "theoretical_loss": 3.401043324946538,
      "tokens_seen": 2219442176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033309090909090913,
      "loss": 2.7846,
      "theoretical_loss": 3.40103512214226,
      "tokens_seen": 2219507712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003330707070707071,
      "loss": 2.826,
      "theoretical_loss": 3.4010269196480003,
      "tokens_seen": 2219573248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033305050505050504,
      "loss": 2.6933,
      "theoretical_loss": 3.401018717463738,
      "tokens_seen": 2219638784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333030303030303,
      "loss": 2.6265,
      "theoretical_loss": 3.4010105155894523,
      "tokens_seen": 2219704320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033301010101010107,
      "loss": 2.4605,
      "theoretical_loss": 3.401002314025122,
      "tokens_seen": 2219769856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332989898989899,
      "loss": 2.7084,
      "theoretical_loss": 3.400994112770727,
      "tokens_seen": 2219835392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332969696969697,
      "loss": 2.6849,
      "theoretical_loss": 3.4009859118262455,
      "tokens_seen": 2219900928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033294949494949494,
      "loss": 2.57,
      "theoretical_loss": 3.4009777111916577,
      "tokens_seen": 2219966464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329292929292929,
      "loss": 2.3423,
      "theoretical_loss": 3.400969510866942,
      "tokens_seen": 2220032000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033290909090909096,
      "loss": 2.6705,
      "theoretical_loss": 3.4009613108520775,
      "tokens_seen": 2220097536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328888888888889,
      "loss": 2.5765,
      "theoretical_loss": 3.400953111147044,
      "tokens_seen": 2220163072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033286868686868687,
      "loss": 2.7474,
      "theoretical_loss": 3.40094491175182,
      "tokens_seen": 2220228608
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1224894,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.333254337310791,
      "objective/train/theoretical_loss": 3.4009367126663848,
      "objective/train/tokens_used": 579153376,
      "theoretical_loss": 3.4009367126663848,
      "tokens_seen": 2220294144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033284848484848483,
      "loss": 2.6027,
      "theoretical_loss": 3.4009367126663848,
      "tokens_seen": 2220294144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033282828282828284,
      "loss": 2.6271,
      "theoretical_loss": 3.400928513890718,
      "tokens_seen": 2220359680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033280808080808085,
      "loss": 2.6168,
      "theoretical_loss": 3.4009203154247976,
      "tokens_seen": 2220425216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327878787878788,
      "loss": 2.7739,
      "theoretical_loss": 3.400912117268604,
      "tokens_seen": 2220490752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033276767676767676,
      "loss": 2.774,
      "theoretical_loss": 3.4009039194221162,
      "tokens_seen": 2220556288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327474747474747,
      "loss": 2.5009,
      "theoretical_loss": 3.4008957218853126,
      "tokens_seen": 2220621824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033272727272727273,
      "loss": 2.8622,
      "theoretical_loss": 3.4008875246581733,
      "tokens_seen": 2220687360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033270707070707074,
      "loss": 2.7762,
      "theoretical_loss": 3.4008793277406766,
      "tokens_seen": 2220752896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326868686868687,
      "loss": 2.5999,
      "theoretical_loss": 3.400871131132802,
      "tokens_seen": 2220818432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033266666666666666,
      "loss": 2.4279,
      "theoretical_loss": 3.400862934834529,
      "tokens_seen": 2220883968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033264646464646467,
      "loss": 2.5061,
      "theoretical_loss": 3.4008547388458363,
      "tokens_seen": 2220949504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326262626262626,
      "loss": 2.5058,
      "theoretical_loss": 3.400846543166703,
      "tokens_seen": 2221015040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033260606060606064,
      "loss": 2.3584,
      "theoretical_loss": 3.400838347797109,
      "tokens_seen": 2221080576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325858585858586,
      "loss": 2.3845,
      "theoretical_loss": 3.4008301527370333,
      "tokens_seen": 2221146112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033256565656565655,
      "loss": 2.5764,
      "theoretical_loss": 3.4008219579864543,
      "tokens_seen": 2221211648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033254545454545456,
      "loss": 2.646,
      "theoretical_loss": 3.4008137635453517,
      "tokens_seen": 2221277184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325252525252525,
      "loss": 2.5971,
      "theoretical_loss": 3.4008055694137047,
      "tokens_seen": 2221342720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033250505050505053,
      "loss": 2.6967,
      "theoretical_loss": 3.4007973755914924,
      "tokens_seen": 2221408256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324848484848485,
      "loss": 2.727,
      "theoretical_loss": 3.400789182078694,
      "tokens_seen": 2221473792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033246464646464644,
      "loss": 2.7374,
      "theoretical_loss": 3.400780988875289,
      "tokens_seen": 2221539328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033244444444444445,
      "loss": 2.5549,
      "theoretical_loss": 3.400772795981256,
      "tokens_seen": 2221604864
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324242424242424,
      "loss": 2.7307,
      "theoretical_loss": 3.4007646033965746,
      "tokens_seen": 2221670400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324040404040404,
      "loss": 2.5225,
      "theoretical_loss": 3.4007564111212236,
      "tokens_seen": 2221735936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323838383838384,
      "loss": 2.3545,
      "theoretical_loss": 3.400748219155183,
      "tokens_seen": 2221801472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323636363636364,
      "loss": 2.5976,
      "theoretical_loss": 3.400740027498431,
      "tokens_seen": 2221867008
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1225752,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6411614418029785,
      "objective/train/theoretical_loss": 3.4007318361509475,
      "objective/train/tokens_used": 580791776,
      "theoretical_loss": 3.4007318361509475,
      "tokens_seen": 2221932544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033234343434343435,
      "loss": 2.7676,
      "theoretical_loss": 3.4007318361509475,
      "tokens_seen": 2221932544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323232323232323,
      "loss": 2.787,
      "theoretical_loss": 3.400723645112712,
      "tokens_seen": 2221998080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323030303030303,
      "loss": 2.5546,
      "theoretical_loss": 3.400715454383702,
      "tokens_seen": 2222063616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033228282828282827,
      "loss": 2.5891,
      "theoretical_loss": 3.400707263963899,
      "tokens_seen": 2222129152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322626262626263,
      "loss": 2.5609,
      "theoretical_loss": 3.4006990738532807,
      "tokens_seen": 2222194688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033224242424242424,
      "loss": 2.4844,
      "theoretical_loss": 3.400690884051827,
      "tokens_seen": 2222260224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322222222222222,
      "loss": 2.8366,
      "theoretical_loss": 3.4006826945595163,
      "tokens_seen": 2222325760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322020202020202,
      "loss": 2.6884,
      "theoretical_loss": 3.4006745053763288,
      "tokens_seen": 2222391296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321818181818182,
      "loss": 2.5898,
      "theoretical_loss": 3.400666316502243,
      "tokens_seen": 2222456832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321616161616162,
      "loss": 2.6397,
      "theoretical_loss": 3.4006581279372385,
      "tokens_seen": 2222522368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033214141414141413,
      "loss": 2.7176,
      "theoretical_loss": 3.400649939681294,
      "tokens_seen": 2222587904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321212121212121,
      "loss": 2.7142,
      "theoretical_loss": 3.4006417517343897,
      "tokens_seen": 2222653440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321010101010101,
      "loss": 2.6039,
      "theoretical_loss": 3.4006335640965037,
      "tokens_seen": 2222718976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003320808080808081,
      "loss": 2.6894,
      "theoretical_loss": 3.4006253767676164,
      "tokens_seen": 2222784512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033206060606060607,
      "loss": 2.5729,
      "theoretical_loss": 3.400617189747706,
      "tokens_seen": 2222850048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332040404040404,
      "loss": 2.486,
      "theoretical_loss": 3.4006090030367524,
      "tokens_seen": 2222915584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332020202020202,
      "loss": 2.3843,
      "theoretical_loss": 3.4006008166347343,
      "tokens_seen": 2222981120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033200000000000005,
      "loss": 2.7624,
      "theoretical_loss": 3.4005926305416314,
      "tokens_seen": 2223046656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331979797979798,
      "loss": 2.5662,
      "theoretical_loss": 3.4005844447574227,
      "tokens_seen": 2223112192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033195959595959596,
      "loss": 2.5462,
      "theoretical_loss": 3.4005762592820874,
      "tokens_seen": 2223177728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319393939393939,
      "loss": 2.6776,
      "theoretical_loss": 3.400568074115605,
      "tokens_seen": 2223243264
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319191919191919,
      "loss": 2.4568,
      "theoretical_loss": 3.4005598892579543,
      "tokens_seen": 2223308800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033189898989898994,
      "loss": 2.4441,
      "theoretical_loss": 3.4005517047091147,
      "tokens_seen": 2223374336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318787878787879,
      "loss": 2.8038,
      "theoretical_loss": 3.400543520469066,
      "tokens_seen": 2223439872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033185858585858585,
      "loss": 2.3434,
      "theoretical_loss": 3.4005353365377866,
      "tokens_seen": 2223505408
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1226492,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.975679636001587,
      "objective/train/theoretical_loss": 3.4005271529152563,
      "objective/train/tokens_used": 582430176,
      "theoretical_loss": 3.4005271529152563,
      "tokens_seen": 2223570944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318383838383838,
      "loss": 2.5701,
      "theoretical_loss": 3.4005271529152563,
      "tokens_seen": 2223570944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318181818181819,
      "loss": 2.6182,
      "theoretical_loss": 3.400518969601454,
      "tokens_seen": 2223636480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033179797979797983,
      "loss": 2.5841,
      "theoretical_loss": 3.4005107865963597,
      "tokens_seen": 2223702016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317777777777778,
      "loss": 2.6592,
      "theoretical_loss": 3.4005026038999517,
      "tokens_seen": 2223767552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033175757575757575,
      "loss": 2.7015,
      "theoretical_loss": 3.4004944215122097,
      "tokens_seen": 2223833088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317373737373737,
      "loss": 2.647,
      "theoretical_loss": 3.4004862394331132,
      "tokens_seen": 2223898624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033171717171717177,
      "loss": 2.629,
      "theoretical_loss": 3.400478057662641,
      "tokens_seen": 2223964160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316969696969697,
      "loss": 2.6428,
      "theoretical_loss": 3.4004698762007726,
      "tokens_seen": 2224029696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316767676767677,
      "loss": 2.5553,
      "theoretical_loss": 3.400461695047487,
      "tokens_seen": 2224095232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033165656565656564,
      "loss": 2.7775,
      "theoretical_loss": 3.400453514202764,
      "tokens_seen": 2224160768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033163636363636365,
      "loss": 2.5992,
      "theoretical_loss": 3.4004453336665827,
      "tokens_seen": 2224226304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033161616161616166,
      "loss": 2.5181,
      "theoretical_loss": 3.4004371534389217,
      "tokens_seen": 2224291840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315959595959596,
      "loss": 2.8832,
      "theoretical_loss": 3.4004289735197615,
      "tokens_seen": 2224357376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315757575757576,
      "loss": 2.6969,
      "theoretical_loss": 3.4004207939090803,
      "tokens_seen": 2224422912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033155555555555553,
      "loss": 2.5016,
      "theoretical_loss": 3.4004126146068576,
      "tokens_seen": 2224488448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033153535353535354,
      "loss": 2.6914,
      "theoretical_loss": 3.400404435613073,
      "tokens_seen": 2224553984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033151515151515155,
      "loss": 2.6616,
      "theoretical_loss": 3.4003962569277055,
      "tokens_seen": 2224619520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314949494949495,
      "loss": 2.5709,
      "theoretical_loss": 3.400388078550735,
      "tokens_seen": 2224685056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033147474747474747,
      "loss": 2.5039,
      "theoretical_loss": 3.40037990048214,
      "tokens_seen": 2224750592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314545454545455,
      "loss": 2.6247,
      "theoretical_loss": 3.4003717227219,
      "tokens_seen": 2224816128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033143434343434344,
      "loss": 2.8543,
      "theoretical_loss": 3.4003635452699945,
      "tokens_seen": 2224881664
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033141414141414145,
      "loss": 2.8326,
      "theoretical_loss": 3.4003553681264025,
      "tokens_seen": 2224947200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313939393939394,
      "loss": 2.6964,
      "theoretical_loss": 3.400347191291104,
      "tokens_seen": 2225012736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033137373737373736,
      "loss": 2.7465,
      "theoretical_loss": 3.4003390147640773,
      "tokens_seen": 2225078272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033135353535353537,
      "loss": 2.8577,
      "theoretical_loss": 3.4003308385453024,
      "tokens_seen": 2225143808
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1227694,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9130122661590576,
      "objective/train/theoretical_loss": 3.4003226626347582,
      "objective/train/tokens_used": 584068576,
      "theoretical_loss": 3.4003226626347582,
      "tokens_seen": 2225209344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033133333333333333,
      "loss": 2.6959,
      "theoretical_loss": 3.4003226626347582,
      "tokens_seen": 2225209344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033131313131313134,
      "loss": 2.7366,
      "theoretical_loss": 3.400314487032424,
      "tokens_seen": 2225274880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312929292929293,
      "loss": 2.5441,
      "theoretical_loss": 3.4003063117382797,
      "tokens_seen": 2225340416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312727272727273,
      "loss": 2.48,
      "theoretical_loss": 3.400298136752304,
      "tokens_seen": 2225405952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033125252525252526,
      "loss": 2.715,
      "theoretical_loss": 3.4002899620744764,
      "tokens_seen": 2225471488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312323232323232,
      "loss": 2.5601,
      "theoretical_loss": 3.400281787704776,
      "tokens_seen": 2225537024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033121212121212123,
      "loss": 2.5835,
      "theoretical_loss": 3.400273613643183,
      "tokens_seen": 2225602560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311919191919192,
      "loss": 2.7125,
      "theoretical_loss": 3.4002654398896754,
      "tokens_seen": 2225668096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311717171717172,
      "loss": 2.5304,
      "theoretical_loss": 3.400257266444233,
      "tokens_seen": 2225733632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033115151515151516,
      "loss": 2.4401,
      "theoretical_loss": 3.4002490933068357,
      "tokens_seen": 2225799168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311313131313131,
      "loss": 2.6642,
      "theoretical_loss": 3.4002409204774624,
      "tokens_seen": 2225864704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311111111111111,
      "loss": 2.7592,
      "theoretical_loss": 3.4002327479560925,
      "tokens_seen": 2225930240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033109090909090914,
      "loss": 2.5815,
      "theoretical_loss": 3.4002245757427048,
      "tokens_seen": 2225995776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003310707070707071,
      "loss": 2.6235,
      "theoretical_loss": 3.4002164038372795,
      "tokens_seen": 2226061312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033105050505050505,
      "loss": 2.478,
      "theoretical_loss": 3.400208232239795,
      "tokens_seen": 2226126848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331030303030303,
      "loss": 2.8519,
      "theoretical_loss": 3.4002000609502314,
      "tokens_seen": 2226192384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331010101010101,
      "loss": 2.5961,
      "theoretical_loss": 3.4001918899685677,
      "tokens_seen": 2226257920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033098989898989903,
      "loss": 2.6191,
      "theoretical_loss": 3.4001837192947835,
      "tokens_seen": 2226323456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000330969696969697,
      "loss": 2.6254,
      "theoretical_loss": 3.4001755489288574,
      "tokens_seen": 2226388992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033094949494949494,
      "loss": 2.5801,
      "theoretical_loss": 3.40016737887077,
      "tokens_seen": 2226454528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309292929292929,
      "loss": 2.4333,
      "theoretical_loss": 3.400159209120499,
      "tokens_seen": 2226520064
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033090909090909096,
      "loss": 2.7306,
      "theoretical_loss": 3.4001510396780255,
      "tokens_seen": 2226585600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308888888888889,
      "loss": 2.7405,
      "theoretical_loss": 3.400142870543328,
      "tokens_seen": 2226651136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308686868686869,
      "loss": 2.6858,
      "theoretical_loss": 3.400134701716385,
      "tokens_seen": 2226716672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033084848484848483,
      "loss": 2.6114,
      "theoretical_loss": 3.4001265331971773,
      "tokens_seen": 2226782208
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 1229090,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5205607414245605,
      "objective/train/theoretical_loss": 3.400118364985684,
      "objective/train/tokens_used": 585706976,
      "theoretical_loss": 3.400118364985684,
      "tokens_seen": 2226847744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308282828282828,
      "loss": 2.6387,
      "theoretical_loss": 3.400118364985684,
      "tokens_seen": 2226847744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033080808080808086,
      "loss": 2.3516,
      "theoretical_loss": 3.4001101970818834,
      "tokens_seen": 2226913280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307878787878788,
      "loss": 2.6906,
      "theoretical_loss": 3.4001020294857556,
      "tokens_seen": 2226978816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033076767676767677,
      "loss": 2.5544,
      "theoretical_loss": 3.4000938621972803,
      "tokens_seen": 2227044352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033074747474747473,
      "loss": 2.7082,
      "theoretical_loss": 3.400085695216436,
      "tokens_seen": 2227109888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307272727272727,
      "loss": 2.7331,
      "theoretical_loss": 3.400077528543203,
      "tokens_seen": 2227175424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033070707070707075,
      "loss": 2.6219,
      "theoretical_loss": 3.40006936217756,
      "tokens_seen": 2227240960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003306868686868687,
      "loss": 2.7171,
      "theoretical_loss": 3.4000611961194864,
      "tokens_seen": 2227306496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033066666666666666,
      "loss": 2.5398,
      "theoretical_loss": 3.400053030368962,
      "tokens_seen": 2227372032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003306464646464646,
      "loss": 2.5068,
      "theoretical_loss": 3.400044864925966,
      "tokens_seen": 2227437568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033062626262626263,
      "loss": 2.6589,
      "theoretical_loss": 3.400036699790477,
      "tokens_seen": 2227503104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033060606060606064,
      "loss": 2.5127,
      "theoretical_loss": 3.4000285349624755,
      "tokens_seen": 2227568640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305858585858586,
      "loss": 2.5166,
      "theoretical_loss": 3.4000203704419403,
      "tokens_seen": 2227634176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033056565656565656,
      "loss": 2.5547,
      "theoretical_loss": 3.4000122062288507,
      "tokens_seen": 2227699712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305454545454545,
      "loss": 2.6215,
      "theoretical_loss": 3.4000040423231863,
      "tokens_seen": 2227765248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305252525252525,
      "loss": 2.7859,
      "theoretical_loss": 3.3999958787249267,
      "tokens_seen": 2227830784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033050505050505053,
      "loss": 2.5376,
      "theoretical_loss": 3.399987715434051,
      "tokens_seen": 2227896320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304848484848485,
      "loss": 2.513,
      "theoretical_loss": 3.3999795524505387,
      "tokens_seen": 2227961856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033046464646464645,
      "loss": 2.4192,
      "theoretical_loss": 3.3999713897743686,
      "tokens_seen": 2228027392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033044444444444446,
      "loss": 2.7682,
      "theoretical_loss": 3.3999632274055207,
      "tokens_seen": 2228092928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304242424242424,
      "loss": 2.7186,
      "theoretical_loss": 3.399955065343975,
      "tokens_seen": 2228158464
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033040404040404043,
      "loss": 2.6879,
      "theoretical_loss": 3.399946903589709,
      "tokens_seen": 2228224000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303838383838384,
      "loss": 2.5556,
      "theoretical_loss": 3.3999387421427043,
      "tokens_seen": 2228289536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033036363636363634,
      "loss": 2.6052,
      "theoretical_loss": 3.3999305810029385,
      "tokens_seen": 2228355072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033034343434343435,
      "loss": 2.5945,
      "theoretical_loss": 3.3999224201703924,
      "tokens_seen": 2228420608
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1229755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9436287879943848,
      "objective/train/theoretical_loss": 3.3999142596450445,
      "objective/train/tokens_used": 587345376,
      "theoretical_loss": 3.3999142596450445,
      "tokens_seen": 2228486144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303232323232323,
      "loss": 2.7154,
      "theoretical_loss": 3.3999142596450445,
      "tokens_seen": 2228486144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303030303030303,
      "loss": 2.5713,
      "theoretical_loss": 3.3999060994268744,
      "tokens_seen": 2228551680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302828282828283,
      "loss": 2.6092,
      "theoretical_loss": 3.3998979395158613,
      "tokens_seen": 2228617216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302626262626263,
      "loss": 2.5645,
      "theoretical_loss": 3.399889779911985,
      "tokens_seen": 2228682752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033024242424242424,
      "loss": 2.5129,
      "theoretical_loss": 3.399881620615225,
      "tokens_seen": 2228748288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302222222222222,
      "loss": 2.5417,
      "theoretical_loss": 3.3998734616255604,
      "tokens_seen": 2228813824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302020202020202,
      "loss": 2.5971,
      "theoretical_loss": 3.399865302942971,
      "tokens_seen": 2228879360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033018181818181817,
      "loss": 2.5154,
      "theoretical_loss": 3.3998571445674353,
      "tokens_seen": 2228944896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301616161616162,
      "loss": 2.5734,
      "theoretical_loss": 3.3998489864989336,
      "tokens_seen": 2229010432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033014141414141414,
      "loss": 2.5061,
      "theoretical_loss": 3.399840828737445,
      "tokens_seen": 2229075968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033012121212121215,
      "loss": 2.5298,
      "theoretical_loss": 3.399832671282949,
      "tokens_seen": 2229141504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301010101010101,
      "loss": 2.5969,
      "theoretical_loss": 3.3998245141354255,
      "tokens_seen": 2229207040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300808080808081,
      "loss": 2.5479,
      "theoretical_loss": 3.3998163572948528,
      "tokens_seen": 2229272576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300606060606061,
      "loss": 2.7211,
      "theoretical_loss": 3.3998082007612114,
      "tokens_seen": 2229338112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033004040404040403,
      "loss": 2.4358,
      "theoretical_loss": 3.3998000445344796,
      "tokens_seen": 2229403648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033002020202020204,
      "loss": 2.7302,
      "theoretical_loss": 3.3997918886146383,
      "tokens_seen": 2229469184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033,
      "loss": 2.5349,
      "theoretical_loss": 3.3997837330016654,
      "tokens_seen": 2229534720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329979797979798,
      "loss": 2.5502,
      "theoretical_loss": 3.399775577695541,
      "tokens_seen": 2229600256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032995959595959597,
      "loss": 2.807,
      "theoretical_loss": 3.3997674226962453,
      "tokens_seen": 2229665792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299393939393939,
      "loss": 2.5437,
      "theoretical_loss": 3.3997592680037565,
      "tokens_seen": 2229731328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032991919191919193,
      "loss": 2.6995,
      "theoretical_loss": 3.399751113618055,
      "tokens_seen": 2229796864
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032989898989898995,
      "loss": 2.5486,
      "theoretical_loss": 3.3997429595391195,
      "tokens_seen": 2229862400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298787878787879,
      "loss": 2.5481,
      "theoretical_loss": 3.39973480576693,
      "tokens_seen": 2229927936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032985858585858586,
      "loss": 2.4959,
      "theoretical_loss": 3.3997266523014655,
      "tokens_seen": 2229993472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298383838383838,
      "loss": 2.6664,
      "theoretical_loss": 3.399718499142706,
      "tokens_seen": 2230059008
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1230979,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6503684520721436,
      "objective/train/theoretical_loss": 3.39971034629063,
      "objective/train/tokens_used": 588983776,
      "theoretical_loss": 3.39971034629063,
      "tokens_seen": 2230124544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298181818181818,
      "loss": 2.78,
      "theoretical_loss": 3.39971034629063,
      "tokens_seen": 2230124544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032979797979797984,
      "loss": 2.6312,
      "theoretical_loss": 3.399702193745218,
      "tokens_seen": 2230190080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297777777777778,
      "loss": 2.632,
      "theoretical_loss": 3.399694041506449,
      "tokens_seen": 2230255616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032975757575757575,
      "loss": 2.5803,
      "theoretical_loss": 3.3996858895743025,
      "tokens_seen": 2230321152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297373737373737,
      "loss": 2.44,
      "theoretical_loss": 3.3996777379487577,
      "tokens_seen": 2230386688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297171717171718,
      "loss": 2.4862,
      "theoretical_loss": 3.399669586629795,
      "tokens_seen": 2230452224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032969696969696973,
      "loss": 2.5718,
      "theoretical_loss": 3.3996614356173924,
      "tokens_seen": 2230517760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296767676767677,
      "loss": 2.5235,
      "theoretical_loss": 3.39965328491153,
      "tokens_seen": 2230583296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032965656565656564,
      "loss": 2.5377,
      "theoretical_loss": 3.3996451345121876,
      "tokens_seen": 2230648832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296363636363636,
      "loss": 2.6386,
      "theoretical_loss": 3.399636984419345,
      "tokens_seen": 2230714368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032961616161616167,
      "loss": 2.6945,
      "theoretical_loss": 3.3996288346329804,
      "tokens_seen": 2230779904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295959595959596,
      "loss": 2.7157,
      "theoretical_loss": 3.3996206851530744,
      "tokens_seen": 2230845440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295757575757576,
      "loss": 2.434,
      "theoretical_loss": 3.399612535979606,
      "tokens_seen": 2230910976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032955555555555554,
      "loss": 2.6753,
      "theoretical_loss": 3.3996043871125545,
      "tokens_seen": 2230976512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032953535353535355,
      "loss": 2.6915,
      "theoretical_loss": 3.3995962385519,
      "tokens_seen": 2231042048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032951515151515156,
      "loss": 2.7246,
      "theoretical_loss": 3.3995880902976214,
      "tokens_seen": 2231107584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294949494949495,
      "loss": 2.5912,
      "theoretical_loss": 3.399579942349698,
      "tokens_seen": 2231173120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032947474747474747,
      "loss": 2.8135,
      "theoretical_loss": 3.3995717947081103,
      "tokens_seen": 2231238656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032945454545454543,
      "loss": 2.8094,
      "theoretical_loss": 3.399563647372837,
      "tokens_seen": 2231304192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032943434343434344,
      "loss": 2.6365,
      "theoretical_loss": 3.3995555003438573,
      "tokens_seen": 2231369728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032941414141414145,
      "loss": 2.6156,
      "theoretical_loss": 3.3995473536211516,
      "tokens_seen": 2231435264
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293939393939394,
      "loss": 2.5703,
      "theoretical_loss": 3.3995392072046986,
      "tokens_seen": 2231500800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032937373737373737,
      "loss": 2.7854,
      "theoretical_loss": 3.3995310610944784,
      "tokens_seen": 2231566336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293535353535354,
      "loss": 2.7232,
      "theoretical_loss": 3.3995229152904702,
      "tokens_seen": 2231631872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032933333333333333,
      "loss": 2.4552,
      "theoretical_loss": 3.3995147697926535,
      "tokens_seen": 2231697408
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1231622,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8242321014404297,
      "objective/train/theoretical_loss": 3.3995066246010075,
      "objective/train/tokens_used": 590622176,
      "theoretical_loss": 3.3995066246010075,
      "tokens_seen": 2231762944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032931313131313134,
      "loss": 2.7745,
      "theoretical_loss": 3.3995066246010075,
      "tokens_seen": 2231762944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292929292929293,
      "loss": 2.5516,
      "theoretical_loss": 3.399498479715512,
      "tokens_seen": 2231828480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032927272727272726,
      "loss": 2.6586,
      "theoretical_loss": 3.399490335136147,
      "tokens_seen": 2231894016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032925252525252527,
      "loss": 2.5252,
      "theoretical_loss": 3.399482190862891,
      "tokens_seen": 2231959552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292323232323232,
      "loss": 2.4889,
      "theoretical_loss": 3.3994740468957243,
      "tokens_seen": 2232025088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032921212121212124,
      "loss": 2.6499,
      "theoretical_loss": 3.3994659032346264,
      "tokens_seen": 2232090624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291919191919192,
      "loss": 2.6323,
      "theoretical_loss": 3.399457759879576,
      "tokens_seen": 2232156160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032917171717171715,
      "loss": 2.5841,
      "theoretical_loss": 3.3994496168305535,
      "tokens_seen": 2232221696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032915151515151516,
      "loss": 2.6963,
      "theoretical_loss": 3.399441474087538,
      "tokens_seen": 2232287232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291313131313131,
      "loss": 2.5118,
      "theoretical_loss": 3.399433331650509,
      "tokens_seen": 2232352768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032911111111111113,
      "loss": 2.4879,
      "theoretical_loss": 3.3994251895194463,
      "tokens_seen": 2232418304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290909090909091,
      "loss": 2.5526,
      "theoretical_loss": 3.399417047694329,
      "tokens_seen": 2232483840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290707070707071,
      "loss": 2.5397,
      "theoretical_loss": 3.399408906175137,
      "tokens_seen": 2232549376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032905050505050505,
      "loss": 2.6151,
      "theoretical_loss": 3.3994007649618494,
      "tokens_seen": 2232614912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329030303030303,
      "loss": 2.6126,
      "theoretical_loss": 3.3993926240544465,
      "tokens_seen": 2232680448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329010101010101,
      "loss": 2.6836,
      "theoretical_loss": 3.399384483452907,
      "tokens_seen": 2232745984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328989898989899,
      "loss": 2.5229,
      "theoretical_loss": 3.399376343157211,
      "tokens_seen": 2232811520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328969696969697,
      "loss": 2.3784,
      "theoretical_loss": 3.3993682031673376,
      "tokens_seen": 2232877056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032894949494949495,
      "loss": 2.7474,
      "theoretical_loss": 3.3993600634832664,
      "tokens_seen": 2232942592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289292929292929,
      "loss": 2.6362,
      "theoretical_loss": 3.399351924104977,
      "tokens_seen": 2233008128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289090909090909,
      "loss": 2.5158,
      "theoretical_loss": 3.3993437850324497,
      "tokens_seen": 2233073664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288888888888889,
      "loss": 2.3868,
      "theoretical_loss": 3.399335646265663,
      "tokens_seen": 2233139200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288686868686869,
      "loss": 2.5855,
      "theoretical_loss": 3.3993275078045966,
      "tokens_seen": 2233204736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032884848484848484,
      "loss": 2.5106,
      "theoretical_loss": 3.3993193696492305,
      "tokens_seen": 2233270272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288282828282828,
      "loss": 2.6512,
      "theoretical_loss": 3.3993112317995435,
      "tokens_seen": 2233335808
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1232627,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4419264793395996,
      "objective/train/theoretical_loss": 3.399303094255516,
      "objective/train/tokens_used": 592260576,
      "theoretical_loss": 3.399303094255516,
      "tokens_seen": 2233401344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288080808080808,
      "loss": 2.5994,
      "theoretical_loss": 3.399303094255516,
      "tokens_seen": 2233401344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287878787878788,
      "loss": 2.585,
      "theoretical_loss": 3.399294957017127,
      "tokens_seen": 2233466880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287676767676768,
      "loss": 2.6031,
      "theoretical_loss": 3.3992868200843565,
      "tokens_seen": 2233532416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032874747474747473,
      "loss": 2.4659,
      "theoretical_loss": 3.3992786834571835,
      "tokens_seen": 2233597952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287272727272727,
      "loss": 2.4812,
      "theoretical_loss": 3.399270547135588,
      "tokens_seen": 2233663488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032870707070707075,
      "loss": 2.5293,
      "theoretical_loss": 3.3992624111195497,
      "tokens_seen": 2233729024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286868686868687,
      "loss": 2.5891,
      "theoretical_loss": 3.3992542754090476,
      "tokens_seen": 2233794560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032866666666666667,
      "loss": 2.5712,
      "theoretical_loss": 3.3992461400040614,
      "tokens_seen": 2233860096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286464646464646,
      "loss": 2.5188,
      "theoretical_loss": 3.399238004904571,
      "tokens_seen": 2233925632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286262626262626,
      "loss": 2.4667,
      "theoretical_loss": 3.3992298701105557,
      "tokens_seen": 2233991168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032860606060606065,
      "loss": 2.5163,
      "theoretical_loss": 3.399221735621995,
      "tokens_seen": 2234056704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285858585858586,
      "loss": 2.867,
      "theoretical_loss": 3.399213601438869,
      "tokens_seen": 2234122240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032856565656565656,
      "loss": 2.527,
      "theoretical_loss": 3.3992054675611563,
      "tokens_seen": 2234187776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285454545454545,
      "loss": 2.6442,
      "theoretical_loss": 3.3991973339888375,
      "tokens_seen": 2234253312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032852525252525253,
      "loss": 2.5602,
      "theoretical_loss": 3.3991892007218913,
      "tokens_seen": 2234318848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032850505050505054,
      "loss": 2.6464,
      "theoretical_loss": 3.3991810677602983,
      "tokens_seen": 2234384384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284848484848485,
      "loss": 2.549,
      "theoretical_loss": 3.399172935104037,
      "tokens_seen": 2234449920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032846464646464645,
      "loss": 2.7451,
      "theoretical_loss": 3.399164802753088,
      "tokens_seen": 2234515456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284444444444444,
      "loss": 2.5182,
      "theoretical_loss": 3.3991566707074297,
      "tokens_seen": 2234580992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284242424242425,
      "loss": 2.8511,
      "theoretical_loss": 3.3991485389670424,
      "tokens_seen": 2234646528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032840404040404043,
      "loss": 2.4048,
      "theoretical_loss": 3.399140407531906,
      "tokens_seen": 2234712064
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283838383838384,
      "loss": 2.6627,
      "theoretical_loss": 3.3991322764019993,
      "tokens_seen": 2234777600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032836363636363635,
      "loss": 2.6656,
      "theoretical_loss": 3.3991241455773027,
      "tokens_seen": 2234843136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032834343434343436,
      "loss": 2.5002,
      "theoretical_loss": 3.3991160150577953,
      "tokens_seen": 2234908672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032832323232323237,
      "loss": 2.6034,
      "theoretical_loss": 3.3991078848434566,
      "tokens_seen": 2234974208
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1233498,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7271065711975098,
      "objective/train/theoretical_loss": 3.3990997549342667,
      "objective/train/tokens_used": 593898976,
      "theoretical_loss": 3.3990997549342667,
      "tokens_seen": 2235039744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283030303030303,
      "loss": 2.5348,
      "theoretical_loss": 3.3990997549342667,
      "tokens_seen": 2235039744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282828282828283,
      "loss": 2.5156,
      "theoretical_loss": 3.399091625330205,
      "tokens_seen": 2235105280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032826262626262624,
      "loss": 2.818,
      "theoretical_loss": 3.3990834960312504,
      "tokens_seen": 2235170816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032824242424242425,
      "loss": 2.6451,
      "theoretical_loss": 3.3990753670373834,
      "tokens_seen": 2235236352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032822222222222226,
      "loss": 2.2549,
      "theoretical_loss": 3.399067238348583,
      "tokens_seen": 2235301888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282020202020202,
      "loss": 2.4261,
      "theoretical_loss": 3.3990591099648295,
      "tokens_seen": 2235367424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281818181818182,
      "loss": 2.6187,
      "theoretical_loss": 3.399050981886102,
      "tokens_seen": 2235432960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281616161616162,
      "loss": 2.442,
      "theoretical_loss": 3.3990428541123805,
      "tokens_seen": 2235498496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032814141414141414,
      "loss": 2.6671,
      "theoretical_loss": 3.3990347266436443,
      "tokens_seen": 2235564032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032812121212121215,
      "loss": 2.6224,
      "theoretical_loss": 3.399026599479873,
      "tokens_seen": 2235629568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281010101010101,
      "loss": 2.8171,
      "theoretical_loss": 3.399018472621046,
      "tokens_seen": 2235695104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032808080808080807,
      "loss": 2.487,
      "theoretical_loss": 3.3990103460671435,
      "tokens_seen": 2235760640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003280606060606061,
      "loss": 2.6067,
      "theoretical_loss": 3.3990022198181444,
      "tokens_seen": 2235826176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032804040404040404,
      "loss": 2.5759,
      "theoretical_loss": 3.398994093874029,
      "tokens_seen": 2235891712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032802020202020205,
      "loss": 2.4193,
      "theoretical_loss": 3.3989859682347765,
      "tokens_seen": 2235957248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328,
      "loss": 2.6055,
      "theoretical_loss": 3.398977842900367,
      "tokens_seen": 2236022784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327979797979798,
      "loss": 2.8154,
      "theoretical_loss": 3.39896971787078,
      "tokens_seen": 2236088320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032795959595959597,
      "loss": 2.6897,
      "theoretical_loss": 3.3989615931459944,
      "tokens_seen": 2236153856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032793939393939393,
      "loss": 2.665,
      "theoretical_loss": 3.398953468725991,
      "tokens_seen": 2236219392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032791919191919194,
      "loss": 2.5713,
      "theoretical_loss": 3.3989453446107483,
      "tokens_seen": 2236284928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278989898989899,
      "loss": 2.6761,
      "theoretical_loss": 3.3989372208002466,
      "tokens_seen": 2236350464
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278787878787879,
      "loss": 2.6578,
      "theoretical_loss": 3.3989290972944657,
      "tokens_seen": 2236416000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032785858585858586,
      "loss": 2.6511,
      "theoretical_loss": 3.398920974093385,
      "tokens_seen": 2236481536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278383838383838,
      "loss": 2.6278,
      "theoretical_loss": 3.3989128511969833,
      "tokens_seen": 2236547072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032781818181818183,
      "loss": 2.5598,
      "theoretical_loss": 3.398904728605242,
      "tokens_seen": 2236612608
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1234140,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.896017551422119,
      "objective/train/theoretical_loss": 3.398896606318139,
      "objective/train/tokens_used": 595537376,
      "theoretical_loss": 3.398896606318139,
      "tokens_seen": 2236678144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032779797979797984,
      "loss": 2.4891,
      "theoretical_loss": 3.398896606318139,
      "tokens_seen": 2236678144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277777777777778,
      "loss": 2.7951,
      "theoretical_loss": 3.398888484335655,
      "tokens_seen": 2236743680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032775757575757576,
      "loss": 2.6724,
      "theoretical_loss": 3.3988803626577697,
      "tokens_seen": 2236809216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277373737373737,
      "loss": 2.452,
      "theoretical_loss": 3.398872241284462,
      "tokens_seen": 2236874752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277171717171717,
      "loss": 2.7162,
      "theoretical_loss": 3.398864120215712,
      "tokens_seen": 2236940288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032769696969696974,
      "loss": 2.6553,
      "theoretical_loss": 3.3988559994514995,
      "tokens_seen": 2237005824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276767676767677,
      "loss": 2.6449,
      "theoretical_loss": 3.398847878991804,
      "tokens_seen": 2237071360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032765656565656565,
      "loss": 2.5481,
      "theoretical_loss": 3.3988397588366053,
      "tokens_seen": 2237136896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276363636363636,
      "loss": 2.4543,
      "theoretical_loss": 3.398831638985883,
      "tokens_seen": 2237202432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032761616161616167,
      "loss": 2.4903,
      "theoretical_loss": 3.398823519439616,
      "tokens_seen": 2237267968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032759595959595963,
      "loss": 2.4446,
      "theoretical_loss": 3.398815400197785,
      "tokens_seen": 2237333504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275757575757576,
      "loss": 2.487,
      "theoretical_loss": 3.3988072812603694,
      "tokens_seen": 2237399040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032755555555555554,
      "loss": 2.4346,
      "theoretical_loss": 3.3987991626273493,
      "tokens_seen": 2237464576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275353535353535,
      "loss": 2.681,
      "theoretical_loss": 3.398791044298703,
      "tokens_seen": 2237530112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032751515151515156,
      "loss": 2.6709,
      "theoretical_loss": 3.3987829262744116,
      "tokens_seen": 2237595648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274949494949495,
      "loss": 2.8194,
      "theoretical_loss": 3.398774808554454,
      "tokens_seen": 2237661184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274747474747475,
      "loss": 2.5114,
      "theoretical_loss": 3.3987666911388104,
      "tokens_seen": 2237726720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032745454545454544,
      "loss": 2.6755,
      "theoretical_loss": 3.3987585740274593,
      "tokens_seen": 2237792256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274343434343434,
      "loss": 2.6678,
      "theoretical_loss": 3.398750457220382,
      "tokens_seen": 2237857792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032741414141414146,
      "loss": 2.4602,
      "theoretical_loss": 3.3987423407175577,
      "tokens_seen": 2237923328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003273939393939394,
      "loss": 2.3153,
      "theoretical_loss": 3.3987342245189653,
      "tokens_seen": 2237988864
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032737373737373737,
      "loss": 2.6252,
      "theoretical_loss": 3.398726108624585,
      "tokens_seen": 2238054400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032735353535353533,
      "loss": 2.5537,
      "theoretical_loss": 3.398717993034397,
      "tokens_seen": 2238119936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032733333333333334,
      "loss": 2.5703,
      "theoretical_loss": 3.39870987774838,
      "tokens_seen": 2238185472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032731313131313135,
      "loss": 2.8515,
      "theoretical_loss": 3.3987017627665144,
      "tokens_seen": 2238251008
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1235278,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7147154808044434,
      "objective/train/theoretical_loss": 3.3986936480887793,
      "objective/train/tokens_used": 597175776,
      "theoretical_loss": 3.3986936480887793,
      "tokens_seen": 2238316544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272929292929293,
      "loss": 2.6231,
      "theoretical_loss": 3.3986936480887793,
      "tokens_seen": 2238316544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032727272727272726,
      "loss": 2.7508,
      "theoretical_loss": 3.398685533715155,
      "tokens_seen": 2238382080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272525252525252,
      "loss": 2.6738,
      "theoretical_loss": 3.3986774196456215,
      "tokens_seen": 2238447616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032723232323232323,
      "loss": 2.7251,
      "theoretical_loss": 3.3986693058801576,
      "tokens_seen": 2238513152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032721212121212124,
      "loss": 2.6471,
      "theoretical_loss": 3.3986611924187433,
      "tokens_seen": 2238578688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271919191919192,
      "loss": 2.704,
      "theoretical_loss": 3.3986530792613583,
      "tokens_seen": 2238644224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032717171717171716,
      "loss": 2.7981,
      "theoretical_loss": 3.3986449664079825,
      "tokens_seen": 2238709760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032715151515151517,
      "loss": 2.683,
      "theoretical_loss": 3.3986368538585956,
      "tokens_seen": 2238775296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271313131313131,
      "loss": 2.5068,
      "theoretical_loss": 3.398628741613177,
      "tokens_seen": 2238840832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032711111111111114,
      "loss": 2.5823,
      "theoretical_loss": 3.398620629671707,
      "tokens_seen": 2238906368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270909090909091,
      "loss": 2.5502,
      "theoretical_loss": 3.398612518034165,
      "tokens_seen": 2238971904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032707070707070705,
      "loss": 2.5768,
      "theoretical_loss": 3.3986044067005303,
      "tokens_seen": 2239037440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032705050505050506,
      "loss": 2.4648,
      "theoretical_loss": 3.3985962956707834,
      "tokens_seen": 2239102976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327030303030303,
      "loss": 2.5492,
      "theoretical_loss": 3.398588184944903,
      "tokens_seen": 2239168512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032701010101010103,
      "loss": 2.6662,
      "theoretical_loss": 3.39858007452287,
      "tokens_seen": 2239234048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326989898989899,
      "loss": 2.637,
      "theoretical_loss": 3.3985719644046632,
      "tokens_seen": 2239299584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326969696969697,
      "loss": 2.5445,
      "theoretical_loss": 3.398563854590263,
      "tokens_seen": 2239365120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032694949494949495,
      "loss": 2.5525,
      "theoretical_loss": 3.3985557450796486,
      "tokens_seen": 2239430656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269292929292929,
      "loss": 2.8025,
      "theoretical_loss": 3.3985476358728,
      "tokens_seen": 2239496192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269090909090909,
      "loss": 2.6419,
      "theoretical_loss": 3.3985395269696967,
      "tokens_seen": 2239561728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268888888888889,
      "loss": 2.4434,
      "theoretical_loss": 3.398531418370319,
      "tokens_seen": 2239627264
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268686868686869,
      "loss": 2.4144,
      "theoretical_loss": 3.398523310074646,
      "tokens_seen": 2239692800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032684848484848485,
      "loss": 2.6922,
      "theoretical_loss": 3.3985152020826574,
      "tokens_seen": 2239758336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268282828282828,
      "loss": 2.5596,
      "theoretical_loss": 3.3985070943943336,
      "tokens_seen": 2239823872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268080808080808,
      "loss": 2.5989,
      "theoretical_loss": 3.3984989870096536,
      "tokens_seen": 2239889408
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1235978,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.565056800842285,
      "objective/train/theoretical_loss": 3.398490879928598,
      "objective/train/tokens_used": 598814176,
      "theoretical_loss": 3.398490879928598,
      "tokens_seen": 2239954944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267878787878788,
      "loss": 2.6903,
      "theoretical_loss": 3.398490879928598,
      "tokens_seen": 2239954944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267676767676768,
      "loss": 2.5615,
      "theoretical_loss": 3.3984827731511458,
      "tokens_seen": 2240020480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032674747474747474,
      "loss": 2.5883,
      "theoretical_loss": 3.398474666677277,
      "tokens_seen": 2240086016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267272727272727,
      "loss": 2.466,
      "theoretical_loss": 3.3984665605069715,
      "tokens_seen": 2240151552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267070707070707,
      "loss": 2.5902,
      "theoretical_loss": 3.398458454640209,
      "tokens_seen": 2240217088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266868686868687,
      "loss": 2.3745,
      "theoretical_loss": 3.3984503490769686,
      "tokens_seen": 2240282624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266666666666667,
      "loss": 2.5181,
      "theoretical_loss": 3.3984422438172315,
      "tokens_seen": 2240348160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032664646464646463,
      "loss": 2.486,
      "theoretical_loss": 3.3984341388609756,
      "tokens_seen": 2240413696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032662626262626264,
      "loss": 2.594,
      "theoretical_loss": 3.3984260342081822,
      "tokens_seen": 2240479232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032660606060606065,
      "loss": 2.8373,
      "theoretical_loss": 3.3984179298588306,
      "tokens_seen": 2240544768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265858585858586,
      "loss": 2.4156,
      "theoretical_loss": 3.3984098258129003,
      "tokens_seen": 2240610304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032656565656565657,
      "loss": 2.7603,
      "theoretical_loss": 3.3984017220703713,
      "tokens_seen": 2240675840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265454545454545,
      "loss": 2.7061,
      "theoretical_loss": 3.398393618631223,
      "tokens_seen": 2240741376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032652525252525253,
      "loss": 2.7418,
      "theoretical_loss": 3.398385515495436,
      "tokens_seen": 2240806912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032650505050505055,
      "loss": 2.6515,
      "theoretical_loss": 3.398377412662989,
      "tokens_seen": 2240872448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264848484848485,
      "loss": 2.6523,
      "theoretical_loss": 3.3983693101338623,
      "tokens_seen": 2240937984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032646464646464646,
      "loss": 2.4976,
      "theoretical_loss": 3.398361207908036,
      "tokens_seen": 2241003520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264444444444444,
      "loss": 2.7018,
      "theoretical_loss": 3.3983531059854895,
      "tokens_seen": 2241069056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264242424242425,
      "loss": 2.5705,
      "theoretical_loss": 3.3983450043662025,
      "tokens_seen": 2241134592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032640404040404044,
      "loss": 2.7231,
      "theoretical_loss": 3.398336903050155,
      "tokens_seen": 2241200128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263838383838384,
      "loss": 2.5503,
      "theoretical_loss": 3.398328802037327,
      "tokens_seen": 2241265664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032636363636363635,
      "loss": 2.4337,
      "theoretical_loss": 3.3983207013276977,
      "tokens_seen": 2241331200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263434343434343,
      "loss": 2.8032,
      "theoretical_loss": 3.398312600921247,
      "tokens_seen": 2241396736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263232323232324,
      "loss": 2.5385,
      "theoretical_loss": 3.3983045008179555,
      "tokens_seen": 2241462272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032630303030303033,
      "loss": 2.2273,
      "theoretical_loss": 3.3982964010178014,
      "tokens_seen": 2241527808
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1237176,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4149110317230225,
      "objective/train/theoretical_loss": 3.3982883015207666,
      "objective/train/tokens_used": 600452576,
      "theoretical_loss": 3.3982883015207666,
      "tokens_seen": 2241593344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262828282828283,
      "loss": 2.4795,
      "theoretical_loss": 3.3982883015207666,
      "tokens_seen": 2241593344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032626262626262624,
      "loss": 2.603,
      "theoretical_loss": 3.398280202326829,
      "tokens_seen": 2241658880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032624242424242426,
      "loss": 2.5269,
      "theoretical_loss": 3.398272103435969,
      "tokens_seen": 2241724416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032622222222222227,
      "loss": 2.5095,
      "theoretical_loss": 3.398264004848167,
      "tokens_seen": 2241789952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262020202020202,
      "loss": 2.4808,
      "theoretical_loss": 3.398255906563402,
      "tokens_seen": 2241855488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261818181818182,
      "loss": 2.65,
      "theoretical_loss": 3.3982478085816545,
      "tokens_seen": 2241921024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032616161616161614,
      "loss": 2.4468,
      "theoretical_loss": 3.398239710902904,
      "tokens_seen": 2241986560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032614141414141415,
      "loss": 2.4904,
      "theoretical_loss": 3.3982316135271295,
      "tokens_seen": 2242052096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032612121212121216,
      "loss": 2.5256,
      "theoretical_loss": 3.3982235164543124,
      "tokens_seen": 2242117632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261010101010101,
      "loss": 2.7606,
      "theoretical_loss": 3.3982154196844316,
      "tokens_seen": 2242183168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260808080808081,
      "loss": 2.6065,
      "theoretical_loss": 3.3982073232174663,
      "tokens_seen": 2242248704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260606060606061,
      "loss": 2.6405,
      "theoretical_loss": 3.398199227053398,
      "tokens_seen": 2242314240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032604040404040404,
      "loss": 2.554,
      "theoretical_loss": 3.398191131192205,
      "tokens_seen": 2242379776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032602020202020205,
      "loss": 2.6045,
      "theoretical_loss": 3.3981830356338674,
      "tokens_seen": 2242445312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326,
      "loss": 2.5911,
      "theoretical_loss": 3.3981749403783654,
      "tokens_seen": 2242510848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032597979797979797,
      "loss": 2.4874,
      "theoretical_loss": 3.3981668454256786,
      "tokens_seen": 2242576384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000325959595959596,
      "loss": 2.5332,
      "theoretical_loss": 3.3981587507757873,
      "tokens_seen": 2242641920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032593939393939393,
      "loss": 2.6727,
      "theoretical_loss": 3.3981506564286708,
      "tokens_seen": 2242707456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032591919191919195,
      "loss": 2.688,
      "theoretical_loss": 3.398142562384309,
      "tokens_seen": 2242772992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258989898989899,
      "loss": 2.6281,
      "theoretical_loss": 3.3981344686426818,
      "tokens_seen": 2242838528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032587878787878786,
      "loss": 2.3954,
      "theoretical_loss": 3.398126375203769,
      "tokens_seen": 2242904064
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032585858585858587,
      "loss": 2.5233,
      "theoretical_loss": 3.3981182820675504,
      "tokens_seen": 2242969600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258383838383838,
      "loss": 2.6406,
      "theoretical_loss": 3.3981101892340058,
      "tokens_seen": 2243035136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032581818181818184,
      "loss": 2.7894,
      "theoretical_loss": 3.3981020967031155,
      "tokens_seen": 2243100672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003257979797979798,
      "loss": 2.6066,
      "theoretical_loss": 3.398094004474858,
      "tokens_seen": 2243166208
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 1238466,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8873229026794434,
      "objective/train/theoretical_loss": 3.3980859125492153,
      "objective/train/tokens_used": 602090976,
      "theoretical_loss": 3.3980859125492153,
      "tokens_seen": 2243231744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003257777777777778,
      "loss": 2.6947,
      "theoretical_loss": 3.3980859125492153,
      "tokens_seen": 2243231744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032575757575757576,
      "loss": 2.8324,
      "theoretical_loss": 3.3980778209261655,
      "tokens_seen": 2243297280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257373737373737,
      "loss": 2.6064,
      "theoretical_loss": 3.398069729605689,
      "tokens_seen": 2243362816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032571717171717173,
      "loss": 2.7454,
      "theoretical_loss": 3.3980616385877656,
      "tokens_seen": 2243428352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256969696969697,
      "loss": 2.7327,
      "theoretical_loss": 3.3980535478723755,
      "tokens_seen": 2243493888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256767676767677,
      "loss": 2.7072,
      "theoretical_loss": 3.398045457459498,
      "tokens_seen": 2243559424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032565656565656566,
      "loss": 2.7652,
      "theoretical_loss": 3.3980373673491133,
      "tokens_seen": 2243624960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256363636363636,
      "loss": 2.4725,
      "theoretical_loss": 3.398029277541201,
      "tokens_seen": 2243690496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256161616161616,
      "loss": 2.3954,
      "theoretical_loss": 3.398021188035741,
      "tokens_seen": 2243756032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032559595959595963,
      "loss": 2.2931,
      "theoretical_loss": 3.3980130988327133,
      "tokens_seen": 2243821568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255757575757576,
      "loss": 2.7455,
      "theoretical_loss": 3.398005009932098,
      "tokens_seen": 2243887104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032555555555555555,
      "loss": 2.5907,
      "theoretical_loss": 3.3979969213338745,
      "tokens_seen": 2243952640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255353535353535,
      "loss": 2.7126,
      "theoretical_loss": 3.3979888330380232,
      "tokens_seen": 2244018176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255151515151515,
      "loss": 2.6299,
      "theoretical_loss": 3.397980745044523,
      "tokens_seen": 2244083712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032549494949494953,
      "loss": 2.564,
      "theoretical_loss": 3.3979726573533546,
      "tokens_seen": 2244149248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254747474747475,
      "loss": 2.5305,
      "theoretical_loss": 3.397964569964498,
      "tokens_seen": 2244214784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032545454545454544,
      "loss": 2.6263,
      "theoretical_loss": 3.3979564828779325,
      "tokens_seen": 2244280320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254343434343434,
      "loss": 2.7636,
      "theoretical_loss": 3.3979483960936383,
      "tokens_seen": 2244345856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032541414141414146,
      "loss": 2.5791,
      "theoretical_loss": 3.3979403096115948,
      "tokens_seen": 2244411392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253939393939394,
      "loss": 2.6575,
      "theoretical_loss": 3.397932223431783,
      "tokens_seen": 2244476928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253737373737374,
      "loss": 2.4282,
      "theoretical_loss": 3.397924137554181,
      "tokens_seen": 2244542464
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032535353535353533,
      "loss": 2.4786,
      "theoretical_loss": 3.3979160519787706,
      "tokens_seen": 2244608000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253333333333333,
      "loss": 2.6864,
      "theoretical_loss": 3.3979079667055303,
      "tokens_seen": 2244673536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032531313131313136,
      "loss": 2.6525,
      "theoretical_loss": 3.397899881734441,
      "tokens_seen": 2244739072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003252929292929293,
      "loss": 2.7195,
      "theoretical_loss": 3.3978917970654816,
      "tokens_seen": 2244804608
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1238885,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1809070110321045,
      "objective/train/theoretical_loss": 3.3978837126986328,
      "objective/train/tokens_used": 603729376,
      "theoretical_loss": 3.3978837126986328,
      "tokens_seen": 2244870144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032527272727272727,
      "loss": 2.5608,
      "theoretical_loss": 3.3978837126986328,
      "tokens_seen": 2244870144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003252525252525252,
      "loss": 2.7261,
      "theoretical_loss": 3.397875628633874,
      "tokens_seen": 2244935680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032523232323232324,
      "loss": 2.6386,
      "theoretical_loss": 3.397867544871185,
      "tokens_seen": 2245001216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032521212121212125,
      "loss": 2.4322,
      "theoretical_loss": 3.397859461410546,
      "tokens_seen": 2245066752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003251919191919192,
      "loss": 2.5448,
      "theoretical_loss": 3.397851378251937,
      "tokens_seen": 2245132288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032517171717171716,
      "loss": 2.7242,
      "theoretical_loss": 3.397843295395338,
      "tokens_seen": 2245197824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003251515151515151,
      "loss": 2.48,
      "theoretical_loss": 3.3978352128407283,
      "tokens_seen": 2245263360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032513131313131313,
      "loss": 2.5817,
      "theoretical_loss": 3.3978271305880883,
      "tokens_seen": 2245328896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032511111111111114,
      "loss": 2.7349,
      "theoretical_loss": 3.3978190486373974,
      "tokens_seen": 2245394432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250909090909091,
      "loss": 2.56,
      "theoretical_loss": 3.3978109669886365,
      "tokens_seen": 2245459968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032507070707070705,
      "loss": 2.7129,
      "theoretical_loss": 3.3978028856417843,
      "tokens_seen": 2245525504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032505050505050507,
      "loss": 2.4765,
      "theoretical_loss": 3.3977948045968214,
      "tokens_seen": 2245591040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000325030303030303,
      "loss": 2.5874,
      "theoretical_loss": 3.397786723853728,
      "tokens_seen": 2245656576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032501010101010103,
      "loss": 2.4721,
      "theoretical_loss": 3.397778643412483,
      "tokens_seen": 2245722112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324989898989899,
      "loss": 2.6481,
      "theoretical_loss": 3.3977705632730673,
      "tokens_seen": 2245787648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032496969696969695,
      "loss": 2.6252,
      "theoretical_loss": 3.3977624834354603,
      "tokens_seen": 2245853184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032494949494949496,
      "loss": 2.3284,
      "theoretical_loss": 3.397754403899642,
      "tokens_seen": 2245918720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032492929292929297,
      "loss": 2.6229,
      "theoretical_loss": 3.397746324665593,
      "tokens_seen": 2245984256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003249090909090909,
      "loss": 2.6977,
      "theoretical_loss": 3.397738245733292,
      "tokens_seen": 2246049792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248888888888889,
      "loss": 2.5982,
      "theoretical_loss": 3.3977301671027194,
      "tokens_seen": 2246115328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248686868686869,
      "loss": 2.6435,
      "theoretical_loss": 3.3977220887738557,
      "tokens_seen": 2246180864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032484848484848485,
      "loss": 2.5501,
      "theoretical_loss": 3.39771401074668,
      "tokens_seen": 2246246400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032482828282828286,
      "loss": 2.6576,
      "theoretical_loss": 3.3977059330211725,
      "tokens_seen": 2246311936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248080808080808,
      "loss": 2.6838,
      "theoretical_loss": 3.397697855597314,
      "tokens_seen": 2246377472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247878787878788,
      "loss": 2.4642,
      "theoretical_loss": 3.3976897784750832,
      "tokens_seen": 2246443008
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1239988,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6558778285980225,
      "objective/train/theoretical_loss": 3.3976817016544603,
      "objective/train/tokens_used": 605367776,
      "theoretical_loss": 3.3976817016544603,
      "tokens_seen": 2246508544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247676767676768,
      "loss": 2.5854,
      "theoretical_loss": 3.3976817016544603,
      "tokens_seen": 2246508544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032474747474747474,
      "loss": 2.5479,
      "theoretical_loss": 3.397673625135426,
      "tokens_seen": 2246574080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032472727272727276,
      "loss": 2.4586,
      "theoretical_loss": 3.3976655489179595,
      "tokens_seen": 2246639616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247070707070707,
      "loss": 2.7558,
      "theoretical_loss": 3.397657473002041,
      "tokens_seen": 2246705152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246868686868687,
      "loss": 2.6734,
      "theoretical_loss": 3.3976493973876503,
      "tokens_seen": 2246770688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246666666666667,
      "loss": 2.5989,
      "theoretical_loss": 3.397641322074767,
      "tokens_seen": 2246836224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032464646464646464,
      "loss": 2.4503,
      "theoretical_loss": 3.397633247063372,
      "tokens_seen": 2246901760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032462626262626265,
      "loss": 2.4268,
      "theoretical_loss": 3.397625172353445,
      "tokens_seen": 2246967296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246060606060606,
      "loss": 2.7411,
      "theoretical_loss": 3.3976170979449654,
      "tokens_seen": 2247032832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245858585858586,
      "loss": 2.5106,
      "theoretical_loss": 3.397609023837913,
      "tokens_seen": 2247098368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032456565656565657,
      "loss": 2.5413,
      "theoretical_loss": 3.397600950032269,
      "tokens_seen": 2247163904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032454545454545453,
      "loss": 2.5425,
      "theoretical_loss": 3.397592876528012,
      "tokens_seen": 2247229440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032452525252525254,
      "loss": 2.6852,
      "theoretical_loss": 3.3975848033251226,
      "tokens_seen": 2247294976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032450505050505055,
      "loss": 2.5082,
      "theoretical_loss": 3.397576730423581,
      "tokens_seen": 2247360512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244848484848485,
      "loss": 2.6075,
      "theoretical_loss": 3.3975686578233666,
      "tokens_seen": 2247426048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032446464646464647,
      "loss": 2.5743,
      "theoretical_loss": 3.39756058552446,
      "tokens_seen": 2247491584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244444444444444,
      "loss": 2.566,
      "theoretical_loss": 3.3975525135268403,
      "tokens_seen": 2247557120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032442424242424243,
      "loss": 2.6145,
      "theoretical_loss": 3.397544441830488,
      "tokens_seen": 2247622656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032440404040404044,
      "loss": 2.585,
      "theoretical_loss": 3.3975363704353834,
      "tokens_seen": 2247688192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243838383838384,
      "loss": 2.4886,
      "theoretical_loss": 3.3975282993415057,
      "tokens_seen": 2247753728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032436363636363636,
      "loss": 2.6176,
      "theoretical_loss": 3.397520228548836,
      "tokens_seen": 2247819264
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243434343434343,
      "loss": 2.6563,
      "theoretical_loss": 3.397512158057353,
      "tokens_seen": 2247884800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243232323232324,
      "loss": 2.6648,
      "theoretical_loss": 3.397504087867037,
      "tokens_seen": 2247950336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032430303030303034,
      "loss": 2.6209,
      "theoretical_loss": 3.3974960179778684,
      "tokens_seen": 2248015872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242828282828283,
      "loss": 2.6033,
      "theoretical_loss": 3.397487948389827,
      "tokens_seen": 2248081408
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1241252,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.037294864654541,
      "objective/train/theoretical_loss": 3.3974798791028933,
      "objective/train/tokens_used": 607006176,
      "theoretical_loss": 3.3974798791028933,
      "tokens_seen": 2248146944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032426262626262625,
      "loss": 2.7454,
      "theoretical_loss": 3.3974798791028933,
      "tokens_seen": 2248146944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242424242424242,
      "loss": 2.5246,
      "theoretical_loss": 3.3974718101170462,
      "tokens_seen": 2248212480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032422222222222227,
      "loss": 2.6371,
      "theoretical_loss": 3.3974637414322664,
      "tokens_seen": 2248278016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032420202020202023,
      "loss": 2.8248,
      "theoretical_loss": 3.397455673048534,
      "tokens_seen": 2248343552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241818181818182,
      "loss": 2.7675,
      "theoretical_loss": 3.397447604965828,
      "tokens_seen": 2248409088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032416161616161614,
      "loss": 2.5905,
      "theoretical_loss": 3.3974395371841304,
      "tokens_seen": 2248474624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241414141414141,
      "loss": 2.6519,
      "theoretical_loss": 3.397431469703419,
      "tokens_seen": 2248540160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032412121212121217,
      "loss": 2.4771,
      "theoretical_loss": 3.397423402523675,
      "tokens_seen": 2248605696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241010101010101,
      "loss": 2.5812,
      "theoretical_loss": 3.397415335644878,
      "tokens_seen": 2248671232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240808080808081,
      "loss": 2.7362,
      "theoretical_loss": 3.3974072690670085,
      "tokens_seen": 2248736768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032406060606060604,
      "loss": 2.7021,
      "theoretical_loss": 3.397399202790046,
      "tokens_seen": 2248802304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032404040404040405,
      "loss": 2.5885,
      "theoretical_loss": 3.3973911368139706,
      "tokens_seen": 2248867840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032402020202020206,
      "loss": 2.7223,
      "theoretical_loss": 3.3973830711387625,
      "tokens_seen": 2248933376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324,
      "loss": 2.7474,
      "theoretical_loss": 3.397375005764401,
      "tokens_seen": 2248998912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032397979797979797,
      "loss": 2.5893,
      "theoretical_loss": 3.3973669406908673,
      "tokens_seen": 2249064448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032395959595959593,
      "loss": 2.6812,
      "theoretical_loss": 3.397358875918141,
      "tokens_seen": 2249129984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032393939393939394,
      "loss": 2.6245,
      "theoretical_loss": 3.397350811446201,
      "tokens_seen": 2249195520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032391919191919195,
      "loss": 2.246,
      "theoretical_loss": 3.397342747275029,
      "tokens_seen": 2249261056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238989898989899,
      "loss": 2.5908,
      "theoretical_loss": 3.397334683404604,
      "tokens_seen": 2249326592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032387878787878786,
      "loss": 2.6178,
      "theoretical_loss": 3.3973266198349066,
      "tokens_seen": 2249392128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238585858585859,
      "loss": 2.6177,
      "theoretical_loss": 3.3973185565659163,
      "tokens_seen": 2249457664
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032383838383838383,
      "loss": 2.5661,
      "theoretical_loss": 3.397310493597613,
      "tokens_seen": 2249523200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032381818181818184,
      "loss": 2.6387,
      "theoretical_loss": 3.3973024309299773,
      "tokens_seen": 2249588736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237979797979798,
      "loss": 2.6505,
      "theoretical_loss": 3.397294368562989,
      "tokens_seen": 2249654272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032377777777777776,
      "loss": 2.6009,
      "theoretical_loss": 3.3972863064966283,
      "tokens_seen": 2249719808
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1241776,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0098400115966797,
      "objective/train/theoretical_loss": 3.3972782447308747,
      "objective/train/tokens_used": 608644576,
      "theoretical_loss": 3.3972782447308747,
      "tokens_seen": 2249785344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032375757575757577,
      "loss": 2.7253,
      "theoretical_loss": 3.3972782447308747,
      "tokens_seen": 2249785344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237373737373737,
      "loss": 2.6584,
      "theoretical_loss": 3.397270183265709,
      "tokens_seen": 2249850880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032371717171717174,
      "loss": 2.6867,
      "theoretical_loss": 3.3972621221011106,
      "tokens_seen": 2249916416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236969696969697,
      "loss": 2.6644,
      "theoretical_loss": 3.3972540612370596,
      "tokens_seen": 2249981952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236767676767677,
      "loss": 2.7775,
      "theoretical_loss": 3.3972460006735363,
      "tokens_seen": 2250047488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032365656565656566,
      "loss": 2.5241,
      "theoretical_loss": 3.3972379404105206,
      "tokens_seen": 2250113024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236363636363636,
      "loss": 2.5691,
      "theoretical_loss": 3.397229880447993,
      "tokens_seen": 2250178560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032361616161616163,
      "loss": 2.5051,
      "theoretical_loss": 3.3972218207859326,
      "tokens_seen": 2250244096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235959595959596,
      "loss": 2.7053,
      "theoretical_loss": 3.3972137614243203,
      "tokens_seen": 2250309632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235757575757576,
      "loss": 2.6714,
      "theoretical_loss": 3.3972057023631357,
      "tokens_seen": 2250375168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032355555555555555,
      "loss": 2.5531,
      "theoretical_loss": 3.397197643602359,
      "tokens_seen": 2250440704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235353535353535,
      "loss": 2.1475,
      "theoretical_loss": 3.3971895851419704,
      "tokens_seen": 2250506240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235151515151515,
      "loss": 2.6421,
      "theoretical_loss": 3.3971815269819494,
      "tokens_seen": 2250571776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032349494949494953,
      "loss": 2.7133,
      "theoretical_loss": 3.397173469122277,
      "tokens_seen": 2250637312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234747474747475,
      "loss": 2.6396,
      "theoretical_loss": 3.3971654115629324,
      "tokens_seen": 2250702848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032345454545454545,
      "loss": 2.7516,
      "theoretical_loss": 3.397157354303896,
      "tokens_seen": 2250768384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234343434343434,
      "loss": 2.5693,
      "theoretical_loss": 3.397149297345148,
      "tokens_seen": 2250833920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234141414141414,
      "loss": 2.5616,
      "theoretical_loss": 3.397141240686668,
      "tokens_seen": 2250899456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233939393939394,
      "loss": 2.5604,
      "theoretical_loss": 3.3971331843284367,
      "tokens_seen": 2250964992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233737373737374,
      "loss": 2.647,
      "theoretical_loss": 3.3971251282704333,
      "tokens_seen": 2251030528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032335353535353534,
      "loss": 2.6121,
      "theoretical_loss": 3.397117072512639,
      "tokens_seen": 2251096064
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233333333333333,
      "loss": 2.7934,
      "theoretical_loss": 3.3971090170550333,
      "tokens_seen": 2251161600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032331313131313136,
      "loss": 2.7487,
      "theoretical_loss": 3.397100961897596,
      "tokens_seen": 2251227136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003232929292929293,
      "loss": 2.6383,
      "theoretical_loss": 3.3970929070403075,
      "tokens_seen": 2251292672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003232727272727273,
      "loss": 2.6685,
      "theoretical_loss": 3.3970848524831476,
      "tokens_seen": 2251358208
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1242897,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5388667583465576,
      "objective/train/theoretical_loss": 3.397076798226097,
      "objective/train/tokens_used": 610282976,
      "theoretical_loss": 3.397076798226097,
      "tokens_seen": 2251423744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032325252525252523,
      "loss": 2.8134,
      "theoretical_loss": 3.397076798226097,
      "tokens_seen": 2251423744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032323232323232324,
      "loss": 2.8143,
      "theoretical_loss": 3.397068744269135,
      "tokens_seen": 2251489280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032321212121212125,
      "loss": 2.8504,
      "theoretical_loss": 3.3970606906122423,
      "tokens_seen": 2251554816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231919191919192,
      "loss": 2.746,
      "theoretical_loss": 3.397052637255399,
      "tokens_seen": 2251620352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032317171717171717,
      "loss": 2.5253,
      "theoretical_loss": 3.3970445841985843,
      "tokens_seen": 2251685888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231515151515151,
      "loss": 2.6397,
      "theoretical_loss": 3.3970365314417794,
      "tokens_seen": 2251751424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231313131313132,
      "loss": 2.4531,
      "theoretical_loss": 3.397028478984964,
      "tokens_seen": 2251816960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032311111111111115,
      "loss": 2.5787,
      "theoretical_loss": 3.397020426828118,
      "tokens_seen": 2251882496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230909090909091,
      "loss": 2.3132,
      "theoretical_loss": 3.3970123749712213,
      "tokens_seen": 2251948032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032307070707070706,
      "loss": 2.6369,
      "theoretical_loss": 3.397004323414255,
      "tokens_seen": 2252013568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323050505050505,
      "loss": 2.7356,
      "theoretical_loss": 3.396996272157198,
      "tokens_seen": 2252079104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230303030303031,
      "loss": 2.5593,
      "theoretical_loss": 3.396988221200031,
      "tokens_seen": 2252144640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032301010101010104,
      "loss": 2.5869,
      "theoretical_loss": 3.396980170542734,
      "tokens_seen": 2252210176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322989898989899,
      "loss": 2.5806,
      "theoretical_loss": 3.3969721201852874,
      "tokens_seen": 2252275712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032296969696969695,
      "loss": 2.6232,
      "theoretical_loss": 3.3969640701276704,
      "tokens_seen": 2252341248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032294949494949496,
      "loss": 2.517,
      "theoretical_loss": 3.3969560203698643,
      "tokens_seen": 2252406784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322929292929293,
      "loss": 2.5175,
      "theoretical_loss": 3.3969479709118486,
      "tokens_seen": 2252472320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032290909090909093,
      "loss": 2.609,
      "theoretical_loss": 3.3969399217536034,
      "tokens_seen": 2252537856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228888888888889,
      "loss": 2.6138,
      "theoretical_loss": 3.3969318728951086,
      "tokens_seen": 2252603392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032286868686868685,
      "loss": 2.8452,
      "theoretical_loss": 3.396923824336345,
      "tokens_seen": 2252668928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032284848484848486,
      "loss": 2.9954,
      "theoretical_loss": 3.396915776077292,
      "tokens_seen": 2252734464
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032282828282828287,
      "loss": 2.5626,
      "theoretical_loss": 3.39690772811793,
      "tokens_seen": 2252800000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228080808080808,
      "loss": 2.6104,
      "theoretical_loss": 3.39689968045824,
      "tokens_seen": 2252865536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227878787878788,
      "loss": 2.5124,
      "theoretical_loss": 3.3968916330982006,
      "tokens_seen": 2252931072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032276767676767674,
      "loss": 2.7513,
      "theoretical_loss": 3.3968835860377924,
      "tokens_seen": 2252996608
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1243463,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8337297439575195,
      "objective/train/theoretical_loss": 3.396875539276996,
      "objective/train/tokens_used": 611921376,
      "theoretical_loss": 3.396875539276996,
      "tokens_seen": 2253062144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032274747474747475,
      "loss": 2.6388,
      "theoretical_loss": 3.396875539276996,
      "tokens_seen": 2253062144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032272727272727276,
      "loss": 2.8779,
      "theoretical_loss": 3.396867492815791,
      "tokens_seen": 2253127680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227070707070707,
      "loss": 2.5202,
      "theoretical_loss": 3.3968594466541586,
      "tokens_seen": 2253193216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226868686868687,
      "loss": 2.6369,
      "theoretical_loss": 3.3968514007920776,
      "tokens_seen": 2253258752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226666666666667,
      "loss": 2.5562,
      "theoretical_loss": 3.3968433552295285,
      "tokens_seen": 2253324288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032264646464646464,
      "loss": 2.785,
      "theoretical_loss": 3.396835309966492,
      "tokens_seen": 2253389824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032262626262626265,
      "loss": 2.5689,
      "theoretical_loss": 3.3968272650029476,
      "tokens_seen": 2253455360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226060606060606,
      "loss": 2.7323,
      "theoretical_loss": 3.396819220338876,
      "tokens_seen": 2253520896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032258585858585857,
      "loss": 2.6385,
      "theoretical_loss": 3.3968111759742565,
      "tokens_seen": 2253586432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225656565656566,
      "loss": 2.5644,
      "theoretical_loss": 3.39680313190907,
      "tokens_seen": 2253651968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032254545454545453,
      "loss": 2.5264,
      "theoretical_loss": 3.3967950881432967,
      "tokens_seen": 2253717504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032252525252525255,
      "loss": 2.5731,
      "theoretical_loss": 3.396787044676916,
      "tokens_seen": 2253783040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225050505050505,
      "loss": 2.7812,
      "theoretical_loss": 3.396779001509909,
      "tokens_seen": 2253848576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224848484848485,
      "loss": 2.7024,
      "theoretical_loss": 3.3967709586422554,
      "tokens_seen": 2253914112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032246464646464647,
      "loss": 2.8075,
      "theoretical_loss": 3.3967629160739348,
      "tokens_seen": 2253979648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032244444444444443,
      "loss": 2.526,
      "theoretical_loss": 3.3967548738049285,
      "tokens_seen": 2254045184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032242424242424244,
      "loss": 2.5916,
      "theoretical_loss": 3.3967468318352156,
      "tokens_seen": 2254110720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224040404040404,
      "loss": 2.9027,
      "theoretical_loss": 3.3967387901647768,
      "tokens_seen": 2254176256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223838383838384,
      "loss": 2.3612,
      "theoretical_loss": 3.3967307487935923,
      "tokens_seen": 2254241792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032236363636363636,
      "loss": 2.4982,
      "theoretical_loss": 3.396722707721642,
      "tokens_seen": 2254307328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223434343434343,
      "loss": 2.6329,
      "theoretical_loss": 3.3967146669489066,
      "tokens_seen": 2254372864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032232323232323233,
      "loss": 2.5713,
      "theoretical_loss": 3.396706626475366,
      "tokens_seen": 2254438400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032230303030303034,
      "loss": 2.6205,
      "theoretical_loss": 3.3966985863009995,
      "tokens_seen": 2254503936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222828282828283,
      "loss": 2.6128,
      "theoretical_loss": 3.396690546425788,
      "tokens_seen": 2254569472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032226262626262626,
      "loss": 2.625,
      "theoretical_loss": 3.396682506849712,
      "tokens_seen": 2254635008
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1244562,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.452125072479248,
      "objective/train/theoretical_loss": 3.3966744675727516,
      "objective/train/tokens_used": 613559776,
      "theoretical_loss": 3.3966744675727516,
      "tokens_seen": 2254700544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222424242424242,
      "loss": 2.7192,
      "theoretical_loss": 3.3966744675727516,
      "tokens_seen": 2254700544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222222222222222,
      "loss": 2.8083,
      "theoretical_loss": 3.3966664285948864,
      "tokens_seen": 2254766080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032220202020202024,
      "loss": 2.3679,
      "theoretical_loss": 3.396658389916097,
      "tokens_seen": 2254831616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221818181818182,
      "loss": 2.6082,
      "theoretical_loss": 3.3966503515363637,
      "tokens_seen": 2254897152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032216161616161615,
      "loss": 2.6304,
      "theoretical_loss": 3.3966423134556667,
      "tokens_seen": 2254962688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221414141414141,
      "loss": 2.5595,
      "theoretical_loss": 3.3966342756739856,
      "tokens_seen": 2255028224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032212121212121217,
      "loss": 2.6471,
      "theoretical_loss": 3.396626238191301,
      "tokens_seen": 2255093760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032210101010101013,
      "loss": 2.7279,
      "theoretical_loss": 3.396618201007593,
      "tokens_seen": 2255159296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220808080808081,
      "loss": 2.5665,
      "theoretical_loss": 3.396610164122842,
      "tokens_seen": 2255224832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032206060606060604,
      "loss": 2.573,
      "theoretical_loss": 3.3966021275370277,
      "tokens_seen": 2255290368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322040404040404,
      "loss": 2.3949,
      "theoretical_loss": 3.396594091250131,
      "tokens_seen": 2255355904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032202020202020206,
      "loss": 2.5372,
      "theoretical_loss": 3.3965860552621314,
      "tokens_seen": 2255421440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322,
      "loss": 2.5403,
      "theoretical_loss": 3.3965780195730098,
      "tokens_seen": 2255486976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321979797979798,
      "loss": 2.629,
      "theoretical_loss": 3.3965699841827455,
      "tokens_seen": 2255552512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032195959595959593,
      "loss": 2.5865,
      "theoretical_loss": 3.39656194909132,
      "tokens_seen": 2255618048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032193939393939395,
      "loss": 2.8571,
      "theoretical_loss": 3.3965539142987122,
      "tokens_seen": 2255683584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032191919191919196,
      "loss": 2.6604,
      "theoretical_loss": 3.396545879804903,
      "tokens_seen": 2255749120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218989898989899,
      "loss": 2.7366,
      "theoretical_loss": 3.3965378456098723,
      "tokens_seen": 2255814656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032187878787878787,
      "loss": 2.4626,
      "theoretical_loss": 3.3965298117136005,
      "tokens_seen": 2255880192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218585858585858,
      "loss": 2.5269,
      "theoretical_loss": 3.3965217781160675,
      "tokens_seen": 2255945728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032183838383838384,
      "loss": 2.773,
      "theoretical_loss": 3.3965137448172538,
      "tokens_seen": 2256011264
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032181818181818185,
      "loss": 2.769,
      "theoretical_loss": 3.39650571181714,
      "tokens_seen": 2256076800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003217979797979798,
      "loss": 2.6846,
      "theoretical_loss": 3.396497679115706,
      "tokens_seen": 2256142336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032177777777777776,
      "loss": 2.6453,
      "theoretical_loss": 3.3964896467129315,
      "tokens_seen": 2256207872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003217575757575758,
      "loss": 2.6132,
      "theoretical_loss": 3.3964816146087973,
      "tokens_seen": 2256273408
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1245051,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4939658641815186,
      "objective/train/theoretical_loss": 3.3964735828032833,
      "objective/train/tokens_used": 615198176,
      "theoretical_loss": 3.3964735828032833,
      "tokens_seen": 2256338944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032173737373737373,
      "loss": 2.5558,
      "theoretical_loss": 3.3964735828032833,
      "tokens_seen": 2256338944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032171717171717174,
      "loss": 2.6278,
      "theoretical_loss": 3.3964655512963704,
      "tokens_seen": 2256404480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216969696969697,
      "loss": 2.7621,
      "theoretical_loss": 3.396457520088038,
      "tokens_seen": 2256470016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032167676767676766,
      "loss": 2.4765,
      "theoretical_loss": 3.3964494891782664,
      "tokens_seen": 2256535552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032165656565656567,
      "loss": 2.5545,
      "theoretical_loss": 3.3964414585670366,
      "tokens_seen": 2256601088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216363636363636,
      "loss": 2.6117,
      "theoretical_loss": 3.396433428254328,
      "tokens_seen": 2256666624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032161616161616163,
      "loss": 2.4907,
      "theoretical_loss": 3.396425398240121,
      "tokens_seen": 2256732160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215959595959596,
      "loss": 2.5676,
      "theoretical_loss": 3.3964173685243964,
      "tokens_seen": 2256797696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215757575757576,
      "loss": 2.568,
      "theoretical_loss": 3.3964093391071337,
      "tokens_seen": 2256863232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032155555555555556,
      "loss": 2.6658,
      "theoretical_loss": 3.3964013099883137,
      "tokens_seen": 2256928768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032153535353535357,
      "loss": 2.5634,
      "theoretical_loss": 3.3963932811679163,
      "tokens_seen": 2256994304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032151515151515153,
      "loss": 2.587,
      "theoretical_loss": 3.3963852526459215,
      "tokens_seen": 2257059840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214949494949495,
      "loss": 2.6791,
      "theoretical_loss": 3.3963772244223107,
      "tokens_seen": 2257125376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214747474747475,
      "loss": 2.5604,
      "theoretical_loss": 3.3963691964970626,
      "tokens_seen": 2257190912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032145454545454545,
      "loss": 2.5811,
      "theoretical_loss": 3.3963611688701585,
      "tokens_seen": 2257256448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032143434343434346,
      "loss": 2.4722,
      "theoretical_loss": 3.3963531415415784,
      "tokens_seen": 2257321984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214141414141414,
      "loss": 2.7084,
      "theoretical_loss": 3.3963451145113024,
      "tokens_seen": 2257387520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032139393939393943,
      "loss": 2.6442,
      "theoretical_loss": 3.396337087779311,
      "tokens_seen": 2257453056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213737373737374,
      "loss": 2.6079,
      "theoretical_loss": 3.396329061345584,
      "tokens_seen": 2257518592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032135353535353534,
      "loss": 2.5841,
      "theoretical_loss": 3.396321035210102,
      "tokens_seen": 2257584128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032133333333333336,
      "loss": 2.6716,
      "theoretical_loss": 3.3963130093728453,
      "tokens_seen": 2257649664
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213131313131313,
      "loss": 2.5498,
      "theoretical_loss": 3.396304983833794,
      "tokens_seen": 2257715200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212929292929293,
      "loss": 2.5937,
      "theoretical_loss": 3.396296958592929,
      "tokens_seen": 2257780736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212727272727273,
      "loss": 2.5268,
      "theoretical_loss": 3.3962889336502293,
      "tokens_seen": 2257846272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032125252525252524,
      "loss": 2.4315,
      "theoretical_loss": 3.396280909005676,
      "tokens_seen": 2257911808
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1246329,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1417624950408936,
      "objective/train/theoretical_loss": 3.3962728846592496,
      "objective/train/tokens_used": 616836576,
      "theoretical_loss": 3.3962728846592496,
      "tokens_seen": 2257977344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032123232323232325,
      "loss": 2.5338,
      "theoretical_loss": 3.3962728846592496,
      "tokens_seen": 2257977344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032121212121212126,
      "loss": 2.4146,
      "theoretical_loss": 3.39626486061093,
      "tokens_seen": 2258042880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211919191919192,
      "loss": 2.5625,
      "theoretical_loss": 3.3962568368606973,
      "tokens_seen": 2258108416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211717171717172,
      "loss": 2.498,
      "theoretical_loss": 3.396248813408532,
      "tokens_seen": 2258173952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032115151515151513,
      "loss": 2.4951,
      "theoretical_loss": 3.3962407902544145,
      "tokens_seen": 2258239488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032113131313131314,
      "loss": 2.3995,
      "theoretical_loss": 3.396232767398325,
      "tokens_seen": 2258305024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032111111111111115,
      "loss": 2.4617,
      "theoretical_loss": 3.3962247448402434,
      "tokens_seen": 2258370560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210909090909091,
      "loss": 2.5313,
      "theoretical_loss": 3.3962167225801503,
      "tokens_seen": 2258436096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032107070707070707,
      "loss": 2.3825,
      "theoretical_loss": 3.3962087006180264,
      "tokens_seen": 2258501632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321050505050505,
      "loss": 2.6772,
      "theoretical_loss": 3.3962006789538512,
      "tokens_seen": 2258567168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210303030303031,
      "loss": 2.4764,
      "theoretical_loss": 3.3961926575876054,
      "tokens_seen": 2258632704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032101010101010104,
      "loss": 2.6091,
      "theoretical_loss": 3.396184636519269,
      "tokens_seen": 2258698240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000320989898989899,
      "loss": 2.4795,
      "theoretical_loss": 3.396176615748823,
      "tokens_seen": 2258763776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032096969696969696,
      "loss": 2.6561,
      "theoretical_loss": 3.3961685952762473,
      "tokens_seen": 2258829312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209494949494949,
      "loss": 2.4325,
      "theoretical_loss": 3.396160575101522,
      "tokens_seen": 2258894848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000320929292929293,
      "loss": 2.713,
      "theoretical_loss": 3.3961525552246274,
      "tokens_seen": 2258960384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032090909090909094,
      "loss": 2.5842,
      "theoretical_loss": 3.396144535645544,
      "tokens_seen": 2259025920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208888888888889,
      "loss": 2.8008,
      "theoretical_loss": 3.396136516364252,
      "tokens_seen": 2259091456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032086868686868685,
      "loss": 2.6915,
      "theoretical_loss": 3.3961284973807313,
      "tokens_seen": 2259156992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208484848484848,
      "loss": 2.4567,
      "theoretical_loss": 3.3961204786949635,
      "tokens_seen": 2259222528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208282828282829,
      "loss": 2.6053,
      "theoretical_loss": 3.3961124603069277,
      "tokens_seen": 2259288064
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032080808080808083,
      "loss": 2.7385,
      "theoretical_loss": 3.3961044422166045,
      "tokens_seen": 2259353600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003207878787878788,
      "loss": 2.425,
      "theoretical_loss": 3.396096424423974,
      "tokens_seen": 2259419136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032076767676767674,
      "loss": 2.6334,
      "theoretical_loss": 3.3960884069290174,
      "tokens_seen": 2259484672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032074747474747476,
      "loss": 2.5307,
      "theoretical_loss": 3.3960803897317136,
      "tokens_seen": 2259550208
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 1246940,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8410935401916504,
      "objective/train/theoretical_loss": 3.396072372832044,
      "objective/train/tokens_used": 618474976,
      "theoretical_loss": 3.396072372832044,
      "tokens_seen": 2259615744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032072727272727277,
      "loss": 2.6552,
      "theoretical_loss": 3.396072372832044,
      "tokens_seen": 2259615744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003207070707070707,
      "loss": 2.4731,
      "theoretical_loss": 3.396064356229989,
      "tokens_seen": 2259681280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206868686868687,
      "loss": 2.6754,
      "theoretical_loss": 3.3960563399255284,
      "tokens_seen": 2259746816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032066666666666664,
      "loss": 2.6882,
      "theoretical_loss": 3.3960483239186425,
      "tokens_seen": 2259812352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032064646464646465,
      "loss": 2.5887,
      "theoretical_loss": 3.396040308209312,
      "tokens_seen": 2259877888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032062626262626266,
      "loss": 2.593,
      "theoretical_loss": 3.396032292797517,
      "tokens_seen": 2259943424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206060606060606,
      "loss": 2.6753,
      "theoretical_loss": 3.396024277683238,
      "tokens_seen": 2260008960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032058585858585857,
      "loss": 2.5601,
      "theoretical_loss": 3.3960162628664547,
      "tokens_seen": 2260074496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205656565656566,
      "loss": 2.3922,
      "theoretical_loss": 3.3960082483471483,
      "tokens_seen": 2260140032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032054545454545454,
      "loss": 2.7442,
      "theoretical_loss": 3.3960002341252986,
      "tokens_seen": 2260205568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032052525252525255,
      "loss": 2.6287,
      "theoretical_loss": 3.395992220200886,
      "tokens_seen": 2260271104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205050505050505,
      "loss": 2.7109,
      "theoretical_loss": 3.3959842065738908,
      "tokens_seen": 2260336640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032048484848484847,
      "loss": 2.6713,
      "theoretical_loss": 3.395976193244294,
      "tokens_seen": 2260402176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204646464646465,
      "loss": 2.7537,
      "theoretical_loss": 3.395968180212075,
      "tokens_seen": 2260467712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032044444444444443,
      "loss": 2.5858,
      "theoretical_loss": 3.395960167477215,
      "tokens_seen": 2260533248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032042424242424244,
      "loss": 2.5759,
      "theoretical_loss": 3.3959521550396934,
      "tokens_seen": 2260598784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204040404040404,
      "loss": 2.4519,
      "theoretical_loss": 3.395944142899491,
      "tokens_seen": 2260664320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203838383838384,
      "loss": 2.3593,
      "theoretical_loss": 3.395936131056588,
      "tokens_seen": 2260729856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032036363636363637,
      "loss": 2.6424,
      "theoretical_loss": 3.3959281195109656,
      "tokens_seen": 2260795392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203434343434343,
      "loss": 2.7046,
      "theoretical_loss": 3.395920108262603,
      "tokens_seen": 2260860928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032032323232323234,
      "loss": 2.6292,
      "theoretical_loss": 3.3959120973114816,
      "tokens_seen": 2260926464
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203030303030303,
      "loss": 2.3901,
      "theoretical_loss": 3.3959040866575805,
      "tokens_seen": 2260992000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202828282828283,
      "loss": 2.607,
      "theoretical_loss": 3.395896076300881,
      "tokens_seen": 2261057536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032026262626262626,
      "loss": 2.6355,
      "theoretical_loss": 3.3958880662413633,
      "tokens_seen": 2261123072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202424242424242,
      "loss": 2.6611,
      "theoretical_loss": 3.3958800564790077,
      "tokens_seen": 2261188608
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1248049,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4141077995300293,
      "objective/train/theoretical_loss": 3.3958720470137944,
      "objective/train/tokens_used": 620113376,
      "theoretical_loss": 3.3958720470137944,
      "tokens_seen": 2261254144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032022222222222223,
      "loss": 2.3786,
      "theoretical_loss": 3.3958720470137944,
      "tokens_seen": 2261254144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032020202020202024,
      "loss": 2.4124,
      "theoretical_loss": 3.3958640378457043,
      "tokens_seen": 2261319680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201818181818182,
      "loss": 2.6535,
      "theoretical_loss": 3.395856028974717,
      "tokens_seen": 2261385216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032016161616161615,
      "loss": 2.5763,
      "theoretical_loss": 3.3958480204008135,
      "tokens_seen": 2261450752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201414141414141,
      "loss": 2.6341,
      "theoretical_loss": 3.3958400121239736,
      "tokens_seen": 2261516288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201212121212121,
      "loss": 2.666,
      "theoretical_loss": 3.395832004144178,
      "tokens_seen": 2261581824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032010101010101013,
      "loss": 2.802,
      "theoretical_loss": 3.395823996461407,
      "tokens_seen": 2261647360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003200808080808081,
      "loss": 2.5193,
      "theoretical_loss": 3.3958159890756416,
      "tokens_seen": 2261712896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032006060606060605,
      "loss": 2.5524,
      "theoretical_loss": 3.395807981986861,
      "tokens_seen": 2261778432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320040404040404,
      "loss": 2.6795,
      "theoretical_loss": 3.3957999751950467,
      "tokens_seen": 2261843968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032002020202020207,
      "loss": 2.3938,
      "theoretical_loss": 3.3957919687001783,
      "tokens_seen": 2261909504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032,
      "loss": 2.7354,
      "theoretical_loss": 3.3957839625022364,
      "tokens_seen": 2261975040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319979797979798,
      "loss": 2.5789,
      "theoretical_loss": 3.3957759566012014,
      "tokens_seen": 2262040576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031995959595959594,
      "loss": 2.618,
      "theoretical_loss": 3.395767950997054,
      "tokens_seen": 2262106112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003199393939393939,
      "loss": 2.5708,
      "theoretical_loss": 3.3957599456897745,
      "tokens_seen": 2262171648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031991919191919196,
      "loss": 2.6591,
      "theoretical_loss": 3.3957519406793426,
      "tokens_seen": 2262237184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198989898989899,
      "loss": 2.6823,
      "theoretical_loss": 3.3957439359657395,
      "tokens_seen": 2262302720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198787878787879,
      "loss": 2.8268,
      "theoretical_loss": 3.3957359315489453,
      "tokens_seen": 2262368256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031985858585858583,
      "loss": 2.7604,
      "theoretical_loss": 3.3957279274289403,
      "tokens_seen": 2262433792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031983838383838384,
      "loss": 2.7604,
      "theoretical_loss": 3.395719923605705,
      "tokens_seen": 2262499328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031981818181818185,
      "loss": 2.6101,
      "theoretical_loss": 3.39571192007922,
      "tokens_seen": 2262564864
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197979797979798,
      "loss": 2.6851,
      "theoretical_loss": 3.395703916849465,
      "tokens_seen": 2262630400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031977777777777777,
      "loss": 2.6499,
      "theoretical_loss": 3.3956959139164216,
      "tokens_seen": 2262695936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197575757575757,
      "loss": 2.3554,
      "theoretical_loss": 3.395687911280069,
      "tokens_seen": 2262761472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197373737373738,
      "loss": 2.7722,
      "theoretical_loss": 3.395679908940388,
      "tokens_seen": 2262827008
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1248635,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3614001274108887,
      "objective/train/theoretical_loss": 3.39567190689736,
      "objective/train/tokens_used": 621751776,
      "theoretical_loss": 3.39567190689736,
      "tokens_seen": 2262892544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031971717171717175,
      "loss": 2.4185,
      "theoretical_loss": 3.39567190689736,
      "tokens_seen": 2262892544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196969696969697,
      "loss": 2.6732,
      "theoretical_loss": 3.395663905150964,
      "tokens_seen": 2262958080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031967676767676766,
      "loss": 2.6997,
      "theoretical_loss": 3.3956559037011806,
      "tokens_seen": 2263023616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031965656565656567,
      "loss": 2.3822,
      "theoretical_loss": 3.395647902547991,
      "tokens_seen": 2263089152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196363636363637,
      "loss": 2.6004,
      "theoretical_loss": 3.395639901691375,
      "tokens_seen": 2263154688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031961616161616164,
      "loss": 2.627,
      "theoretical_loss": 3.3956319011313134,
      "tokens_seen": 2263220224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195959595959596,
      "loss": 2.4729,
      "theoretical_loss": 3.395623900867786,
      "tokens_seen": 2263285760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031957575757575755,
      "loss": 2.5198,
      "theoretical_loss": 3.3956159009007743,
      "tokens_seen": 2263351296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031955555555555556,
      "loss": 2.666,
      "theoretical_loss": 3.395607901230258,
      "tokens_seen": 2263416832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195353535353536,
      "loss": 2.3702,
      "theoretical_loss": 3.3955999018562175,
      "tokens_seen": 2263482368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031951515151515153,
      "loss": 2.5827,
      "theoretical_loss": 3.395591902778633,
      "tokens_seen": 2263547904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194949494949495,
      "loss": 2.5407,
      "theoretical_loss": 3.3955839039974856,
      "tokens_seen": 2263613440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031947474747474745,
      "loss": 2.4561,
      "theoretical_loss": 3.3955759055127555,
      "tokens_seen": 2263678976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031945454545454546,
      "loss": 2.701,
      "theoretical_loss": 3.3955679073244225,
      "tokens_seen": 2263744512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031943434343434347,
      "loss": 2.6367,
      "theoretical_loss": 3.3955599094324684,
      "tokens_seen": 2263810048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194141414141414,
      "loss": 2.7365,
      "theoretical_loss": 3.395551911836872,
      "tokens_seen": 2263875584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193939393939394,
      "loss": 2.5264,
      "theoretical_loss": 3.395543914537615,
      "tokens_seen": 2263941120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193737373737374,
      "loss": 2.603,
      "theoretical_loss": 3.3955359175346773,
      "tokens_seen": 2264006656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031935353535353535,
      "loss": 2.5396,
      "theoretical_loss": 3.3955279208280396,
      "tokens_seen": 2264072192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031933333333333336,
      "loss": 2.604,
      "theoretical_loss": 3.395519924417682,
      "tokens_seen": 2264137728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193131313131313,
      "loss": 2.8452,
      "theoretical_loss": 3.395511928303585,
      "tokens_seen": 2264203264
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192929292929293,
      "loss": 2.6547,
      "theoretical_loss": 3.3955039324857292,
      "tokens_seen": 2264268800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192727272727273,
      "loss": 2.3664,
      "theoretical_loss": 3.395495936964095,
      "tokens_seen": 2264334336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031925252525252524,
      "loss": 2.5833,
      "theoretical_loss": 3.395487941738663,
      "tokens_seen": 2264399872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031923232323232325,
      "loss": 2.688,
      "theoretical_loss": 3.395479946809414,
      "tokens_seen": 2264465408
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1249740,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7208354473114014,
      "objective/train/theoretical_loss": 3.3954719521763277,
      "objective/train/tokens_used": 623390176,
      "theoretical_loss": 3.3954719521763277,
      "tokens_seen": 2264530944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192121212121212,
      "loss": 2.5151,
      "theoretical_loss": 3.3954719521763277,
      "tokens_seen": 2264530944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191919191919192,
      "loss": 2.8347,
      "theoretical_loss": 3.3954639578393846,
      "tokens_seen": 2264596480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191717171717172,
      "loss": 2.8171,
      "theoretical_loss": 3.3954559637985655,
      "tokens_seen": 2264662016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031915151515151514,
      "loss": 2.4953,
      "theoretical_loss": 3.3954479700538505,
      "tokens_seen": 2264727552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031913131313131315,
      "loss": 2.7396,
      "theoretical_loss": 3.395439976605221,
      "tokens_seen": 2264793088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191111111111111,
      "loss": 2.5031,
      "theoretical_loss": 3.3954319834526565,
      "tokens_seen": 2264858624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003190909090909091,
      "loss": 2.3527,
      "theoretical_loss": 3.3954239905961376,
      "tokens_seen": 2264924160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031907070707070707,
      "loss": 2.5388,
      "theoretical_loss": 3.395415998035645,
      "tokens_seen": 2264989696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031905050505050503,
      "loss": 2.7247,
      "theoretical_loss": 3.3954080057711593,
      "tokens_seen": 2265055232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031903030303030304,
      "loss": 2.5036,
      "theoretical_loss": 3.3954000138026603,
      "tokens_seen": 2265120768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031901010101010105,
      "loss": 2.5102,
      "theoretical_loss": 3.3953920221301295,
      "tokens_seen": 2265186304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318989898989899,
      "loss": 2.6913,
      "theoretical_loss": 3.395384030753547,
      "tokens_seen": 2265251840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031896969696969696,
      "loss": 2.5066,
      "theoretical_loss": 3.3953760396728923,
      "tokens_seen": 2265317376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189494949494949,
      "loss": 2.7384,
      "theoretical_loss": 3.3953680488881472,
      "tokens_seen": 2265382912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031892929292929293,
      "loss": 2.5828,
      "theoretical_loss": 3.395360058399292,
      "tokens_seen": 2265448448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031890909090909094,
      "loss": 2.5857,
      "theoretical_loss": 3.3953520682063063,
      "tokens_seen": 2265513984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188888888888889,
      "loss": 2.5849,
      "theoretical_loss": 3.3953440783091713,
      "tokens_seen": 2265579520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031886868686868686,
      "loss": 2.5293,
      "theoretical_loss": 3.395336088707867,
      "tokens_seen": 2265645056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188484848484848,
      "loss": 2.6149,
      "theoretical_loss": 3.395328099402375,
      "tokens_seen": 2265710592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188282828282829,
      "loss": 2.7865,
      "theoretical_loss": 3.3953201103926745,
      "tokens_seen": 2265776128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031880808080808084,
      "loss": 2.5799,
      "theoretical_loss": 3.3953121216787467,
      "tokens_seen": 2265841664
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187878787878788,
      "loss": 2.425,
      "theoretical_loss": 3.395304133260572,
      "tokens_seen": 2265907200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031876767676767675,
      "loss": 2.6586,
      "theoretical_loss": 3.3952961451381305,
      "tokens_seen": 2265972736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187474747474747,
      "loss": 2.5189,
      "theoretical_loss": 3.3952881573114033,
      "tokens_seen": 2266038272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031872727272727277,
      "loss": 2.569,
      "theoretical_loss": 3.3952801697803707,
      "tokens_seen": 2266103808
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1250565,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.438335418701172,
      "objective/train/theoretical_loss": 3.3952721825450127,
      "objective/train/tokens_used": 625028576,
      "theoretical_loss": 3.3952721825450127,
      "tokens_seen": 2266169344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031870707070707073,
      "loss": 2.6011,
      "theoretical_loss": 3.3952721825450127,
      "tokens_seen": 2266169344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186868686868687,
      "loss": 2.8084,
      "theoretical_loss": 3.3952641956053102,
      "tokens_seen": 2266234880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031866666666666664,
      "loss": 2.483,
      "theoretical_loss": 3.395256208961244,
      "tokens_seen": 2266300416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031864646464646465,
      "loss": 2.6591,
      "theoretical_loss": 3.3952482226127945,
      "tokens_seen": 2266365952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031862626262626266,
      "loss": 2.486,
      "theoretical_loss": 3.3952402365599417,
      "tokens_seen": 2266431488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186060606060606,
      "loss": 2.6736,
      "theoretical_loss": 3.3952322508026667,
      "tokens_seen": 2266497024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003185858585858586,
      "loss": 2.5937,
      "theoretical_loss": 3.3952242653409495,
      "tokens_seen": 2266562560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031856565656565654,
      "loss": 2.5626,
      "theoretical_loss": 3.395216280174771,
      "tokens_seen": 2266628096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031854545454545455,
      "loss": 2.7643,
      "theoretical_loss": 3.3952082953041116,
      "tokens_seen": 2266693632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031852525252525256,
      "loss": 2.5137,
      "theoretical_loss": 3.395200310728952,
      "tokens_seen": 2266759168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003185050505050505,
      "loss": 2.5405,
      "theoretical_loss": 3.395192326449272,
      "tokens_seen": 2266824704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031848484848484847,
      "loss": 2.6636,
      "theoretical_loss": 3.395184342465053,
      "tokens_seen": 2266890240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184646464646465,
      "loss": 2.8203,
      "theoretical_loss": 3.395176358776275,
      "tokens_seen": 2266955776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031844444444444444,
      "loss": 2.8766,
      "theoretical_loss": 3.3951683753829194,
      "tokens_seen": 2267021312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031842424242424245,
      "loss": 2.7388,
      "theoretical_loss": 3.3951603922849656,
      "tokens_seen": 2267086848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184040404040404,
      "loss": 2.7884,
      "theoretical_loss": 3.395152409482394,
      "tokens_seen": 2267152384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031838383838383836,
      "loss": 2.4545,
      "theoretical_loss": 3.3951444269751865,
      "tokens_seen": 2267217920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183636363636364,
      "loss": 2.6157,
      "theoretical_loss": 3.395136444763323,
      "tokens_seen": 2267283456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031834343434343433,
      "loss": 2.478,
      "theoretical_loss": 3.3951284628467833,
      "tokens_seen": 2267348992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031832323232323234,
      "loss": 2.6954,
      "theoretical_loss": 3.3951204812255487,
      "tokens_seen": 2267414528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183030303030303,
      "loss": 2.6418,
      "theoretical_loss": 3.395112499899599,
      "tokens_seen": 2267480064
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182828282828283,
      "loss": 2.7822,
      "theoretical_loss": 3.395104518868916,
      "tokens_seen": 2267545600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031826262626262627,
      "loss": 2.7872,
      "theoretical_loss": 3.3950965381334797,
      "tokens_seen": 2267611136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182424242424242,
      "loss": 2.7075,
      "theoretical_loss": 3.39508855769327,
      "tokens_seen": 2267676672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031822222222222224,
      "loss": 2.2708,
      "theoretical_loss": 3.395080577548268,
      "tokens_seen": 2267742208
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1251506,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.883519411087036,
      "objective/train/theoretical_loss": 3.3950725976984546,
      "objective/train/tokens_used": 626666976,
      "theoretical_loss": 3.3950725976984546,
      "tokens_seen": 2267807744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182020202020202,
      "loss": 2.7139,
      "theoretical_loss": 3.3950725976984546,
      "tokens_seen": 2267807744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181818181818182,
      "loss": 2.4992,
      "theoretical_loss": 3.3950646181438096,
      "tokens_seen": 2267873280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031816161616161616,
      "loss": 2.5939,
      "theoretical_loss": 3.3950566388843137,
      "tokens_seen": 2267938816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181414141414141,
      "loss": 2.6849,
      "theoretical_loss": 3.395048659919948,
      "tokens_seen": 2268004352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031812121212121213,
      "loss": 2.5904,
      "theoretical_loss": 3.3950406812506926,
      "tokens_seen": 2268069888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031810101010101014,
      "loss": 2.5358,
      "theoretical_loss": 3.395032702876528,
      "tokens_seen": 2268135424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003180808080808081,
      "loss": 2.502,
      "theoretical_loss": 3.3950247247974352,
      "tokens_seen": 2268200960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031806060606060605,
      "loss": 2.4941,
      "theoretical_loss": 3.3950167470133943,
      "tokens_seen": 2268266496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031804040404040406,
      "loss": 2.6482,
      "theoretical_loss": 3.395008769524386,
      "tokens_seen": 2268332032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318020202020202,
      "loss": 2.6035,
      "theoretical_loss": 3.3950007923303906,
      "tokens_seen": 2268397568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031800000000000003,
      "loss": 2.4701,
      "theoretical_loss": 3.3949928154313893,
      "tokens_seen": 2268463104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317979797979798,
      "loss": 2.5139,
      "theoretical_loss": 3.3949848388273627,
      "tokens_seen": 2268528640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031795959595959595,
      "loss": 2.6684,
      "theoretical_loss": 3.3949768625182903,
      "tokens_seen": 2268594176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031793939393939396,
      "loss": 2.6147,
      "theoretical_loss": 3.394968886504154,
      "tokens_seen": 2268659712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031791919191919197,
      "loss": 2.5804,
      "theoretical_loss": 3.3949609107849334,
      "tokens_seen": 2268725248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178989898989899,
      "loss": 2.4854,
      "theoretical_loss": 3.39495293536061,
      "tokens_seen": 2268790784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178787878787879,
      "loss": 2.5769,
      "theoretical_loss": 3.394944960231163,
      "tokens_seen": 2268856320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031785858585858584,
      "loss": 2.6394,
      "theoretical_loss": 3.3949369853965745,
      "tokens_seen": 2268921856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031783838383838385,
      "loss": 2.5502,
      "theoretical_loss": 3.3949290108568237,
      "tokens_seen": 2268987392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031781818181818186,
      "loss": 2.5496,
      "theoretical_loss": 3.3949210366118923,
      "tokens_seen": 2269052928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177979797979798,
      "loss": 2.518,
      "theoretical_loss": 3.3949130626617605,
      "tokens_seen": 2269118464
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177777777777778,
      "loss": 2.6216,
      "theoretical_loss": 3.3949050890064085,
      "tokens_seen": 2269184000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031775757575757573,
      "loss": 2.3819,
      "theoretical_loss": 3.3948971156458176,
      "tokens_seen": 2269249536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177373737373738,
      "loss": 2.8586,
      "theoretical_loss": 3.394889142579968,
      "tokens_seen": 2269315072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031771717171717175,
      "loss": 2.6064,
      "theoretical_loss": 3.39488116980884,
      "tokens_seen": 2269380608
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1252104,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.231213092803955,
      "objective/train/theoretical_loss": 3.394873197332415,
      "objective/train/tokens_used": 628305376,
      "theoretical_loss": 3.394873197332415,
      "tokens_seen": 2269446144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176969696969697,
      "loss": 2.4436,
      "theoretical_loss": 3.394873197332415,
      "tokens_seen": 2269446144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031767676767676767,
      "loss": 2.4369,
      "theoretical_loss": 3.3948652251506726,
      "tokens_seen": 2269511680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176565656565656,
      "loss": 2.7869,
      "theoretical_loss": 3.394857253263594,
      "tokens_seen": 2269577216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176363636363637,
      "loss": 2.379,
      "theoretical_loss": 3.39484928167116,
      "tokens_seen": 2269642752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031761616161616165,
      "loss": 2.6299,
      "theoretical_loss": 3.3948413103733506,
      "tokens_seen": 2269708288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175959595959596,
      "loss": 2.6942,
      "theoretical_loss": 3.394833339370147,
      "tokens_seen": 2269773824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031757575757575756,
      "loss": 2.6772,
      "theoretical_loss": 3.394825368661529,
      "tokens_seen": 2269839360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175555555555555,
      "loss": 2.5731,
      "theoretical_loss": 3.3948173982474783,
      "tokens_seen": 2269904896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175353535353536,
      "loss": 2.7969,
      "theoretical_loss": 3.3948094281279744,
      "tokens_seen": 2269970432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031751515151515154,
      "loss": 2.4478,
      "theoretical_loss": 3.394801458302999,
      "tokens_seen": 2270035968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003174949494949495,
      "loss": 2.6576,
      "theoretical_loss": 3.3947934887725317,
      "tokens_seen": 2270101504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031747474747474745,
      "loss": 2.5791,
      "theoretical_loss": 3.394785519536554,
      "tokens_seen": 2270167040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031745454545454546,
      "loss": 2.6718,
      "theoretical_loss": 3.3947775505950455,
      "tokens_seen": 2270232576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003174343434343435,
      "loss": 2.4232,
      "theoretical_loss": 3.3947695819479877,
      "tokens_seen": 2270298112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031741414141414143,
      "loss": 2.575,
      "theoretical_loss": 3.394761613595361,
      "tokens_seen": 2270363648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173939393939394,
      "loss": 2.5744,
      "theoretical_loss": 3.394753645537146,
      "tokens_seen": 2270429184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031737373737373734,
      "loss": 2.535,
      "theoretical_loss": 3.394745677773323,
      "tokens_seen": 2270494720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031735353535353536,
      "loss": 2.4876,
      "theoretical_loss": 3.3947377103038736,
      "tokens_seen": 2270560256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031733333333333337,
      "loss": 2.6554,
      "theoretical_loss": 3.394729743128777,
      "tokens_seen": 2270625792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173131313131313,
      "loss": 2.7185,
      "theoretical_loss": 3.394721776248015,
      "tokens_seen": 2270691328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172929292929293,
      "loss": 2.4903,
      "theoretical_loss": 3.394713809661568,
      "tokens_seen": 2270756864
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172727272727273,
      "loss": 2.5748,
      "theoretical_loss": 3.3947058433694157,
      "tokens_seen": 2270822400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031725252525252525,
      "loss": 2.3493,
      "theoretical_loss": 3.3946978773715397,
      "tokens_seen": 2270887936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031723232323232326,
      "loss": 2.4245,
      "theoretical_loss": 3.394689911667921,
      "tokens_seen": 2270953472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172121212121212,
      "loss": 2.4038,
      "theoretical_loss": 3.394681946258539,
      "tokens_seen": 2271019008
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1253238,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6408684253692627,
      "objective/train/theoretical_loss": 3.394673981143375,
      "objective/train/tokens_used": 629943776,
      "theoretical_loss": 3.394673981143375,
      "tokens_seen": 2271084544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171919191919192,
      "loss": 2.5847,
      "theoretical_loss": 3.394673981143375,
      "tokens_seen": 2271084544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171717171717172,
      "loss": 2.5202,
      "theoretical_loss": 3.39466601632241,
      "tokens_seen": 2271150080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031715151515151514,
      "loss": 2.4437,
      "theoretical_loss": 3.394658051795624,
      "tokens_seen": 2271215616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031713131313131315,
      "loss": 2.4978,
      "theoretical_loss": 3.3946500875629977,
      "tokens_seen": 2271281152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171111111111111,
      "loss": 2.6541,
      "theoretical_loss": 3.3946421236245126,
      "tokens_seen": 2271346688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170909090909091,
      "loss": 2.6193,
      "theoretical_loss": 3.394634159980148,
      "tokens_seen": 2271412224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170707070707071,
      "loss": 2.3838,
      "theoretical_loss": 3.394626196629886,
      "tokens_seen": 2271477760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031705050505050503,
      "loss": 2.4387,
      "theoretical_loss": 3.3946182335737056,
      "tokens_seen": 2271543296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031703030303030305,
      "loss": 2.5306,
      "theoretical_loss": 3.3946102708115893,
      "tokens_seen": 2271608832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317010101010101,
      "loss": 2.5514,
      "theoretical_loss": 3.3946023083435164,
      "tokens_seen": 2271674368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316989898989899,
      "loss": 2.4871,
      "theoretical_loss": 3.394594346169468,
      "tokens_seen": 2271739904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031696969696969697,
      "loss": 2.6184,
      "theoretical_loss": 3.3945863842894246,
      "tokens_seen": 2271805440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003169494949494949,
      "loss": 2.6487,
      "theoretical_loss": 3.3945784227033675,
      "tokens_seen": 2271870976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031692929292929294,
      "loss": 2.6708,
      "theoretical_loss": 3.394570461411276,
      "tokens_seen": 2271936512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031690909090909095,
      "loss": 2.6389,
      "theoretical_loss": 3.394562500413133,
      "tokens_seen": 2272002048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168888888888889,
      "loss": 2.5883,
      "theoretical_loss": 3.3945545397089165,
      "tokens_seen": 2272067584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031686868686868686,
      "loss": 2.6557,
      "theoretical_loss": 3.394546579298609,
      "tokens_seen": 2272133120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168484848484848,
      "loss": 2.6129,
      "theoretical_loss": 3.3945386191821907,
      "tokens_seen": 2272198656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031682828282828283,
      "loss": 2.5841,
      "theoretical_loss": 3.394530659359642,
      "tokens_seen": 2272264192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031680808080808084,
      "loss": 2.4224,
      "theoretical_loss": 3.3945226998309446,
      "tokens_seen": 2272329728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167878787878788,
      "loss": 2.5636,
      "theoretical_loss": 3.3945147405960774,
      "tokens_seen": 2272395264
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031676767676767676,
      "loss": 2.3931,
      "theoretical_loss": 3.3945067816550223,
      "tokens_seen": 2272460800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167474747474747,
      "loss": 2.6219,
      "theoretical_loss": 3.39449882300776,
      "tokens_seen": 2272526336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167272727272728,
      "loss": 2.2824,
      "theoretical_loss": 3.394490864654271,
      "tokens_seen": 2272591872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031670707070707073,
      "loss": 2.4839,
      "theoretical_loss": 3.394482906594536,
      "tokens_seen": 2272657408
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1253856,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.699746608734131,
      "objective/train/theoretical_loss": 3.394474948828535,
      "objective/train/tokens_used": 631582176,
      "theoretical_loss": 3.394474948828535,
      "tokens_seen": 2272722944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166868686868687,
      "loss": 2.582,
      "theoretical_loss": 3.394474948828535,
      "tokens_seen": 2272722944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031666666666666665,
      "loss": 2.5631,
      "theoretical_loss": 3.39446699135625,
      "tokens_seen": 2272788480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166464646464646,
      "loss": 2.6504,
      "theoretical_loss": 3.39445903417766,
      "tokens_seen": 2272854016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031662626262626267,
      "loss": 2.6006,
      "theoretical_loss": 3.394451077292748,
      "tokens_seen": 2272919552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166060606060606,
      "loss": 2.567,
      "theoretical_loss": 3.3944431207014922,
      "tokens_seen": 2272985088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165858585858586,
      "loss": 2.744,
      "theoretical_loss": 3.394435164403875,
      "tokens_seen": 2273050624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031656565656565654,
      "loss": 2.679,
      "theoretical_loss": 3.3944272083998763,
      "tokens_seen": 2273116160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031654545454545455,
      "loss": 2.3544,
      "theoretical_loss": 3.3944192526894774,
      "tokens_seen": 2273181696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031652525252525256,
      "loss": 2.791,
      "theoretical_loss": 3.394411297272659,
      "tokens_seen": 2273247232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165050505050505,
      "loss": 2.684,
      "theoretical_loss": 3.3944033421494004,
      "tokens_seen": 2273312768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003164848484848485,
      "loss": 2.5749,
      "theoretical_loss": 3.394395387319684,
      "tokens_seen": 2273378304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031646464646464643,
      "loss": 2.6576,
      "theoretical_loss": 3.39438743278349,
      "tokens_seen": 2273443840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031644444444444444,
      "loss": 2.4809,
      "theoretical_loss": 3.394379478540799,
      "tokens_seen": 2273509376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031642424242424246,
      "loss": 2.5898,
      "theoretical_loss": 3.394371524591591,
      "tokens_seen": 2273574912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003164040404040404,
      "loss": 2.5077,
      "theoretical_loss": 3.394363570935848,
      "tokens_seen": 2273640448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031638383838383837,
      "loss": 2.4831,
      "theoretical_loss": 3.39435561757355,
      "tokens_seen": 2273705984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163636363636364,
      "loss": 2.581,
      "theoretical_loss": 3.3943476645046777,
      "tokens_seen": 2273771520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163434343434344,
      "loss": 2.5112,
      "theoretical_loss": 3.394339711729212,
      "tokens_seen": 2273837056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031632323232323235,
      "loss": 2.6041,
      "theoretical_loss": 3.394331759247134,
      "tokens_seen": 2273902592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163030303030303,
      "loss": 2.5754,
      "theoretical_loss": 3.3943238070584236,
      "tokens_seen": 2273968128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031628282828282826,
      "loss": 2.701,
      "theoretical_loss": 3.394315855163062,
      "tokens_seen": 2274033664
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162626262626263,
      "loss": 2.4644,
      "theoretical_loss": 3.39430790356103,
      "tokens_seen": 2274099200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162424242424243,
      "loss": 2.5318,
      "theoretical_loss": 3.394299952252308,
      "tokens_seen": 2274164736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031622222222222224,
      "loss": 2.7892,
      "theoretical_loss": 3.3942920012368765,
      "tokens_seen": 2274230272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162020202020202,
      "loss": 2.5052,
      "theoretical_loss": 3.394284050514717,
      "tokens_seen": 2274295808
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1254938,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9091999530792236,
      "objective/train/theoretical_loss": 3.39427610008581,
      "objective/train/tokens_used": 633220576,
      "theoretical_loss": 3.39427610008581,
      "tokens_seen": 2274361344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031618181818181815,
      "loss": 2.8026,
      "theoretical_loss": 3.39427610008581,
      "tokens_seen": 2274361344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031616161616161617,
      "loss": 2.6426,
      "theoretical_loss": 3.394268149950136,
      "tokens_seen": 2274426880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161414141414142,
      "loss": 2.3218,
      "theoretical_loss": 3.394260200107676,
      "tokens_seen": 2274492416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031612121212121213,
      "loss": 2.4472,
      "theoretical_loss": 3.3942522505584103,
      "tokens_seen": 2274557952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161010101010101,
      "loss": 2.801,
      "theoretical_loss": 3.39424430130232,
      "tokens_seen": 2274623488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160808080808081,
      "loss": 2.6436,
      "theoretical_loss": 3.3942363523393855,
      "tokens_seen": 2274689024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031606060606060606,
      "loss": 2.6681,
      "theoretical_loss": 3.3942284036695884,
      "tokens_seen": 2274754560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031604040404040407,
      "loss": 2.4765,
      "theoretical_loss": 3.394220455292908,
      "tokens_seen": 2274820096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316020202020202,
      "loss": 2.4176,
      "theoretical_loss": 3.3942125072093265,
      "tokens_seen": 2274885632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316,
      "loss": 2.6185,
      "theoretical_loss": 3.394204559418824,
      "tokens_seen": 2274951168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000315979797979798,
      "loss": 2.5395,
      "theoretical_loss": 3.394196611921381,
      "tokens_seen": 2275016704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031595959595959595,
      "loss": 2.8416,
      "theoretical_loss": 3.394188664716979,
      "tokens_seen": 2275082240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031593939393939396,
      "loss": 2.58,
      "theoretical_loss": 3.3941807178055976,
      "tokens_seen": 2275147776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159191919191919,
      "loss": 2.4582,
      "theoretical_loss": 3.3941727711872187,
      "tokens_seen": 2275213312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031589898989898993,
      "loss": 2.5861,
      "theoretical_loss": 3.394164824861823,
      "tokens_seen": 2275278848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158787878787879,
      "loss": 2.6339,
      "theoretical_loss": 3.3941568788293903,
      "tokens_seen": 2275344384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031585858585858584,
      "loss": 2.588,
      "theoretical_loss": 3.3941489330899017,
      "tokens_seen": 2275409920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031583838383838385,
      "loss": 2.5906,
      "theoretical_loss": 3.394140987643339,
      "tokens_seen": 2275475456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158181818181818,
      "loss": 2.6659,
      "theoretical_loss": 3.394133042489681,
      "tokens_seen": 2275540992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157979797979798,
      "loss": 2.6415,
      "theoretical_loss": 3.3941250976289106,
      "tokens_seen": 2275606528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157777777777778,
      "loss": 2.4657,
      "theoretical_loss": 3.3941171530610075,
      "tokens_seen": 2275672064
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031575757575757574,
      "loss": 2.2953,
      "theoretical_loss": 3.3941092087859523,
      "tokens_seen": 2275737600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031573737373737375,
      "loss": 2.6089,
      "theoretical_loss": 3.394101264803726,
      "tokens_seen": 2275803136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031571717171717176,
      "loss": 2.6701,
      "theoretical_loss": 3.3940933211143096,
      "tokens_seen": 2275868672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003156969696969697,
      "loss": 2.6224,
      "theoretical_loss": 3.3940853777176834,
      "tokens_seen": 2275934208
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 1256312,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5594100952148438,
      "objective/train/theoretical_loss": 3.3940774346138287,
      "objective/train/tokens_used": 634858976,
      "theoretical_loss": 3.3940774346138287,
      "tokens_seen": 2275999744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031567676767676767,
      "loss": 2.5569,
      "theoretical_loss": 3.3940774346138287,
      "tokens_seen": 2275999744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031565656565656563,
      "loss": 2.5351,
      "theoretical_loss": 3.394069491802726,
      "tokens_seen": 2276065280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031563636363636364,
      "loss": 2.6004,
      "theoretical_loss": 3.394061549284356,
      "tokens_seen": 2276130816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031561616161616165,
      "loss": 2.5797,
      "theoretical_loss": 3.3940536070586997,
      "tokens_seen": 2276196352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155959595959596,
      "loss": 2.5316,
      "theoretical_loss": 3.394045665125738,
      "tokens_seen": 2276261888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031557575757575757,
      "loss": 2.2698,
      "theoretical_loss": 3.394037723485451,
      "tokens_seen": 2276327424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155555555555555,
      "loss": 2.6141,
      "theoretical_loss": 3.39402978213782,
      "tokens_seen": 2276392960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155353535353536,
      "loss": 2.5721,
      "theoretical_loss": 3.394021841082826,
      "tokens_seen": 2276458496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031551515151515154,
      "loss": 2.562,
      "theoretical_loss": 3.3940139003204495,
      "tokens_seen": 2276524032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154949494949495,
      "loss": 2.5509,
      "theoretical_loss": 3.3940059598506718,
      "tokens_seen": 2276589568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031547474747474746,
      "loss": 2.5295,
      "theoretical_loss": 3.3939980196734725,
      "tokens_seen": 2276655104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154545454545454,
      "loss": 2.5415,
      "theoretical_loss": 3.3939900797888334,
      "tokens_seen": 2276720640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154343434343435,
      "loss": 2.6936,
      "theoretical_loss": 3.3939821401967354,
      "tokens_seen": 2276786176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031541414141414144,
      "loss": 2.5969,
      "theoretical_loss": 3.3939742008971585,
      "tokens_seen": 2276851712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153939393939394,
      "loss": 2.6559,
      "theoretical_loss": 3.393966261890084,
      "tokens_seen": 2276917248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031537373737373735,
      "loss": 2.5614,
      "theoretical_loss": 3.3939583231754926,
      "tokens_seen": 2276982784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031535353535353536,
      "loss": 2.5683,
      "theoretical_loss": 3.3939503847533654,
      "tokens_seen": 2277048320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031533333333333337,
      "loss": 2.746,
      "theoretical_loss": 3.393942446623683,
      "tokens_seen": 2277113856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031531313131313133,
      "loss": 2.7708,
      "theoretical_loss": 3.3939345087864257,
      "tokens_seen": 2277179392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152929292929293,
      "loss": 2.586,
      "theoretical_loss": 3.393926571241575,
      "tokens_seen": 2277244928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031527272727272724,
      "loss": 2.5205,
      "theoretical_loss": 3.3939186339891116,
      "tokens_seen": 2277310464
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031525252525252525,
      "loss": 2.6002,
      "theoretical_loss": 3.3939106970290163,
      "tokens_seen": 2277376000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031523232323232327,
      "loss": 2.538,
      "theoretical_loss": 3.3939027603612697,
      "tokens_seen": 2277441536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152121212121212,
      "loss": 2.5522,
      "theoretical_loss": 3.3938948239858524,
      "tokens_seen": 2277507072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151919191919192,
      "loss": 2.4125,
      "theoretical_loss": 3.393886887902746,
      "tokens_seen": 2277572608
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1256724,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7904176712036133,
      "objective/train/theoretical_loss": 3.3938789521119306,
      "objective/train/tokens_used": 636497376,
      "theoretical_loss": 3.3938789521119306,
      "tokens_seen": 2277638144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151717171717172,
      "loss": 2.5879,
      "theoretical_loss": 3.3938789521119306,
      "tokens_seen": 2277638144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031515151515151515,
      "loss": 2.6244,
      "theoretical_loss": 3.393871016613388,
      "tokens_seen": 2277703680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031513131313131316,
      "loss": 2.824,
      "theoretical_loss": 3.3938630814070976,
      "tokens_seen": 2277769216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151111111111111,
      "loss": 2.6226,
      "theoretical_loss": 3.393855146493041,
      "tokens_seen": 2277834752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031509090909090907,
      "loss": 2.5232,
      "theoretical_loss": 3.3938472118711993,
      "tokens_seen": 2277900288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150707070707071,
      "loss": 2.5386,
      "theoretical_loss": 3.393839277541553,
      "tokens_seen": 2277965824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031505050505050504,
      "loss": 2.4923,
      "theoretical_loss": 3.393831343504083,
      "tokens_seen": 2278031360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031503030303030305,
      "loss": 2.3651,
      "theoretical_loss": 3.3938234097587703,
      "tokens_seen": 2278096896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000315010101010101,
      "loss": 2.5789,
      "theoretical_loss": 3.393815476305595,
      "tokens_seen": 2278162432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314989898989899,
      "loss": 2.4717,
      "theoretical_loss": 3.393807543144539,
      "tokens_seen": 2278227968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314969696969697,
      "loss": 2.3875,
      "theoretical_loss": 3.393799610275582,
      "tokens_seen": 2278293504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031494949494949493,
      "loss": 2.6223,
      "theoretical_loss": 3.3937916776987063,
      "tokens_seen": 2278359040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031492929292929294,
      "loss": 2.7733,
      "theoretical_loss": 3.3937837454138915,
      "tokens_seen": 2278424576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149090909090909,
      "loss": 2.5053,
      "theoretical_loss": 3.3937758134211187,
      "tokens_seen": 2278490112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148888888888889,
      "loss": 2.2793,
      "theoretical_loss": 3.393767881720369,
      "tokens_seen": 2278555648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031486868686868687,
      "loss": 2.5787,
      "theoretical_loss": 3.3937599503116234,
      "tokens_seen": 2278621184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148484848484848,
      "loss": 2.6155,
      "theoretical_loss": 3.3937520191948622,
      "tokens_seen": 2278686720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031482828282828284,
      "loss": 2.5502,
      "theoretical_loss": 3.3937440883700667,
      "tokens_seen": 2278752256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031480808080808085,
      "loss": 2.6447,
      "theoretical_loss": 3.3937361578372176,
      "tokens_seen": 2278817792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147878787878788,
      "loss": 2.3951,
      "theoretical_loss": 3.393728227596296,
      "tokens_seen": 2278883328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031476767676767676,
      "loss": 2.685,
      "theoretical_loss": 3.393720297647282,
      "tokens_seen": 2278948864
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147474747474747,
      "loss": 2.6433,
      "theoretical_loss": 3.3937123679901573,
      "tokens_seen": 2279014400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031472727272727273,
      "loss": 2.5549,
      "theoretical_loss": 3.3937044386249022,
      "tokens_seen": 2279079936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031470707070707074,
      "loss": 2.6744,
      "theoretical_loss": 3.393696509551498,
      "tokens_seen": 2279145472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146868686868687,
      "loss": 2.6514,
      "theoretical_loss": 3.3936885807699255,
      "tokens_seen": 2279211008
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1257757,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0417563915252686,
      "objective/train/theoretical_loss": 3.3936806522801657,
      "objective/train/tokens_used": 638135776,
      "theoretical_loss": 3.3936806522801657,
      "tokens_seen": 2279276544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031466666666666665,
      "loss": 2.3856,
      "theoretical_loss": 3.3936806522801657,
      "tokens_seen": 2279276544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031464646464646466,
      "loss": 2.771,
      "theoretical_loss": 3.3936727240821987,
      "tokens_seen": 2279342080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146262626262627,
      "loss": 2.5876,
      "theoretical_loss": 3.3936647961760062,
      "tokens_seen": 2279407616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031460606060606063,
      "loss": 2.7307,
      "theoretical_loss": 3.393656868561569,
      "tokens_seen": 2279473152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145858585858586,
      "loss": 2.5587,
      "theoretical_loss": 3.393648941238867,
      "tokens_seen": 2279538688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031456565656565655,
      "loss": 2.5594,
      "theoretical_loss": 3.3936410142078826,
      "tokens_seen": 2279604224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031454545454545456,
      "loss": 2.6003,
      "theoretical_loss": 3.3936330874685954,
      "tokens_seen": 2279669760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031452525252525257,
      "loss": 2.6426,
      "theoretical_loss": 3.393625161020987,
      "tokens_seen": 2279735296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145050505050505,
      "loss": 2.613,
      "theoretical_loss": 3.393617234865038,
      "tokens_seen": 2279800832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144848484848485,
      "loss": 2.761,
      "theoretical_loss": 3.3936093090007295,
      "tokens_seen": 2279866368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031446464646464644,
      "loss": 2.478,
      "theoretical_loss": 3.393601383428042,
      "tokens_seen": 2279931904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144444444444445,
      "loss": 2.5253,
      "theoretical_loss": 3.393593458146957,
      "tokens_seen": 2279997440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031442424242424246,
      "loss": 2.5476,
      "theoretical_loss": 3.393585533157455,
      "tokens_seen": 2280062976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144040404040404,
      "loss": 2.3795,
      "theoretical_loss": 3.393577608459516,
      "tokens_seen": 2280128512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143838383838384,
      "loss": 2.6271,
      "theoretical_loss": 3.393569684053123,
      "tokens_seen": 2280194048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031436363636363633,
      "loss": 2.5094,
      "theoretical_loss": 3.393561759938255,
      "tokens_seen": 2280259584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143434343434344,
      "loss": 2.4484,
      "theoretical_loss": 3.393553836114894,
      "tokens_seen": 2280325120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031432323232323235,
      "loss": 2.5701,
      "theoretical_loss": 3.3935459125830203,
      "tokens_seen": 2280390656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143030303030303,
      "loss": 2.6625,
      "theoretical_loss": 3.393537989342615,
      "tokens_seen": 2280456192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031428282828282827,
      "loss": 2.6173,
      "theoretical_loss": 3.3935300663936587,
      "tokens_seen": 2280521728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142626262626262,
      "loss": 2.6235,
      "theoretical_loss": 3.393522143736133,
      "tokens_seen": 2280587264
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142424242424243,
      "loss": 2.599,
      "theoretical_loss": 3.3935142213700185,
      "tokens_seen": 2280652800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031422222222222225,
      "loss": 2.448,
      "theoretical_loss": 3.3935062992952956,
      "tokens_seen": 2280718336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142020202020202,
      "loss": 2.5578,
      "theoretical_loss": 3.393498377511946,
      "tokens_seen": 2280783872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031418181818181816,
      "loss": 2.6501,
      "theoretical_loss": 3.39349045601995,
      "tokens_seen": 2280849408
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1258472,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2951927185058594,
      "objective/train/theoretical_loss": 3.393482534819289,
      "objective/train/tokens_used": 639774176,
      "theoretical_loss": 3.393482534819289,
      "tokens_seen": 2280914944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031416161616161617,
      "loss": 2.5746,
      "theoretical_loss": 3.393482534819289,
      "tokens_seen": 2280914944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141414141414142,
      "loss": 2.5196,
      "theoretical_loss": 3.393474613909943,
      "tokens_seen": 2280980480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031412121212121214,
      "loss": 2.5398,
      "theoretical_loss": 3.393466693291894,
      "tokens_seen": 2281046016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141010101010101,
      "loss": 2.618,
      "theoretical_loss": 3.393458772965123,
      "tokens_seen": 2281111552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031408080808080805,
      "loss": 2.6383,
      "theoretical_loss": 3.3934508529296097,
      "tokens_seen": 2281177088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031406060606060606,
      "loss": 2.6989,
      "theoretical_loss": 3.3934429331853364,
      "tokens_seen": 2281242624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140404040404041,
      "loss": 2.5702,
      "theoretical_loss": 3.393435013732283,
      "tokens_seen": 2281308160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031402020202020203,
      "loss": 2.9019,
      "theoretical_loss": 3.3934270945704306,
      "tokens_seen": 2281373696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314,
      "loss": 2.5032,
      "theoretical_loss": 3.3934191756997603,
      "tokens_seen": 2281439232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313979797979798,
      "loss": 2.6049,
      "theoretical_loss": 3.3934112571202535,
      "tokens_seen": 2281504768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031395959595959596,
      "loss": 2.5911,
      "theoretical_loss": 3.3934033388318903,
      "tokens_seen": 2281570304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031393939393939397,
      "loss": 2.4743,
      "theoretical_loss": 3.3933954208346524,
      "tokens_seen": 2281635840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139191919191919,
      "loss": 2.7516,
      "theoretical_loss": 3.39338750312852,
      "tokens_seen": 2281701376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138989898989899,
      "loss": 2.5268,
      "theoretical_loss": 3.3933795857134745,
      "tokens_seen": 2281766912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138787878787879,
      "loss": 2.753,
      "theoretical_loss": 3.393371668589497,
      "tokens_seen": 2281832448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031385858585858585,
      "loss": 2.6866,
      "theoretical_loss": 3.3933637517565676,
      "tokens_seen": 2281897984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031383838383838386,
      "loss": 2.6907,
      "theoretical_loss": 3.393355835214668,
      "tokens_seen": 2281963520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138181818181818,
      "loss": 2.49,
      "theoretical_loss": 3.393347918963779,
      "tokens_seen": 2282029056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031379797979797983,
      "loss": 2.5764,
      "theoretical_loss": 3.393340003003882,
      "tokens_seen": 2282094592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137777777777778,
      "loss": 2.6265,
      "theoretical_loss": 3.3933320873349566,
      "tokens_seen": 2282160128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031375757575757574,
      "loss": 2.485,
      "theoretical_loss": 3.393324171956985,
      "tokens_seen": 2282225664
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031373737373737375,
      "loss": 2.4023,
      "theoretical_loss": 3.3933162568699475,
      "tokens_seen": 2282291200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137171717171717,
      "loss": 2.559,
      "theoretical_loss": 3.3933083420738255,
      "tokens_seen": 2282356736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136969696969697,
      "loss": 2.5539,
      "theoretical_loss": 3.3933004275686,
      "tokens_seen": 2282422272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136767676767677,
      "loss": 2.6842,
      "theoretical_loss": 3.3932925133542513,
      "tokens_seen": 2282487808
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1259673,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.045727491378784,
      "objective/train/theoretical_loss": 3.393284599430761,
      "objective/train/tokens_used": 641412576,
      "theoretical_loss": 3.393284599430761,
      "tokens_seen": 2282553344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031365656565656563,
      "loss": 2.9195,
      "theoretical_loss": 3.393284599430761,
      "tokens_seen": 2282553344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031363636363636365,
      "loss": 2.5363,
      "theoretical_loss": 3.39327668579811,
      "tokens_seen": 2282618880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031361616161616166,
      "loss": 2.575,
      "theoretical_loss": 3.3932687724562784,
      "tokens_seen": 2282684416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135959595959596,
      "loss": 2.6229,
      "theoretical_loss": 3.393260859405249,
      "tokens_seen": 2282749952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031357575757575757,
      "loss": 2.5216,
      "theoretical_loss": 3.3932529466450005,
      "tokens_seen": 2282815488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031355555555555553,
      "loss": 2.5041,
      "theoretical_loss": 3.3932450341755156,
      "tokens_seen": 2282881024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031353535353535354,
      "loss": 2.7757,
      "theoretical_loss": 3.3932371219967745,
      "tokens_seen": 2282946560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031351515151515155,
      "loss": 2.5232,
      "theoretical_loss": 3.393229210108758,
      "tokens_seen": 2283012096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134949494949495,
      "loss": 2.4608,
      "theoretical_loss": 3.393221298511448,
      "tokens_seen": 2283077632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031347474747474746,
      "loss": 2.6039,
      "theoretical_loss": 3.3932133872048245,
      "tokens_seen": 2283143168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134545454545454,
      "loss": 2.5171,
      "theoretical_loss": 3.393205476188869,
      "tokens_seen": 2283208704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134343434343435,
      "loss": 2.5416,
      "theoretical_loss": 3.393197565463562,
      "tokens_seen": 2283274240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031341414141414144,
      "loss": 2.6953,
      "theoretical_loss": 3.393189655028885,
      "tokens_seen": 2283339776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133939393939394,
      "loss": 2.5893,
      "theoretical_loss": 3.393181744884819,
      "tokens_seen": 2283405312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031337373737373736,
      "loss": 2.7462,
      "theoretical_loss": 3.3931738350313445,
      "tokens_seen": 2283470848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133535353535353,
      "loss": 2.6833,
      "theoretical_loss": 3.393165925468443,
      "tokens_seen": 2283536384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133333333333334,
      "loss": 2.5097,
      "theoretical_loss": 3.393158016196095,
      "tokens_seen": 2283601920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031331313131313134,
      "loss": 2.5382,
      "theoretical_loss": 3.393150107214282,
      "tokens_seen": 2283667456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132929292929293,
      "loss": 2.5648,
      "theoretical_loss": 3.393142198522984,
      "tokens_seen": 2283732992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031327272727272725,
      "loss": 2.6896,
      "theoretical_loss": 3.3931342901221835,
      "tokens_seen": 2283798528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031325252525252526,
      "loss": 2.566,
      "theoretical_loss": 3.3931263820118605,
      "tokens_seen": 2283864064
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031323232323232327,
      "loss": 2.5805,
      "theoretical_loss": 3.393118474191996,
      "tokens_seen": 2283929600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031321212121212123,
      "loss": 2.615,
      "theoretical_loss": 3.393110566662571,
      "tokens_seen": 2283995136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131919191919192,
      "loss": 2.6735,
      "theoretical_loss": 3.393102659423567,
      "tokens_seen": 2284060672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031317171717171714,
      "loss": 2.6565,
      "theoretical_loss": 3.393094752474965,
      "tokens_seen": 2284126208
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1260242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6039111614227295,
      "objective/train/theoretical_loss": 3.393086845816746,
      "objective/train/tokens_used": 643050976,
      "theoretical_loss": 3.393086845816746,
      "tokens_seen": 2284191744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031315151515151515,
      "loss": 2.5128,
      "theoretical_loss": 3.393086845816746,
      "tokens_seen": 2284191744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031313131313131316,
      "loss": 2.6107,
      "theoretical_loss": 3.39307893944889,
      "tokens_seen": 2284257280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131111111111111,
      "loss": 2.7187,
      "theoretical_loss": 3.3930710333713785,
      "tokens_seen": 2284322816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130909090909091,
      "loss": 2.5281,
      "theoretical_loss": 3.393063127584193,
      "tokens_seen": 2284388352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130707070707071,
      "loss": 2.5965,
      "theoretical_loss": 3.3930552220873147,
      "tokens_seen": 2284453888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031305050505050505,
      "loss": 2.4025,
      "theoretical_loss": 3.3930473168807236,
      "tokens_seen": 2284519424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031303030303030306,
      "loss": 2.7882,
      "theoretical_loss": 3.3930394119644016,
      "tokens_seen": 2284584960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313010101010101,
      "loss": 2.7073,
      "theoretical_loss": 3.3930315073383293,
      "tokens_seen": 2284650496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031298989898989897,
      "loss": 2.5197,
      "theoretical_loss": 3.3930236030024874,
      "tokens_seen": 2284716032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000312969696969697,
      "loss": 2.449,
      "theoretical_loss": 3.393015698956858,
      "tokens_seen": 2284781568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031294949494949494,
      "loss": 2.4184,
      "theoretical_loss": 3.393007795201421,
      "tokens_seen": 2284847104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031292929292929295,
      "loss": 2.5209,
      "theoretical_loss": 3.3929998917361575,
      "tokens_seen": 2284912640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129090909090909,
      "loss": 2.4928,
      "theoretical_loss": 3.3929919885610493,
      "tokens_seen": 2284978176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031288888888888886,
      "loss": 2.4869,
      "theoretical_loss": 3.392984085676077,
      "tokens_seen": 2285043712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128686868686869,
      "loss": 2.4231,
      "theoretical_loss": 3.392976183081222,
      "tokens_seen": 2285109248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128484848484849,
      "loss": 2.5543,
      "theoretical_loss": 3.3929682807764645,
      "tokens_seen": 2285174784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031282828282828284,
      "loss": 2.6339,
      "theoretical_loss": 3.392960378761786,
      "tokens_seen": 2285240320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128080808080808,
      "loss": 2.6631,
      "theoretical_loss": 3.3929524770371677,
      "tokens_seen": 2285305856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127878787878788,
      "loss": 2.5875,
      "theoretical_loss": 3.3929445756025904,
      "tokens_seen": 2285371392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031276767676767677,
      "loss": 2.4776,
      "theoretical_loss": 3.392936674458035,
      "tokens_seen": 2285436928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127474747474748,
      "loss": 2.7435,
      "theoretical_loss": 3.392928773603483,
      "tokens_seen": 2285502464
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031272727272727273,
      "loss": 2.5056,
      "theoretical_loss": 3.392920873038915,
      "tokens_seen": 2285568000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127070707070707,
      "loss": 2.5386,
      "theoretical_loss": 3.3929129727643126,
      "tokens_seen": 2285633536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126868686868687,
      "loss": 2.7133,
      "theoretical_loss": 3.392905072779656,
      "tokens_seen": 2285699072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031266666666666666,
      "loss": 2.7472,
      "theoretical_loss": 3.392897173084927,
      "tokens_seen": 2285764608
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1261616,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.160874605178833,
      "objective/train/theoretical_loss": 3.392889273680106,
      "objective/train/tokens_used": 644689376,
      "theoretical_loss": 3.392889273680106,
      "tokens_seen": 2285830144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031264646464646467,
      "loss": 2.7724,
      "theoretical_loss": 3.392889273680106,
      "tokens_seen": 2285830144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126262626262626,
      "loss": 2.6675,
      "theoretical_loss": 3.3928813745651745,
      "tokens_seen": 2285895680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031260606060606064,
      "loss": 2.5848,
      "theoretical_loss": 3.3928734757401138,
      "tokens_seen": 2285961216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125858585858586,
      "loss": 2.4852,
      "theoretical_loss": 3.3928655772049043,
      "tokens_seen": 2286026752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031256565656565655,
      "loss": 2.459,
      "theoretical_loss": 3.3928576789595275,
      "tokens_seen": 2286092288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031254545454545456,
      "loss": 2.3874,
      "theoretical_loss": 3.3928497810039646,
      "tokens_seen": 2286157824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125252525252525,
      "loss": 2.742,
      "theoretical_loss": 3.3928418833381957,
      "tokens_seen": 2286223360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031250505050505053,
      "loss": 2.5145,
      "theoretical_loss": 3.392833985962203,
      "tokens_seen": 2286288896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124848484848485,
      "loss": 2.6008,
      "theoretical_loss": 3.392826088875967,
      "tokens_seen": 2286354432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031246464646464644,
      "loss": 2.258,
      "theoretical_loss": 3.392818192079469,
      "tokens_seen": 2286419968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031244444444444446,
      "loss": 2.5079,
      "theoretical_loss": 3.3928102955726898,
      "tokens_seen": 2286485504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031242424242424247,
      "loss": 2.5061,
      "theoretical_loss": 3.3928023993556105,
      "tokens_seen": 2286551040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124040404040404,
      "loss": 2.8721,
      "theoretical_loss": 3.3927945034282123,
      "tokens_seen": 2286616576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123838383838384,
      "loss": 2.6242,
      "theoretical_loss": 3.3927866077904767,
      "tokens_seen": 2286682112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031236363636363634,
      "loss": 2.5288,
      "theoretical_loss": 3.3927787124423836,
      "tokens_seen": 2286747648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031234343434343435,
      "loss": 2.5572,
      "theoretical_loss": 3.3927708173839153,
      "tokens_seen": 2286813184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031232323232323236,
      "loss": 2.6219,
      "theoretical_loss": 3.392762922615052,
      "tokens_seen": 2286878720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123030303030303,
      "loss": 2.6096,
      "theoretical_loss": 3.3927550281357757,
      "tokens_seen": 2286944256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122828282828283,
      "loss": 2.4923,
      "theoretical_loss": 3.3927471339460666,
      "tokens_seen": 2287009792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031226262626262623,
      "loss": 2.6047,
      "theoretical_loss": 3.3927392400459055,
      "tokens_seen": 2287075328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122424242424243,
      "loss": 2.5471,
      "theoretical_loss": 3.392731346435275,
      "tokens_seen": 2287140864
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031222222222222225,
      "loss": 2.7516,
      "theoretical_loss": 3.392723453114155,
      "tokens_seen": 2287206400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122020202020202,
      "loss": 2.5674,
      "theoretical_loss": 3.3927155600825265,
      "tokens_seen": 2287271936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031218181818181817,
      "loss": 2.5033,
      "theoretical_loss": 3.3927076673403715,
      "tokens_seen": 2287337472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121616161616161,
      "loss": 2.6554,
      "theoretical_loss": 3.39269977488767,
      "tokens_seen": 2287403008
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1262743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.618569850921631,
      "objective/train/theoretical_loss": 3.392691882724404,
      "objective/train/tokens_used": 646327776,
      "theoretical_loss": 3.392691882724404,
      "tokens_seen": 2287468544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121414141414142,
      "loss": 2.7417,
      "theoretical_loss": 3.392691882724404,
      "tokens_seen": 2287468544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031212121212121214,
      "loss": 2.7722,
      "theoretical_loss": 3.392683990850554,
      "tokens_seen": 2287534080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121010101010101,
      "loss": 2.5364,
      "theoretical_loss": 3.3926760992661014,
      "tokens_seen": 2287599616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031208080808080806,
      "loss": 2.5532,
      "theoretical_loss": 3.3926682079710275,
      "tokens_seen": 2287665152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031206060606060607,
      "loss": 2.5556,
      "theoretical_loss": 3.3926603169653125,
      "tokens_seen": 2287730688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003120404040404041,
      "loss": 2.549,
      "theoretical_loss": 3.3926524262489384,
      "tokens_seen": 2287796224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031202020202020204,
      "loss": 2.4129,
      "theoretical_loss": 3.392644535821886,
      "tokens_seen": 2287861760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000312,
      "loss": 2.4981,
      "theoretical_loss": 3.392636645684137,
      "tokens_seen": 2287927296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031197979797979795,
      "loss": 2.5447,
      "theoretical_loss": 3.392628755835671,
      "tokens_seen": 2287992832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031195959595959596,
      "loss": 2.584,
      "theoretical_loss": 3.3926208662764705,
      "tokens_seen": 2288058368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311939393939394,
      "loss": 2.6751,
      "theoretical_loss": 3.392612977006516,
      "tokens_seen": 2288123904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031191919191919193,
      "loss": 2.6755,
      "theoretical_loss": 3.392605088025789,
      "tokens_seen": 2288189440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118989898989899,
      "loss": 2.5299,
      "theoretical_loss": 3.3925971993342703,
      "tokens_seen": 2288254976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118787878787879,
      "loss": 2.415,
      "theoretical_loss": 3.3925893109319407,
      "tokens_seen": 2288320512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031185858585858585,
      "loss": 2.5081,
      "theoretical_loss": 3.392581422818782,
      "tokens_seen": 2288386048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031183838383838387,
      "loss": 2.6131,
      "theoretical_loss": 3.392573534994775,
      "tokens_seen": 2288451584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118181818181818,
      "loss": 2.6848,
      "theoretical_loss": 3.392565647459901,
      "tokens_seen": 2288517120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117979797979798,
      "loss": 2.5136,
      "theoretical_loss": 3.3925577602141406,
      "tokens_seen": 2288582656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117777777777778,
      "loss": 2.4394,
      "theoretical_loss": 3.392549873257476,
      "tokens_seen": 2288648192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031175757575757575,
      "loss": 2.5428,
      "theoretical_loss": 3.3925419865898867,
      "tokens_seen": 2288713728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031173737373737376,
      "loss": 2.6776,
      "theoretical_loss": 3.392534100211355,
      "tokens_seen": 2288779264
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117171717171717,
      "loss": 2.4304,
      "theoretical_loss": 3.392526214121862,
      "tokens_seen": 2288844800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116969696969697,
      "loss": 2.4941,
      "theoretical_loss": 3.3925183283213887,
      "tokens_seen": 2288910336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116767676767677,
      "loss": 2.6966,
      "theoretical_loss": 3.392510442809916,
      "tokens_seen": 2288975872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031165656565656564,
      "loss": 2.648,
      "theoretical_loss": 3.3925025575874255,
      "tokens_seen": 2289041408
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1263359,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.576524019241333,
      "objective/train/theoretical_loss": 3.392494672653897,
      "objective/train/tokens_used": 647966176,
      "theoretical_loss": 3.392494672653897,
      "tokens_seen": 2289106944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031163636363636365,
      "loss": 2.7716,
      "theoretical_loss": 3.392494672653897,
      "tokens_seen": 2289106944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116161616161616,
      "loss": 2.6873,
      "theoretical_loss": 3.3924867880093137,
      "tokens_seen": 2289172480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003115959595959596,
      "loss": 2.5042,
      "theoretical_loss": 3.392478903653655,
      "tokens_seen": 2289238016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003115757575757576,
      "loss": 2.4627,
      "theoretical_loss": 3.3924710195869032,
      "tokens_seen": 2289303552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031155555555555553,
      "loss": 2.3332,
      "theoretical_loss": 3.392463135809039,
      "tokens_seen": 2289369088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031153535353535354,
      "loss": 2.6457,
      "theoretical_loss": 3.392455252320043,
      "tokens_seen": 2289434624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031151515151515156,
      "loss": 2.3721,
      "theoretical_loss": 3.392447369119897,
      "tokens_seen": 2289500160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114949494949495,
      "loss": 2.8084,
      "theoretical_loss": 3.3924394862085823,
      "tokens_seen": 2289565696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031147474747474747,
      "loss": 2.4445,
      "theoretical_loss": 3.3924316035860795,
      "tokens_seen": 2289631232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114545454545454,
      "loss": 2.4715,
      "theoretical_loss": 3.3924237212523702,
      "tokens_seen": 2289696768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031143434343434344,
      "loss": 2.6434,
      "theoretical_loss": 3.3924158392074353,
      "tokens_seen": 2289762304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031141414141414145,
      "loss": 2.5525,
      "theoretical_loss": 3.392407957451256,
      "tokens_seen": 2289827840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113939393939394,
      "loss": 2.5955,
      "theoretical_loss": 3.3924000759838133,
      "tokens_seen": 2289893376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031137373737373736,
      "loss": 2.6712,
      "theoretical_loss": 3.392392194805089,
      "tokens_seen": 2289958912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113535353535353,
      "loss": 2.5935,
      "theoretical_loss": 3.3923843139150636,
      "tokens_seen": 2290024448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113333333333334,
      "loss": 2.5464,
      "theoretical_loss": 3.3923764333137183,
      "tokens_seen": 2290089984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031131313131313134,
      "loss": 2.4847,
      "theoretical_loss": 3.3923685530010346,
      "tokens_seen": 2290155520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112929292929293,
      "loss": 2.7156,
      "theoretical_loss": 3.3923606729769933,
      "tokens_seen": 2290221056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031127272727272725,
      "loss": 2.509,
      "theoretical_loss": 3.3923527932415762,
      "tokens_seen": 2290286592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112525252525252,
      "loss": 2.9099,
      "theoretical_loss": 3.3923449137947634,
      "tokens_seen": 2290352128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112323232323233,
      "loss": 2.4327,
      "theoretical_loss": 3.392337034636537,
      "tokens_seen": 2290417664
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031121212121212123,
      "loss": 2.863,
      "theoretical_loss": 3.3923291557668778,
      "tokens_seen": 2290483200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111919191919192,
      "loss": 2.5169,
      "theoretical_loss": 3.3923212771857676,
      "tokens_seen": 2290548736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031117171717171715,
      "loss": 2.6663,
      "theoretical_loss": 3.3923133988931866,
      "tokens_seen": 2290614272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111515151515152,
      "loss": 2.4132,
      "theoretical_loss": 3.392305520889116,
      "tokens_seen": 2290679808
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1264250,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7503297328948975,
      "objective/train/theoretical_loss": 3.392297643173538,
      "objective/train/tokens_used": 649604576,
      "theoretical_loss": 3.392297643173538,
      "tokens_seen": 2290745344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031113131313131317,
      "loss": 2.5558,
      "theoretical_loss": 3.392297643173538,
      "tokens_seen": 2290745344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111111111111111,
      "loss": 2.8482,
      "theoretical_loss": 3.3922897657464333,
      "tokens_seen": 2290810880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003110909090909091,
      "loss": 2.6156,
      "theoretical_loss": 3.3922818886077826,
      "tokens_seen": 2290876416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031107070707070704,
      "loss": 2.6552,
      "theoretical_loss": 3.392274011757567,
      "tokens_seen": 2290941952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003110505050505051,
      "loss": 2.5031,
      "theoretical_loss": 3.3922661351957686,
      "tokens_seen": 2291007488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031103030303030306,
      "loss": 2.6322,
      "theoretical_loss": 3.3922582589223684,
      "tokens_seen": 2291073024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311010101010101,
      "loss": 2.5946,
      "theoretical_loss": 3.392250382937347,
      "tokens_seen": 2291138560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000310989898989899,
      "loss": 2.6454,
      "theoretical_loss": 3.392242507240686,
      "tokens_seen": 2291204096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031096969696969693,
      "loss": 2.4293,
      "theoretical_loss": 3.392234631832366,
      "tokens_seen": 2291269632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000310949494949495,
      "loss": 2.4459,
      "theoretical_loss": 3.3922267567123696,
      "tokens_seen": 2291335168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031092929292929295,
      "loss": 2.6167,
      "theoretical_loss": 3.3922188818806767,
      "tokens_seen": 2291400704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109090909090909,
      "loss": 2.5918,
      "theoretical_loss": 3.3922110073372687,
      "tokens_seen": 2291466240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031088888888888887,
      "loss": 2.6297,
      "theoretical_loss": 3.392203133082127,
      "tokens_seen": 2291531776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108686868686869,
      "loss": 2.7358,
      "theoretical_loss": 3.392195259115233,
      "tokens_seen": 2291597312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108484848484849,
      "loss": 2.5908,
      "theoretical_loss": 3.392187385436568,
      "tokens_seen": 2291662848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031082828282828285,
      "loss": 2.8274,
      "theoretical_loss": 3.3921795120461122,
      "tokens_seen": 2291728384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108080808080808,
      "loss": 2.6506,
      "theoretical_loss": 3.392171638943848,
      "tokens_seen": 2291793920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031078787878787876,
      "loss": 2.5598,
      "theoretical_loss": 3.392163766129756,
      "tokens_seen": 2291859456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031076767676767677,
      "loss": 2.8199,
      "theoretical_loss": 3.3921558936038174,
      "tokens_seen": 2291924992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107474747474748,
      "loss": 2.602,
      "theoretical_loss": 3.3921480213660136,
      "tokens_seen": 2291990528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031072727272727274,
      "loss": 2.6791,
      "theoretical_loss": 3.392140149416326,
      "tokens_seen": 2292056064
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107070707070707,
      "loss": 2.6833,
      "theoretical_loss": 3.3921322777547354,
      "tokens_seen": 2292121600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003106868686868687,
      "loss": 2.4246,
      "theoretical_loss": 3.3921244063812233,
      "tokens_seen": 2292187136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031066666666666666,
      "loss": 2.6528,
      "theoretical_loss": 3.392116535295771,
      "tokens_seen": 2292252672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003106464646464647,
      "loss": 2.6907,
      "theoretical_loss": 3.3921086644983593,
      "tokens_seen": 2292318208
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 1264872,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.41400146484375,
      "objective/train/theoretical_loss": 3.39210079398897,
      "objective/train/tokens_used": 651242976,
      "theoretical_loss": 3.39210079398897,
      "tokens_seen": 2292383744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031062626262626263,
      "loss": 2.4975,
      "theoretical_loss": 3.39210079398897,
      "tokens_seen": 2292383744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003106060606060606,
      "loss": 2.6498,
      "theoretical_loss": 3.3920929237675836,
      "tokens_seen": 2292449280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105858585858586,
      "loss": 2.5796,
      "theoretical_loss": 3.392085053834182,
      "tokens_seen": 2292514816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031056565656565656,
      "loss": 2.8169,
      "theoretical_loss": 3.392077184188746,
      "tokens_seen": 2292580352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031054545454545457,
      "loss": 2.7175,
      "theoretical_loss": 3.3920693148312573,
      "tokens_seen": 2292645888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105252525252525,
      "loss": 2.669,
      "theoretical_loss": 3.392061445761696,
      "tokens_seen": 2292711424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031050505050505054,
      "loss": 2.5452,
      "theoretical_loss": 3.392053576980045,
      "tokens_seen": 2292776960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104848484848485,
      "loss": 2.669,
      "theoretical_loss": 3.3920457084862843,
      "tokens_seen": 2292842496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031046464646464645,
      "loss": 2.4938,
      "theoretical_loss": 3.3920378402803957,
      "tokens_seen": 2292908032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031044444444444446,
      "loss": 2.647,
      "theoretical_loss": 3.3920299723623604,
      "tokens_seen": 2292973568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104242424242424,
      "loss": 2.6427,
      "theoretical_loss": 3.3920221047321597,
      "tokens_seen": 2293039104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031040404040404043,
      "loss": 2.5542,
      "theoretical_loss": 3.392014237389774,
      "tokens_seen": 2293104640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103838383838384,
      "loss": 2.6219,
      "theoretical_loss": 3.392006370335186,
      "tokens_seen": 2293170176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031036363636363634,
      "loss": 2.8148,
      "theoretical_loss": 3.3919985035683755,
      "tokens_seen": 2293235712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031034343434343435,
      "loss": 2.8275,
      "theoretical_loss": 3.3919906370893247,
      "tokens_seen": 2293301248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031032323232323236,
      "loss": 2.5837,
      "theoretical_loss": 3.391982770898015,
      "tokens_seen": 2293366784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103030303030303,
      "loss": 2.6612,
      "theoretical_loss": 3.3919749049944263,
      "tokens_seen": 2293432320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003102828282828283,
      "loss": 2.5767,
      "theoretical_loss": 3.3919670393785415,
      "tokens_seen": 2293497856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031026262626262624,
      "loss": 2.5719,
      "theoretical_loss": 3.3919591740503408,
      "tokens_seen": 2293563392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031024242424242425,
      "loss": 2.388,
      "theoretical_loss": 3.391951309009806,
      "tokens_seen": 2293628928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031022222222222226,
      "loss": 2.6829,
      "theoretical_loss": 3.391943444256918,
      "tokens_seen": 2293694464
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003102020202020202,
      "loss": 2.6251,
      "theoretical_loss": 3.391935579791658,
      "tokens_seen": 2293760000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031018181818181817,
      "loss": 2.5934,
      "theoretical_loss": 3.3919277156140075,
      "tokens_seen": 2293825536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031016161616161613,
      "loss": 2.4721,
      "theoretical_loss": 3.391919851723948,
      "tokens_seen": 2293891072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101414141414142,
      "loss": 2.6051,
      "theoretical_loss": 3.3919119881214606,
      "tokens_seen": 2293956608
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1266089,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6598479747772217,
      "objective/train/theoretical_loss": 3.3919041248065263,
      "objective/train/tokens_used": 652881376,
      "theoretical_loss": 3.3919041248065263,
      "tokens_seen": 2294022144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031012121212121215,
      "loss": 2.5609,
      "theoretical_loss": 3.3919041248065263,
      "tokens_seen": 2294022144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101010101010101,
      "loss": 2.6973,
      "theoretical_loss": 3.3918962617791264,
      "tokens_seen": 2294087680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031008080808080806,
      "loss": 2.5322,
      "theoretical_loss": 3.3918883990392423,
      "tokens_seen": 2294153216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000310060606060606,
      "loss": 2.6408,
      "theoretical_loss": 3.3918805365868554,
      "tokens_seen": 2294218752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100404040404041,
      "loss": 2.5532,
      "theoretical_loss": 3.391872674421947,
      "tokens_seen": 2294284288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031002020202020204,
      "loss": 2.6585,
      "theoretical_loss": 3.391864812544498,
      "tokens_seen": 2294349824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031,
      "loss": 2.4281,
      "theoretical_loss": 3.39185695095449,
      "tokens_seen": 2294415360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030997979797979796,
      "loss": 2.7691,
      "theoretical_loss": 3.3918490896519042,
      "tokens_seen": 2294480896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030995959595959597,
      "loss": 2.6763,
      "theoretical_loss": 3.391841228636722,
      "tokens_seen": 2294546432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000309939393939394,
      "loss": 2.5084,
      "theoretical_loss": 3.3918333679089243,
      "tokens_seen": 2294611968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030991919191919194,
      "loss": 2.6088,
      "theoretical_loss": 3.391825507468493,
      "tokens_seen": 2294677504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098989898989899,
      "loss": 2.5089,
      "theoretical_loss": 3.391817647315409,
      "tokens_seen": 2294743040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030987878787878785,
      "loss": 2.5429,
      "theoretical_loss": 3.3918097874496533,
      "tokens_seen": 2294808576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030985858585858586,
      "loss": 2.5231,
      "theoretical_loss": 3.391801927871208,
      "tokens_seen": 2294874112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030983838383838387,
      "loss": 2.6438,
      "theoretical_loss": 3.3917940685800536,
      "tokens_seen": 2294939648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030981818181818183,
      "loss": 2.575,
      "theoretical_loss": 3.3917862095761717,
      "tokens_seen": 2295005184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097979797979798,
      "loss": 2.5795,
      "theoretical_loss": 3.3917783508595436,
      "tokens_seen": 2295070720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097777777777778,
      "loss": 2.6748,
      "theoretical_loss": 3.3917704924301506,
      "tokens_seen": 2295136256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030975757575757575,
      "loss": 2.5086,
      "theoretical_loss": 3.391762634287974,
      "tokens_seen": 2295201792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030973737373737376,
      "loss": 2.7201,
      "theoretical_loss": 3.3917547764329954,
      "tokens_seen": 2295267328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097171717171717,
      "loss": 2.7846,
      "theoretical_loss": 3.391746918865196,
      "tokens_seen": 2295332864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096969696969697,
      "loss": 2.6534,
      "theoretical_loss": 3.3917390615845564,
      "tokens_seen": 2295398400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096767676767677,
      "loss": 2.6957,
      "theoretical_loss": 3.3917312045910584,
      "tokens_seen": 2295463936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030965656565656565,
      "loss": 2.3621,
      "theoretical_loss": 3.3917233478846835,
      "tokens_seen": 2295529472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030963636363636366,
      "loss": 2.497,
      "theoretical_loss": 3.391715491465413,
      "tokens_seen": 2295595008
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1266610,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5779285430908203,
      "objective/train/theoretical_loss": 3.3917076353332276,
      "objective/train/tokens_used": 654519776,
      "theoretical_loss": 3.3917076353332276,
      "tokens_seen": 2295660544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096161616161616,
      "loss": 2.6769,
      "theoretical_loss": 3.3917076353332276,
      "tokens_seen": 2295660544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030959595959595957,
      "loss": 2.5646,
      "theoretical_loss": 3.3916997794881096,
      "tokens_seen": 2295726080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095757575757576,
      "loss": 2.5626,
      "theoretical_loss": 3.39169192393004,
      "tokens_seen": 2295791616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030955555555555554,
      "loss": 2.5432,
      "theoretical_loss": 3.391684068658999,
      "tokens_seen": 2295857152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030953535353535355,
      "loss": 2.6588,
      "theoretical_loss": 3.391676213674969,
      "tokens_seen": 2295922688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095151515151515,
      "loss": 2.6954,
      "theoretical_loss": 3.391668358977932,
      "tokens_seen": 2295988224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094949494949495,
      "loss": 2.705,
      "theoretical_loss": 3.391660504567868,
      "tokens_seen": 2296053760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094747474747475,
      "loss": 2.5259,
      "theoretical_loss": 3.391652650444758,
      "tokens_seen": 2296119296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094545454545455,
      "loss": 2.713,
      "theoretical_loss": 3.391644796608585,
      "tokens_seen": 2296184832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030943434343434344,
      "loss": 2.759,
      "theoretical_loss": 3.3916369430593294,
      "tokens_seen": 2296250368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094141414141414,
      "loss": 2.4451,
      "theoretical_loss": 3.3916290897969725,
      "tokens_seen": 2296315904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003093939393939394,
      "loss": 2.7109,
      "theoretical_loss": 3.3916212368214955,
      "tokens_seen": 2296381440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030937373737373737,
      "loss": 2.7819,
      "theoretical_loss": 3.39161338413288,
      "tokens_seen": 2296446976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003093535353535354,
      "loss": 2.7268,
      "theoretical_loss": 3.391605531731108,
      "tokens_seen": 2296512512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030933333333333334,
      "loss": 2.6137,
      "theoretical_loss": 3.3915976796161593,
      "tokens_seen": 2296578048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030931313131313135,
      "loss": 2.8267,
      "theoretical_loss": 3.391589827788016,
      "tokens_seen": 2296643584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003092929292929293,
      "loss": 2.8173,
      "theoretical_loss": 3.39158197624666,
      "tokens_seen": 2296709120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030927272727272726,
      "loss": 2.6758,
      "theoretical_loss": 3.3915741249920717,
      "tokens_seen": 2296774656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030925252525252527,
      "loss": 2.4154,
      "theoretical_loss": 3.391566274024233,
      "tokens_seen": 2296840192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030923232323232323,
      "loss": 2.6211,
      "theoretical_loss": 3.3915584233431253,
      "tokens_seen": 2296905728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030921212121212124,
      "loss": 2.5942,
      "theoretical_loss": 3.39155057294873,
      "tokens_seen": 2296971264
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091919191919192,
      "loss": 2.6141,
      "theoretical_loss": 3.3915427228410278,
      "tokens_seen": 2297036800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030917171717171715,
      "loss": 2.5705,
      "theoretical_loss": 3.3915348730200003,
      "tokens_seen": 2297102336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030915151515151516,
      "loss": 2.8109,
      "theoretical_loss": 3.3915270234856294,
      "tokens_seen": 2297167872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091313131313132,
      "loss": 2.5373,
      "theoretical_loss": 3.3915191742378963,
      "tokens_seen": 2297233408
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1267818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8280587196350098,
      "objective/train/theoretical_loss": 3.391511325276782,
      "objective/train/tokens_used": 656158176,
      "theoretical_loss": 3.391511325276782,
      "tokens_seen": 2297298944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030911111111111113,
      "loss": 2.7048,
      "theoretical_loss": 3.391511325276782,
      "tokens_seen": 2297298944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090909090909091,
      "loss": 2.4976,
      "theoretical_loss": 3.3915034766022676,
      "tokens_seen": 2297364480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030907070707070705,
      "loss": 2.6139,
      "theoretical_loss": 3.3914956282143347,
      "tokens_seen": 2297430016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030905050505050506,
      "loss": 2.4821,
      "theoretical_loss": 3.3914877801129655,
      "tokens_seen": 2297495552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030903030303030307,
      "loss": 2.5747,
      "theoretical_loss": 3.3914799322981404,
      "tokens_seen": 2297561088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000309010101010101,
      "loss": 2.6617,
      "theoretical_loss": 3.391472084769841,
      "tokens_seen": 2297626624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308989898989899,
      "loss": 2.589,
      "theoretical_loss": 3.391464237528049,
      "tokens_seen": 2297692160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030896969696969694,
      "loss": 2.6321,
      "theoretical_loss": 3.391456390572745,
      "tokens_seen": 2297757696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308949494949495,
      "loss": 2.6723,
      "theoretical_loss": 3.391448543903911,
      "tokens_seen": 2297823232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030892929292929296,
      "loss": 2.6143,
      "theoretical_loss": 3.3914406975215283,
      "tokens_seen": 2297888768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003089090909090909,
      "loss": 2.6186,
      "theoretical_loss": 3.391432851425578,
      "tokens_seen": 2297954304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088888888888889,
      "loss": 2.6953,
      "theoretical_loss": 3.391425005616042,
      "tokens_seen": 2298019840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030886868686868683,
      "loss": 2.4945,
      "theoretical_loss": 3.3914171600929013,
      "tokens_seen": 2298085376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088484848484849,
      "loss": 2.6718,
      "theoretical_loss": 3.3914093148561375,
      "tokens_seen": 2298150912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030882828282828285,
      "loss": 2.7374,
      "theoretical_loss": 3.3914014699057313,
      "tokens_seen": 2298216448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088080808080808,
      "loss": 2.5461,
      "theoretical_loss": 3.391393625241665,
      "tokens_seen": 2298281984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030878787878787877,
      "loss": 2.5156,
      "theoretical_loss": 3.3913857808639194,
      "tokens_seen": 2298347520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087676767676768,
      "loss": 2.8457,
      "theoretical_loss": 3.391377936772476,
      "tokens_seen": 2298413056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087474747474748,
      "loss": 2.4838,
      "theoretical_loss": 3.3913700929673163,
      "tokens_seen": 2298478592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030872727272727275,
      "loss": 2.7192,
      "theoretical_loss": 3.391362249448422,
      "tokens_seen": 2298544128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087070707070707,
      "loss": 2.7096,
      "theoretical_loss": 3.3913544062157737,
      "tokens_seen": 2298609664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030868686868686866,
      "loss": 2.6366,
      "theoretical_loss": 3.3913465632693534,
      "tokens_seen": 2298675200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030866666666666667,
      "loss": 2.4499,
      "theoretical_loss": 3.391338720609142,
      "tokens_seen": 2298740736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086464646464647,
      "loss": 2.5605,
      "theoretical_loss": 3.3913308782351215,
      "tokens_seen": 2298806272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030862626262626264,
      "loss": 2.8173,
      "theoretical_loss": 3.391323036147273,
      "tokens_seen": 2298871808
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1268401,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1293718814849854,
      "objective/train/theoretical_loss": 3.391315194345578,
      "objective/train/tokens_used": 657796576,
      "theoretical_loss": 3.391315194345578,
      "tokens_seen": 2298937344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086060606060606,
      "loss": 2.3771,
      "theoretical_loss": 3.391315194345578,
      "tokens_seen": 2298937344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085858585858586,
      "loss": 2.8841,
      "theoretical_loss": 3.3913073528300175,
      "tokens_seen": 2299002880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030856565656565656,
      "loss": 2.6842,
      "theoretical_loss": 3.3912995116005735,
      "tokens_seen": 2299068416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085454545454546,
      "loss": 2.5984,
      "theoretical_loss": 3.391291670657227,
      "tokens_seen": 2299133952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030852525252525253,
      "loss": 2.5788,
      "theoretical_loss": 3.3912838299999595,
      "tokens_seen": 2299199488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085050505050505,
      "loss": 2.4784,
      "theoretical_loss": 3.3912759896287525,
      "tokens_seen": 2299265024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084848484848485,
      "loss": 2.7181,
      "theoretical_loss": 3.391268149543587,
      "tokens_seen": 2299330560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030846464646464646,
      "loss": 2.5747,
      "theoretical_loss": 3.391260309744445,
      "tokens_seen": 2299396096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030844444444444447,
      "loss": 2.4784,
      "theoretical_loss": 3.391252470231308,
      "tokens_seen": 2299461632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084242424242424,
      "loss": 2.4493,
      "theoretical_loss": 3.391244631004157,
      "tokens_seen": 2299527168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030840404040404043,
      "loss": 2.5682,
      "theoretical_loss": 3.391236792062973,
      "tokens_seen": 2299592704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003083838383838384,
      "loss": 2.3225,
      "theoretical_loss": 3.3912289534077384,
      "tokens_seen": 2299658240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030836363636363635,
      "loss": 2.6492,
      "theoretical_loss": 3.391221115038434,
      "tokens_seen": 2299723776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030834343434343436,
      "loss": 2.4695,
      "theoretical_loss": 3.3912132769550416,
      "tokens_seen": 2299789312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003083232323232323,
      "loss": 2.2739,
      "theoretical_loss": 3.3912054391575417,
      "tokens_seen": 2299854848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030830303030303033,
      "loss": 2.5069,
      "theoretical_loss": 3.391197601645917,
      "tokens_seen": 2299920384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082828282828283,
      "loss": 2.435,
      "theoretical_loss": 3.3911897644201483,
      "tokens_seen": 2299985920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030826262626262624,
      "loss": 2.6603,
      "theoretical_loss": 3.391181927480217,
      "tokens_seen": 2300051456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030824242424242425,
      "loss": 2.6379,
      "theoretical_loss": 3.3911740908261043,
      "tokens_seen": 2300116992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030822222222222226,
      "loss": 2.4672,
      "theoretical_loss": 3.3911662544577927,
      "tokens_seen": 2300182528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082020202020202,
      "loss": 2.5501,
      "theoretical_loss": 3.3911584183752623,
      "tokens_seen": 2300248064
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081818181818182,
      "loss": 2.4988,
      "theoretical_loss": 3.391150582578495,
      "tokens_seen": 2300313600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030816161616161613,
      "loss": 2.7939,
      "theoretical_loss": 3.3911427470674727,
      "tokens_seen": 2300379136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030814141414141414,
      "loss": 2.6005,
      "theoretical_loss": 3.391134911842176,
      "tokens_seen": 2300444672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030812121212121216,
      "loss": 2.5733,
      "theoretical_loss": 3.3911270769025874,
      "tokens_seen": 2300510208
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1269713,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.40155291557312,
      "objective/train/theoretical_loss": 3.3911192422486875,
      "objective/train/tokens_used": 659434976,
      "theoretical_loss": 3.3911192422486875,
      "tokens_seen": 2300575744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081010101010101,
      "loss": 2.5398,
      "theoretical_loss": 3.3911192422486875,
      "tokens_seen": 2300575744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030808080808080807,
      "loss": 2.5321,
      "theoretical_loss": 3.391111407880458,
      "tokens_seen": 2300641280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308060606060606,
      "loss": 2.77,
      "theoretical_loss": 3.3911035737978805,
      "tokens_seen": 2300706816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003080404040404041,
      "loss": 2.622,
      "theoretical_loss": 3.391095740000936,
      "tokens_seen": 2300772352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030802020202020205,
      "loss": 2.5246,
      "theoretical_loss": 3.3910879064896067,
      "tokens_seen": 2300837888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308,
      "loss": 2.6132,
      "theoretical_loss": 3.3910800732638733,
      "tokens_seen": 2300903424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030797979797979796,
      "loss": 2.5629,
      "theoretical_loss": 3.3910722403237177,
      "tokens_seen": 2300968960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079595959595959,
      "loss": 2.7191,
      "theoretical_loss": 3.3910644076691208,
      "tokens_seen": 2301034496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000307939393939394,
      "loss": 2.535,
      "theoretical_loss": 3.391056575300065,
      "tokens_seen": 2301100032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030791919191919194,
      "loss": 2.4897,
      "theoretical_loss": 3.391048743216531,
      "tokens_seen": 2301165568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078989898989899,
      "loss": 2.72,
      "theoretical_loss": 3.3910409114185005,
      "tokens_seen": 2301231104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030787878787878786,
      "loss": 2.4124,
      "theoretical_loss": 3.391033079905955,
      "tokens_seen": 2301296640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078585858585858,
      "loss": 2.4873,
      "theoretical_loss": 3.3910252486788757,
      "tokens_seen": 2301362176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078383838383839,
      "loss": 2.3633,
      "theoretical_loss": 3.3910174177372445,
      "tokens_seen": 2301427712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030781818181818183,
      "loss": 2.597,
      "theoretical_loss": 3.391009587081043,
      "tokens_seen": 2301493248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077979797979798,
      "loss": 2.5238,
      "theoretical_loss": 3.391001756710252,
      "tokens_seen": 2301558784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030777777777777775,
      "loss": 2.4102,
      "theoretical_loss": 3.390993926624853,
      "tokens_seen": 2301624320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077575757575758,
      "loss": 2.6404,
      "theoretical_loss": 3.390986096824828,
      "tokens_seen": 2301689856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030773737373737377,
      "loss": 2.5578,
      "theoretical_loss": 3.3909782673101585,
      "tokens_seen": 2301755392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077171717171717,
      "loss": 2.6326,
      "theoretical_loss": 3.390970438080825,
      "tokens_seen": 2301820928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076969696969697,
      "loss": 2.451,
      "theoretical_loss": 3.390962609136811,
      "tokens_seen": 2301886464
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030767676767676764,
      "loss": 2.4678,
      "theoretical_loss": 3.3909547804780953,
      "tokens_seen": 2301952000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076565656565657,
      "loss": 2.2928,
      "theoretical_loss": 3.3909469521046613,
      "tokens_seen": 2302017536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030763636363636366,
      "loss": 2.4691,
      "theoretical_loss": 3.39093912401649,
      "tokens_seen": 2302083072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076161616161616,
      "loss": 2.3614,
      "theoretical_loss": 3.390931296213563,
      "tokens_seen": 2302148608
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1270398,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.385732412338257,
      "objective/train/theoretical_loss": 3.3909234686958616,
      "objective/train/tokens_used": 661073376,
      "theoretical_loss": 3.3909234686958616,
      "tokens_seen": 2302214144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075959595959596,
      "loss": 2.6673,
      "theoretical_loss": 3.3909234686958616,
      "tokens_seen": 2302214144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075757575757576,
      "loss": 2.7333,
      "theoretical_loss": 3.3909156414633665,
      "tokens_seen": 2302279680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075555555555556,
      "loss": 2.4734,
      "theoretical_loss": 3.390907814516061,
      "tokens_seen": 2302345216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030753535353535356,
      "loss": 2.6199,
      "theoretical_loss": 3.390899987853925,
      "tokens_seen": 2302410752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075151515151515,
      "loss": 2.6377,
      "theoretical_loss": 3.3908921614769407,
      "tokens_seen": 2302476288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030749494949494947,
      "loss": 2.499,
      "theoretical_loss": 3.3908843353850893,
      "tokens_seen": 2302541824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074747474747475,
      "loss": 2.448,
      "theoretical_loss": 3.390876509578353,
      "tokens_seen": 2302607360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074545454545455,
      "loss": 2.4048,
      "theoretical_loss": 3.3908686840567124,
      "tokens_seen": 2302672896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030743434343434345,
      "loss": 2.537,
      "theoretical_loss": 3.390860858820149,
      "tokens_seen": 2302738432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074141414141414,
      "loss": 2.5484,
      "theoretical_loss": 3.3908530338686456,
      "tokens_seen": 2302803968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073939393939394,
      "loss": 2.4612,
      "theoretical_loss": 3.390845209202182,
      "tokens_seen": 2302869504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030737373737373737,
      "loss": 2.5923,
      "theoretical_loss": 3.390837384820741,
      "tokens_seen": 2302935040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073535353535354,
      "loss": 2.436,
      "theoretical_loss": 3.390829560724303,
      "tokens_seen": 2303000576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030733333333333334,
      "loss": 2.648,
      "theoretical_loss": 3.3908217369128506,
      "tokens_seen": 2303066112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073131313131313,
      "loss": 2.5518,
      "theoretical_loss": 3.3908139133863644,
      "tokens_seen": 2303131648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072929292929293,
      "loss": 2.6575,
      "theoretical_loss": 3.3908060901448263,
      "tokens_seen": 2303197184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030727272727272727,
      "loss": 2.4643,
      "theoretical_loss": 3.390798267188218,
      "tokens_seen": 2303262720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072525252525253,
      "loss": 2.5023,
      "theoretical_loss": 3.390790444516521,
      "tokens_seen": 2303328256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030723232323232323,
      "loss": 2.5001,
      "theoretical_loss": 3.3907826221297164,
      "tokens_seen": 2303393792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030721212121212124,
      "loss": 2.4754,
      "theoretical_loss": 3.390774800027786,
      "tokens_seen": 2303459328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071919191919192,
      "loss": 2.616,
      "theoretical_loss": 3.3907669782107113,
      "tokens_seen": 2303524864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030717171717171716,
      "loss": 2.377,
      "theoretical_loss": 3.390759156678474,
      "tokens_seen": 2303590400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030715151515151517,
      "loss": 2.2582,
      "theoretical_loss": 3.3907513354310557,
      "tokens_seen": 2303655936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071313131313131,
      "loss": 2.6651,
      "theoretical_loss": 3.3907435144684372,
      "tokens_seen": 2303721472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030711111111111114,
      "loss": 2.512,
      "theoretical_loss": 3.3907356937906004,
      "tokens_seen": 2303787008
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1271635,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6408653259277344,
      "objective/train/theoretical_loss": 3.3907278733975272,
      "objective/train/tokens_used": 662711776,
      "theoretical_loss": 3.3907278733975272,
      "tokens_seen": 2303852544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070909090909091,
      "loss": 2.3809,
      "theoretical_loss": 3.3907278733975272,
      "tokens_seen": 2303852544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030707070707070705,
      "loss": 2.4918,
      "theoretical_loss": 3.390720053289199,
      "tokens_seen": 2303918080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030705050505050506,
      "loss": 2.3329,
      "theoretical_loss": 3.390712233465597,
      "tokens_seen": 2303983616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070303030303031,
      "loss": 2.6718,
      "theoretical_loss": 3.390704413926703,
      "tokens_seen": 2304049152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030701010101010103,
      "loss": 2.5126,
      "theoretical_loss": 3.390696594672498,
      "tokens_seen": 2304114688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306989898989899,
      "loss": 2.2646,
      "theoretical_loss": 3.390688775702965,
      "tokens_seen": 2304180224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030696969696969694,
      "loss": 2.5245,
      "theoretical_loss": 3.3906809570180836,
      "tokens_seen": 2304245760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030694949494949495,
      "loss": 2.85,
      "theoretical_loss": 3.390673138617837,
      "tokens_seen": 2304311296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030692929292929297,
      "loss": 2.5639,
      "theoretical_loss": 3.3906653205022055,
      "tokens_seen": 2304376832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069090909090909,
      "loss": 2.4869,
      "theoretical_loss": 3.3906575026711714,
      "tokens_seen": 2304442368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068888888888889,
      "loss": 2.6592,
      "theoretical_loss": 3.390649685124716,
      "tokens_seen": 2304507904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030686868686868684,
      "loss": 2.5078,
      "theoretical_loss": 3.3906418678628207,
      "tokens_seen": 2304573440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068484848484849,
      "loss": 2.5299,
      "theoretical_loss": 3.3906340508854673,
      "tokens_seen": 2304638976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030682828282828286,
      "loss": 2.3896,
      "theoretical_loss": 3.3906262341926374,
      "tokens_seen": 2304704512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068080808080808,
      "loss": 2.4701,
      "theoretical_loss": 3.3906184177843124,
      "tokens_seen": 2304770048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030678787878787877,
      "loss": 2.6287,
      "theoretical_loss": 3.390610601660474,
      "tokens_seen": 2304835584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030676767676767673,
      "loss": 2.583,
      "theoretical_loss": 3.3906027858211036,
      "tokens_seen": 2304901120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067474747474748,
      "loss": 2.8207,
      "theoretical_loss": 3.3905949702661826,
      "tokens_seen": 2304966656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030672727272727275,
      "loss": 2.2541,
      "theoretical_loss": 3.3905871549956927,
      "tokens_seen": 2305032192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067070707070707,
      "loss": 2.465,
      "theoretical_loss": 3.390579340009616,
      "tokens_seen": 2305097728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030668686868686866,
      "loss": 2.2704,
      "theoretical_loss": 3.390571525307933,
      "tokens_seen": 2305163264
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066666666666667,
      "loss": 2.5124,
      "theoretical_loss": 3.390563710890626,
      "tokens_seen": 2305228800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066464646464647,
      "loss": 2.8017,
      "theoretical_loss": 3.3905558967576765,
      "tokens_seen": 2305294336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030662626262626264,
      "loss": 2.7332,
      "theoretical_loss": 3.390548082909066,
      "tokens_seen": 2305359872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066060606060606,
      "loss": 2.5018,
      "theoretical_loss": 3.3905402693447764,
      "tokens_seen": 2305425408
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1272482,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6973299980163574,
      "objective/train/theoretical_loss": 3.3905324560647885,
      "objective/train/tokens_used": 664350176,
      "theoretical_loss": 3.3905324560647885,
      "tokens_seen": 2305490944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030658585858585856,
      "loss": 2.5932,
      "theoretical_loss": 3.3905324560647885,
      "tokens_seen": 2305490944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030656565656565657,
      "loss": 2.5374,
      "theoretical_loss": 3.3905246430690843,
      "tokens_seen": 2305556480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065454545454546,
      "loss": 2.5087,
      "theoretical_loss": 3.3905168303576456,
      "tokens_seen": 2305622016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030652525252525254,
      "loss": 2.5229,
      "theoretical_loss": 3.3905090179304533,
      "tokens_seen": 2305687552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065050505050505,
      "loss": 2.6958,
      "theoretical_loss": 3.39050120578749,
      "tokens_seen": 2305753088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064848484848485,
      "loss": 2.3973,
      "theoretical_loss": 3.3904933939287365,
      "tokens_seen": 2305818624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030646464646464646,
      "loss": 2.3898,
      "theoretical_loss": 3.3904855823541746,
      "tokens_seen": 2305884160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030644444444444447,
      "loss": 2.6596,
      "theoretical_loss": 3.390477771063786,
      "tokens_seen": 2305949696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030642424242424243,
      "loss": 2.5784,
      "theoretical_loss": 3.390469960057552,
      "tokens_seen": 2306015232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064040404040404,
      "loss": 2.5346,
      "theoretical_loss": 3.3904621493354545,
      "tokens_seen": 2306080768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063838383838384,
      "loss": 2.5161,
      "theoretical_loss": 3.3904543388974746,
      "tokens_seen": 2306146304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030636363636363635,
      "loss": 2.4475,
      "theoretical_loss": 3.3904465287435945,
      "tokens_seen": 2306211840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030634343434343437,
      "loss": 2.7658,
      "theoretical_loss": 3.3904387188737957,
      "tokens_seen": 2306277376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063232323232323,
      "loss": 2.4109,
      "theoretical_loss": 3.3904309092880593,
      "tokens_seen": 2306342912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063030303030303,
      "loss": 2.4875,
      "theoretical_loss": 3.390423099986368,
      "tokens_seen": 2306408448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062828282828283,
      "loss": 2.6148,
      "theoretical_loss": 3.3904152909687015,
      "tokens_seen": 2306473984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030626262626262625,
      "loss": 2.8049,
      "theoretical_loss": 3.390407482235043,
      "tokens_seen": 2306539520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030624242424242426,
      "loss": 2.6674,
      "theoretical_loss": 3.3903996737853737,
      "tokens_seen": 2306605056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062222222222222,
      "loss": 2.5598,
      "theoretical_loss": 3.390391865619675,
      "tokens_seen": 2306670592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062020202020202,
      "loss": 2.4708,
      "theoretical_loss": 3.3903840577379287,
      "tokens_seen": 2306736128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061818181818182,
      "loss": 2.6744,
      "theoretical_loss": 3.3903762501401165,
      "tokens_seen": 2306801664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030616161616161614,
      "loss": 2.5458,
      "theoretical_loss": 3.3903684428262197,
      "tokens_seen": 2306867200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030614141414141415,
      "loss": 2.5156,
      "theoretical_loss": 3.3903606357962204,
      "tokens_seen": 2306932736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061212121212121,
      "loss": 2.6062,
      "theoretical_loss": 3.3903528290500993,
      "tokens_seen": 2306998272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061010101010101,
      "loss": 2.6043,
      "theoretical_loss": 3.390345022587839,
      "tokens_seen": 2307063808
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1273129,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3374531269073486,
      "objective/train/theoretical_loss": 3.3903372164094208,
      "objective/train/tokens_used": 665988576,
      "theoretical_loss": 3.3903372164094208,
      "tokens_seen": 2307129344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060808080808081,
      "loss": 2.4758,
      "theoretical_loss": 3.3903372164094208,
      "tokens_seen": 2307129344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030606060606060603,
      "loss": 2.5465,
      "theoretical_loss": 3.390329410514826,
      "tokens_seen": 2307194880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030604040404040404,
      "loss": 2.3914,
      "theoretical_loss": 3.3903216049040363,
      "tokens_seen": 2307260416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030602020202020205,
      "loss": 2.5634,
      "theoretical_loss": 3.3903137995770334,
      "tokens_seen": 2307325952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306,
      "loss": 2.2768,
      "theoretical_loss": 3.3903059945337994,
      "tokens_seen": 2307391488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030597979797979797,
      "loss": 2.7051,
      "theoretical_loss": 3.390298189774315,
      "tokens_seen": 2307457024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000305959595959596,
      "loss": 2.4825,
      "theoretical_loss": 3.3902903852985626,
      "tokens_seen": 2307522560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030593939393939394,
      "loss": 2.4341,
      "theoretical_loss": 3.390282581106524,
      "tokens_seen": 2307588096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030591919191919195,
      "loss": 2.6403,
      "theoretical_loss": 3.39027477719818,
      "tokens_seen": 2307653632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058989898989899,
      "loss": 2.7605,
      "theoretical_loss": 3.3902669735735125,
      "tokens_seen": 2307719168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030587878787878786,
      "loss": 2.4623,
      "theoretical_loss": 3.390259170232503,
      "tokens_seen": 2307784704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030585858585858587,
      "loss": 2.5519,
      "theoretical_loss": 3.390251367175134,
      "tokens_seen": 2307850240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058383838383839,
      "loss": 2.4641,
      "theoretical_loss": 3.390243564401386,
      "tokens_seen": 2307915776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030581818181818184,
      "loss": 2.6751,
      "theoretical_loss": 3.3902357619112413,
      "tokens_seen": 2307981312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057979797979798,
      "loss": 2.5488,
      "theoretical_loss": 3.390227959704682,
      "tokens_seen": 2308046848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030577777777777775,
      "loss": 2.723,
      "theoretical_loss": 3.3902201577816884,
      "tokens_seen": 2308112384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030575757575757576,
      "loss": 2.5285,
      "theoretical_loss": 3.390212356142243,
      "tokens_seen": 2308177920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057373737373738,
      "loss": 2.645,
      "theoretical_loss": 3.3902045547863273,
      "tokens_seen": 2308243456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030571717171717173,
      "loss": 2.6811,
      "theoretical_loss": 3.3901967537139233,
      "tokens_seen": 2308308992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056969696969697,
      "loss": 2.5863,
      "theoretical_loss": 3.390188952925012,
      "tokens_seen": 2308374528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030567676767676765,
      "loss": 2.5006,
      "theoretical_loss": 3.3901811524195757,
      "tokens_seen": 2308440064
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056565656565657,
      "loss": 2.4715,
      "theoretical_loss": 3.3901733521975954,
      "tokens_seen": 2308505600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030563636363636367,
      "loss": 2.5382,
      "theoretical_loss": 3.3901655522590532,
      "tokens_seen": 2308571136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056161616161616,
      "loss": 2.6117,
      "theoretical_loss": 3.3901577526039306,
      "tokens_seen": 2308636672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003055959595959596,
      "loss": 2.6662,
      "theoretical_loss": 3.3901499532322097,
      "tokens_seen": 2308702208
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 1274350,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.294728994369507,
      "objective/train/theoretical_loss": 3.390142154143871,
      "objective/train/tokens_used": 667626976,
      "theoretical_loss": 3.390142154143871,
      "tokens_seen": 2308767744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030557575757575754,
      "loss": 2.5048,
      "theoretical_loss": 3.390142154143871,
      "tokens_seen": 2308767744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055555555555556,
      "loss": 2.6972,
      "theoretical_loss": 3.3901343553388976,
      "tokens_seen": 2308833280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030553535353535356,
      "loss": 2.5423,
      "theoretical_loss": 3.39012655681727,
      "tokens_seen": 2308898816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055151515151515,
      "loss": 2.54,
      "theoretical_loss": 3.3901187585789705,
      "tokens_seen": 2308964352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054949494949495,
      "loss": 2.5325,
      "theoretical_loss": 3.3901109606239808,
      "tokens_seen": 2309029888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054747474747475,
      "loss": 2.5189,
      "theoretical_loss": 3.3901031629522818,
      "tokens_seen": 2309095424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054545454545455,
      "loss": 2.7139,
      "theoretical_loss": 3.390095365563856,
      "tokens_seen": 2309160960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030543434343434345,
      "loss": 2.5577,
      "theoretical_loss": 3.3900875684586853,
      "tokens_seen": 2309226496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054141414141414,
      "loss": 2.5071,
      "theoretical_loss": 3.3900797716367506,
      "tokens_seen": 2309292032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030539393939393937,
      "loss": 2.7558,
      "theoretical_loss": 3.3900719750980333,
      "tokens_seen": 2309357568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053737373737374,
      "loss": 2.3048,
      "theoretical_loss": 3.390064178842516,
      "tokens_seen": 2309423104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053535353535354,
      "loss": 2.6012,
      "theoretical_loss": 3.3900563828701804,
      "tokens_seen": 2309488640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030533333333333335,
      "loss": 2.7033,
      "theoretical_loss": 3.390048587181007,
      "tokens_seen": 2309554176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053131313131313,
      "loss": 2.9447,
      "theoretical_loss": 3.390040791774979,
      "tokens_seen": 2309619712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052929292929293,
      "loss": 2.3706,
      "theoretical_loss": 3.390032996652077,
      "tokens_seen": 2309685248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030527272727272727,
      "loss": 2.5489,
      "theoretical_loss": 3.390025201812283,
      "tokens_seen": 2309750784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052525252525253,
      "loss": 2.6724,
      "theoretical_loss": 3.390017407255579,
      "tokens_seen": 2309816320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030523232323232324,
      "loss": 2.6711,
      "theoretical_loss": 3.3900096129819457,
      "tokens_seen": 2309881856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052121212121212,
      "loss": 2.8219,
      "theoretical_loss": 3.390001818991366,
      "tokens_seen": 2309947392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003051919191919192,
      "loss": 2.6027,
      "theoretical_loss": 3.389994025283821,
      "tokens_seen": 2310012928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030517171717171716,
      "loss": 2.5895,
      "theoretical_loss": 3.3899862318592926,
      "tokens_seen": 2310078464
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003051515151515152,
      "loss": 2.5967,
      "theoretical_loss": 3.3899784387177623,
      "tokens_seen": 2310144000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030513131313131313,
      "loss": 2.5548,
      "theoretical_loss": 3.389970645859212,
      "tokens_seen": 2310209536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030511111111111114,
      "loss": 2.6856,
      "theoretical_loss": 3.389962853283623,
      "tokens_seen": 2310275072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050909090909091,
      "loss": 2.6099,
      "theoretical_loss": 3.389955060990977,
      "tokens_seen": 2310340608
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1274732,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.461940050125122,
      "objective/train/theoretical_loss": 3.389947268981256,
      "objective/train/tokens_used": 669265376,
      "theoretical_loss": 3.389947268981256,
      "tokens_seen": 2310406144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030507070707070706,
      "loss": 2.354,
      "theoretical_loss": 3.389947268981256,
      "tokens_seen": 2310406144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030505050505050507,
      "loss": 2.5882,
      "theoretical_loss": 3.3899394772544422,
      "tokens_seen": 2310471680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305030303030303,
      "loss": 2.517,
      "theoretical_loss": 3.3899316858105166,
      "tokens_seen": 2310537216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030501010101010104,
      "loss": 2.6366,
      "theoretical_loss": 3.3899238946494608,
      "tokens_seen": 2310602752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304989898989899,
      "loss": 2.6087,
      "theoretical_loss": 3.3899161037712573,
      "tokens_seen": 2310668288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030496969696969695,
      "loss": 2.5083,
      "theoretical_loss": 3.3899083131758867,
      "tokens_seen": 2310733824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030494949494949496,
      "loss": 2.4066,
      "theoretical_loss": 3.3899005228633317,
      "tokens_seen": 2310799360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030492929292929297,
      "loss": 2.3075,
      "theoretical_loss": 3.389892732833573,
      "tokens_seen": 2310864896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030490909090909093,
      "loss": 2.4479,
      "theoretical_loss": 3.3898849430865936,
      "tokens_seen": 2310930432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048888888888889,
      "loss": 2.635,
      "theoretical_loss": 3.389877153622374,
      "tokens_seen": 2310995968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030486868686868684,
      "loss": 2.6388,
      "theoretical_loss": 3.389869364440897,
      "tokens_seen": 2311061504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030484848484848485,
      "loss": 2.4726,
      "theoretical_loss": 3.3898615755421435,
      "tokens_seen": 2311127040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030482828282828286,
      "loss": 2.4258,
      "theoretical_loss": 3.389853786926096,
      "tokens_seen": 2311192576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048080808080808,
      "loss": 2.554,
      "theoretical_loss": 3.389845998592735,
      "tokens_seen": 2311258112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047878787878788,
      "loss": 2.6835,
      "theoretical_loss": 3.3898382105420435,
      "tokens_seen": 2311323648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030476767676767673,
      "loss": 2.5725,
      "theoretical_loss": 3.3898304227740024,
      "tokens_seen": 2311389184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047474747474748,
      "loss": 2.4118,
      "theoretical_loss": 3.3898226352885934,
      "tokens_seen": 2311454720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030472727272727276,
      "loss": 2.6774,
      "theoretical_loss": 3.389814848085799,
      "tokens_seen": 2311520256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047070707070707,
      "loss": 2.548,
      "theoretical_loss": 3.3898070611656,
      "tokens_seen": 2311585792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030468686868686867,
      "loss": 2.6404,
      "theoretical_loss": 3.389799274527979,
      "tokens_seen": 2311651328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030466666666666663,
      "loss": 2.3368,
      "theoretical_loss": 3.389791488172917,
      "tokens_seen": 2311716864
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046464646464647,
      "loss": 2.7248,
      "theoretical_loss": 3.3897837021003965,
      "tokens_seen": 2311782400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030462626262626265,
      "loss": 2.5494,
      "theoretical_loss": 3.3897759163103984,
      "tokens_seen": 2311847936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046060606060606,
      "loss": 2.7867,
      "theoretical_loss": 3.3897681308029055,
      "tokens_seen": 2311913472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030458585858585856,
      "loss": 2.5334,
      "theoretical_loss": 3.389760345577898,
      "tokens_seen": 2311979008
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1275645,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.752490520477295,
      "objective/train/theoretical_loss": 3.389752560635359,
      "objective/train/tokens_used": 670903776,
      "theoretical_loss": 3.389752560635359,
      "tokens_seen": 2312044544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045656565656565,
      "loss": 2.6714,
      "theoretical_loss": 3.389752560635359,
      "tokens_seen": 2312044544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045454545454546,
      "loss": 2.6804,
      "theoretical_loss": 3.38974477597527,
      "tokens_seen": 2312110080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030452525252525254,
      "loss": 2.7004,
      "theoretical_loss": 3.389736991597612,
      "tokens_seen": 2312175616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045050505050505,
      "loss": 2.4943,
      "theoretical_loss": 3.3897292075023677,
      "tokens_seen": 2312241152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030448484848484846,
      "loss": 2.5386,
      "theoretical_loss": 3.3897214236895183,
      "tokens_seen": 2312306688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030446464646464647,
      "loss": 2.6384,
      "theoretical_loss": 3.3897136401590457,
      "tokens_seen": 2312372224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044444444444445,
      "loss": 2.5939,
      "theoretical_loss": 3.3897058569109317,
      "tokens_seen": 2312437760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030442424242424243,
      "loss": 2.6308,
      "theoretical_loss": 3.3896980739451577,
      "tokens_seen": 2312503296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044040404040404,
      "loss": 2.5855,
      "theoretical_loss": 3.3896902912617057,
      "tokens_seen": 2312568832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030438383838383835,
      "loss": 2.6419,
      "theoretical_loss": 3.3896825088605578,
      "tokens_seen": 2312634368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030436363636363636,
      "loss": 2.4153,
      "theoretical_loss": 3.3896747267416956,
      "tokens_seen": 2312699904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030434343434343437,
      "loss": 2.6084,
      "theoretical_loss": 3.3896669449051,
      "tokens_seen": 2312765440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030432323232323233,
      "loss": 2.4422,
      "theoretical_loss": 3.3896591633507542,
      "tokens_seen": 2312830976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043030303030303,
      "loss": 2.7781,
      "theoretical_loss": 3.3896513820786387,
      "tokens_seen": 2312896512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042828282828283,
      "loss": 2.6358,
      "theoretical_loss": 3.3896436010887365,
      "tokens_seen": 2312962048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042626262626263,
      "loss": 2.5163,
      "theoretical_loss": 3.389635820381028,
      "tokens_seen": 2313027584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030424242424242426,
      "loss": 2.5995,
      "theoretical_loss": 3.3896280399554963,
      "tokens_seen": 2313093120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042222222222222,
      "loss": 2.9271,
      "theoretical_loss": 3.389620259812122,
      "tokens_seen": 2313158656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042020202020202,
      "loss": 2.6007,
      "theoretical_loss": 3.3896124799508875,
      "tokens_seen": 2313224192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041818181818182,
      "loss": 2.8641,
      "theoretical_loss": 3.3896047003717746,
      "tokens_seen": 2313289728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041616161616162,
      "loss": 2.6169,
      "theoretical_loss": 3.389596921074765,
      "tokens_seen": 2313355264
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030414141414141416,
      "loss": 2.7031,
      "theoretical_loss": 3.3895891420598403,
      "tokens_seen": 2313420800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041212121212121,
      "loss": 2.7717,
      "theoretical_loss": 3.3895813633269825,
      "tokens_seen": 2313486336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041010101010101,
      "loss": 2.7382,
      "theoretical_loss": 3.389573584876173,
      "tokens_seen": 2313551872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040808080808081,
      "loss": 2.6476,
      "theoretical_loss": 3.3895658067073944,
      "tokens_seen": 2313617408
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1276618,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.560056209564209,
      "objective/train/theoretical_loss": 3.3895580288206277,
      "objective/train/tokens_used": 672542176,
      "theoretical_loss": 3.3895580288206277,
      "tokens_seen": 2313682944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040606060606061,
      "loss": 2.5657,
      "theoretical_loss": 3.3895580288206277,
      "tokens_seen": 2313682944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030404040404040405,
      "loss": 2.5529,
      "theoretical_loss": 3.389550251215855,
      "tokens_seen": 2313748480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304020202020202,
      "loss": 2.6101,
      "theoretical_loss": 3.389542473893058,
      "tokens_seen": 2313814016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304,
      "loss": 2.7534,
      "theoretical_loss": 3.389534696852219,
      "tokens_seen": 2313879552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303979797979798,
      "loss": 2.5307,
      "theoretical_loss": 3.3895269200933185,
      "tokens_seen": 2313945088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303959595959596,
      "loss": 2.6574,
      "theoretical_loss": 3.3895191436163397,
      "tokens_seen": 2314010624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030393939393939394,
      "loss": 2.5521,
      "theoretical_loss": 3.3895113674212634,
      "tokens_seen": 2314076160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030391919191919195,
      "loss": 2.3814,
      "theoretical_loss": 3.3895035915080722,
      "tokens_seen": 2314141696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038989898989899,
      "loss": 2.594,
      "theoretical_loss": 3.389495815876747,
      "tokens_seen": 2314207232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030387878787878787,
      "loss": 2.6072,
      "theoretical_loss": 3.389488040527271,
      "tokens_seen": 2314272768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038585858585859,
      "loss": 2.7825,
      "theoretical_loss": 3.3894802654596243,
      "tokens_seen": 2314338304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030383838383838383,
      "loss": 2.6401,
      "theoretical_loss": 3.38947249067379,
      "tokens_seen": 2314403840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030381818181818185,
      "loss": 2.6194,
      "theoretical_loss": 3.389464716169749,
      "tokens_seen": 2314469376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037979797979798,
      "loss": 2.6753,
      "theoretical_loss": 3.3894569419474836,
      "tokens_seen": 2314534912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030377777777777776,
      "loss": 2.8071,
      "theoretical_loss": 3.389449168006976,
      "tokens_seen": 2314600448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030375757575757577,
      "loss": 2.5532,
      "theoretical_loss": 3.389441394348207,
      "tokens_seen": 2314665984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037373737373738,
      "loss": 2.7082,
      "theoretical_loss": 3.3894336209711593,
      "tokens_seen": 2314731520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030371717171717174,
      "loss": 2.7112,
      "theoretical_loss": 3.3894258478758146,
      "tokens_seen": 2314797056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003036969696969697,
      "loss": 2.6809,
      "theoretical_loss": 3.3894180750621543,
      "tokens_seen": 2314862592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030367676767676765,
      "loss": 2.6255,
      "theoretical_loss": 3.3894103025301603,
      "tokens_seen": 2314928128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030365656565656566,
      "loss": 2.6545,
      "theoretical_loss": 3.3894025302798143,
      "tokens_seen": 2314993664
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003036363636363637,
      "loss": 2.8286,
      "theoretical_loss": 3.3893947583110986,
      "tokens_seen": 2315059200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030361616161616163,
      "loss": 2.5691,
      "theoretical_loss": 3.389386986623995,
      "tokens_seen": 2315124736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035959595959596,
      "loss": 2.7007,
      "theoretical_loss": 3.3893792152184847,
      "tokens_seen": 2315190272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030357575757575754,
      "loss": 2.5359,
      "theoretical_loss": 3.3893714440945506,
      "tokens_seen": 2315255808
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1277260,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.181581735610962,
      "objective/train/theoretical_loss": 3.3893636732521735,
      "objective/train/tokens_used": 674180576,
      "theoretical_loss": 3.3893636732521735,
      "tokens_seen": 2315321344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035555555555556,
      "loss": 2.7444,
      "theoretical_loss": 3.3893636732521735,
      "tokens_seen": 2315321344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030353535353535357,
      "loss": 2.5262,
      "theoretical_loss": 3.389355902691335,
      "tokens_seen": 2315386880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035151515151515,
      "loss": 2.6335,
      "theoretical_loss": 3.3893481324120183,
      "tokens_seen": 2315452416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034949494949495,
      "loss": 2.769,
      "theoretical_loss": 3.389340362414204,
      "tokens_seen": 2315517952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030347474747474744,
      "loss": 2.7289,
      "theoretical_loss": 3.389332592697875,
      "tokens_seen": 2315583488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034545454545455,
      "loss": 2.6278,
      "theoretical_loss": 3.3893248232630118,
      "tokens_seen": 2315649024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030343434343434346,
      "loss": 2.651,
      "theoretical_loss": 3.3893170541095974,
      "tokens_seen": 2315714560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034141414141414,
      "loss": 2.7817,
      "theoretical_loss": 3.389309285237613,
      "tokens_seen": 2315780096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033939393939394,
      "loss": 2.4756,
      "theoretical_loss": 3.389301516647041,
      "tokens_seen": 2315845632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033737373737374,
      "loss": 2.7765,
      "theoretical_loss": 3.3892937483378622,
      "tokens_seen": 2315911168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033535353535354,
      "loss": 2.4798,
      "theoretical_loss": 3.38928598031006,
      "tokens_seen": 2315976704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030333333333333335,
      "loss": 2.6061,
      "theoretical_loss": 3.3892782125636147,
      "tokens_seen": 2316042240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033131313131313,
      "loss": 2.6794,
      "theoretical_loss": 3.389270445098509,
      "tokens_seen": 2316107776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030329292929292927,
      "loss": 2.5613,
      "theoretical_loss": 3.389262677914725,
      "tokens_seen": 2316173312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032727272727273,
      "loss": 2.6267,
      "theoretical_loss": 3.3892549110122436,
      "tokens_seen": 2316238848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032525252525253,
      "loss": 2.7271,
      "theoretical_loss": 3.3892471443910472,
      "tokens_seen": 2316304384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030323232323232324,
      "loss": 2.7138,
      "theoretical_loss": 3.3892393780511174,
      "tokens_seen": 2316369920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032121212121212,
      "loss": 2.7729,
      "theoretical_loss": 3.389231611992437,
      "tokens_seen": 2316435456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031919191919192,
      "loss": 2.5525,
      "theoretical_loss": 3.389223846214987,
      "tokens_seen": 2316500992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030317171717171717,
      "loss": 2.6584,
      "theoretical_loss": 3.3892160807187492,
      "tokens_seen": 2316566528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031515151515152,
      "loss": 2.6226,
      "theoretical_loss": 3.389208315503706,
      "tokens_seen": 2316632064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030313131313131314,
      "loss": 2.6754,
      "theoretical_loss": 3.3892005505698384,
      "tokens_seen": 2316697600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031111111111111,
      "loss": 2.5844,
      "theoretical_loss": 3.3891927859171287,
      "tokens_seen": 2316763136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030909090909091,
      "loss": 2.6093,
      "theoretical_loss": 3.3891850215455595,
      "tokens_seen": 2316828672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030307070707070706,
      "loss": 2.695,
      "theoretical_loss": 3.3891772574551116,
      "tokens_seen": 2316894208
    },
    {
      "debugging/Self-BLEU-5": 0.20619410364685406,
      "debugging/distinct-1-grams": 0.7574324808367362,
      "debugging/distinct-2-grams": 0.9451559705015926,
      "debugging/entropy-1-grams": 4.6576422156928405,
      "debugging/entropy-2-grams": 5.24391176970891,
      "debugging/length": 578.6666666666666,
      "debugging/num_segments": 3,
      "epoch": 0.4,
      "objective/train/docs_used": 1278459,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2756106853485107,
      "objective/train/theoretical_loss": 3.3891694936457677,
      "objective/train/tokens_used": 675818976,
      "theoretical_loss": 3.3891694936457677,
      "tokens_seen": 2316959744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030505050505051,
      "loss": 2.4249,
      "theoretical_loss": 3.3891694936457677,
      "tokens_seen": 2316959744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030303030303030303,
      "loss": 2.5714,
      "theoretical_loss": 3.389161730117509,
      "tokens_seen": 2317025280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303010101010101,
      "loss": 2.6624,
      "theoretical_loss": 3.3891539668703174,
      "tokens_seen": 2317090816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302989898989899,
      "loss": 2.5118,
      "theoretical_loss": 3.3891462039041755,
      "tokens_seen": 2317156352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030296969696969695,
      "loss": 2.5119,
      "theoretical_loss": 3.389138441219065,
      "tokens_seen": 2317221888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030294949494949497,
      "loss": 2.5936,
      "theoretical_loss": 3.3891306788149667,
      "tokens_seen": 2317287424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029292929292929,
      "loss": 2.506,
      "theoretical_loss": 3.3891229166918637,
      "tokens_seen": 2317352960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030290909090909093,
      "loss": 2.7018,
      "theoretical_loss": 3.389115154849737,
      "tokens_seen": 2317418496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028888888888889,
      "loss": 2.5806,
      "theoretical_loss": 3.3891073932885694,
      "tokens_seen": 2317484032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030286868686868685,
      "loss": 2.7083,
      "theoretical_loss": 3.3890996320083424,
      "tokens_seen": 2317549568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030284848484848486,
      "loss": 2.5808,
      "theoretical_loss": 3.3890918710090374,
      "tokens_seen": 2317615104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028282828282828,
      "loss": 2.6755,
      "theoretical_loss": 3.3890841102906366,
      "tokens_seen": 2317680640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028080808080808,
      "loss": 2.6778,
      "theoretical_loss": 3.3890763498531227,
      "tokens_seen": 2317746176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027878787878788,
      "loss": 2.6103,
      "theoretical_loss": 3.3890685896964765,
      "tokens_seen": 2317811712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030276767676767674,
      "loss": 2.4966,
      "theoretical_loss": 3.38906082982068,
      "tokens_seen": 2317877248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030274747474747475,
      "loss": 2.5976,
      "theoretical_loss": 3.3890530702257156,
      "tokens_seen": 2317942784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030272727272727276,
      "loss": 2.5301,
      "theoretical_loss": 3.389045310911565,
      "tokens_seen": 2318008320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027070707070707,
      "loss": 2.6838,
      "theoretical_loss": 3.38903755187821,
      "tokens_seen": 2318073856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026868686868687,
      "loss": 2.5929,
      "theoretical_loss": 3.3890297931256326,
      "tokens_seen": 2318139392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030266666666666663,
      "loss": 2.2318,
      "theoretical_loss": 3.3890220346538142,
      "tokens_seen": 2318204928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030264646464646464,
      "loss": 2.4557,
      "theoretical_loss": 3.3890142764627376,
      "tokens_seen": 2318270464
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030262626262626266,
      "loss": 2.6753,
      "theoretical_loss": 3.3890065185523843,
      "tokens_seen": 2318336000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026060606060606,
      "loss": 2.7227,
      "theoretical_loss": 3.3889987609227363,
      "tokens_seen": 2318401536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030258585858585857,
      "loss": 2.6271,
      "theoretical_loss": 3.388991003573775,
      "tokens_seen": 2318467072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025656565656566,
      "loss": 2.6011,
      "theoretical_loss": 3.3889832465054828,
      "tokens_seen": 2318532608
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1279139,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5781471729278564,
      "objective/train/theoretical_loss": 3.3889754897178417,
      "objective/train/tokens_used": 677457376,
      "theoretical_loss": 3.3889754897178417,
      "tokens_seen": 2318598144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025454545454546,
      "loss": 2.513,
      "theoretical_loss": 3.3889754897178417,
      "tokens_seen": 2318598144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030252525252525255,
      "loss": 2.6527,
      "theoretical_loss": 3.3889677332108334,
      "tokens_seen": 2318663680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025050505050505,
      "loss": 2.5918,
      "theoretical_loss": 3.38895997698444,
      "tokens_seen": 2318729216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030248484848484846,
      "loss": 2.5136,
      "theoretical_loss": 3.3889522210386427,
      "tokens_seen": 2318794752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030246464646464647,
      "loss": 2.7214,
      "theoretical_loss": 3.388944465373424,
      "tokens_seen": 2318860288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024444444444445,
      "loss": 2.2814,
      "theoretical_loss": 3.388936709988766,
      "tokens_seen": 2318925824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030242424242424244,
      "loss": 2.3996,
      "theoretical_loss": 3.388928954884651,
      "tokens_seen": 2318991360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024040404040404,
      "loss": 2.6588,
      "theoretical_loss": 3.3889212000610596,
      "tokens_seen": 2319056896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030238383838383835,
      "loss": 2.7012,
      "theoretical_loss": 3.3889134455179746,
      "tokens_seen": 2319122432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023636363636364,
      "loss": 2.488,
      "theoretical_loss": 3.3889056912553777,
      "tokens_seen": 2319187968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023434343434344,
      "loss": 2.5442,
      "theoretical_loss": 3.388897937273251,
      "tokens_seen": 2319253504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030232323232323233,
      "loss": 2.7,
      "theoretical_loss": 3.3888901835715766,
      "tokens_seen": 2319319040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023030303030303,
      "loss": 2.6603,
      "theoretical_loss": 3.388882430150336,
      "tokens_seen": 2319384576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030228282828282825,
      "loss": 2.3551,
      "theoretical_loss": 3.3888746770095115,
      "tokens_seen": 2319450112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022626262626263,
      "loss": 2.6881,
      "theoretical_loss": 3.3888669241490845,
      "tokens_seen": 2319515648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030224242424242427,
      "loss": 2.7914,
      "theoretical_loss": 3.3888591715690377,
      "tokens_seen": 2319581184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022222222222222,
      "loss": 2.6074,
      "theoretical_loss": 3.3888514192693524,
      "tokens_seen": 2319646720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022020202020202,
      "loss": 2.7748,
      "theoretical_loss": 3.388843667250011,
      "tokens_seen": 2319712256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021818181818182,
      "loss": 2.7027,
      "theoretical_loss": 3.388835915510995,
      "tokens_seen": 2319777792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021616161616162,
      "loss": 2.5539,
      "theoretical_loss": 3.3888281640522866,
      "tokens_seen": 2319843328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030214141414141416,
      "loss": 2.4172,
      "theoretical_loss": 3.3888204128738675,
      "tokens_seen": 2319908864
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021212121212121,
      "loss": 2.5263,
      "theoretical_loss": 3.3888126619757197,
      "tokens_seen": 2319974400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021010101010101,
      "loss": 2.5707,
      "theoretical_loss": 3.388804911357826,
      "tokens_seen": 2320039936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020808080808081,
      "loss": 2.6024,
      "theoretical_loss": 3.3887971610201673,
      "tokens_seen": 2320105472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020606060606061,
      "loss": 2.5436,
      "theoretical_loss": 3.3887894109627257,
      "tokens_seen": 2320171008
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1280192,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5417916774749756,
      "objective/train/theoretical_loss": 3.3887816611854835,
      "objective/train/tokens_used": 679095776,
      "theoretical_loss": 3.3887816611854835,
      "tokens_seen": 2320236544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030204040404040405,
      "loss": 2.9523,
      "theoretical_loss": 3.3887816611854835,
      "tokens_seen": 2320236544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302020202020202,
      "loss": 2.6333,
      "theoretical_loss": 3.388773911688422,
      "tokens_seen": 2320302080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302,
      "loss": 2.6155,
      "theoretical_loss": 3.3887661624715246,
      "tokens_seen": 2320367616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301979797979798,
      "loss": 2.6584,
      "theoretical_loss": 3.388758413534772,
      "tokens_seen": 2320433152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301959595959596,
      "loss": 2.4983,
      "theoretical_loss": 3.3887506648781462,
      "tokens_seen": 2320498688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030193939393939395,
      "loss": 2.494,
      "theoretical_loss": 3.38874291650163,
      "tokens_seen": 2320564224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003019191919191919,
      "loss": 2.516,
      "theoretical_loss": 3.388735168405204,
      "tokens_seen": 2320629760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003018989898989899,
      "loss": 2.2484,
      "theoretical_loss": 3.3887274205888516,
      "tokens_seen": 2320695296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030187878787878787,
      "loss": 2.6002,
      "theoretical_loss": 3.388719673052554,
      "tokens_seen": 2320760832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003018585858585859,
      "loss": 2.6474,
      "theoretical_loss": 3.388711925796293,
      "tokens_seen": 2320826368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030183838383838384,
      "loss": 2.4902,
      "theoretical_loss": 3.3887041788200514,
      "tokens_seen": 2320891904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030181818181818185,
      "loss": 2.5106,
      "theoretical_loss": 3.3886964321238104,
      "tokens_seen": 2320957440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017979797979798,
      "loss": 2.64,
      "theoretical_loss": 3.388688685707552,
      "tokens_seen": 2321022976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030177777777777776,
      "loss": 2.699,
      "theoretical_loss": 3.388680939571259,
      "tokens_seen": 2321088512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017575757575758,
      "loss": 2.7201,
      "theoretical_loss": 3.3886731937149124,
      "tokens_seen": 2321154048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030173737373737373,
      "loss": 2.4053,
      "theoretical_loss": 3.388665448138495,
      "tokens_seen": 2321219584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030171717171717174,
      "loss": 2.5413,
      "theoretical_loss": 3.3886577028419875,
      "tokens_seen": 2321285120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016969696969697,
      "loss": 2.3769,
      "theoretical_loss": 3.3886499578253737,
      "tokens_seen": 2321350656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030167676767676766,
      "loss": 2.6089,
      "theoretical_loss": 3.388642213088634,
      "tokens_seen": 2321416192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030165656565656567,
      "loss": 2.4305,
      "theoretical_loss": 3.388634468631751,
      "tokens_seen": 2321481728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016363636363637,
      "loss": 2.586,
      "theoretical_loss": 3.388626724454707,
      "tokens_seen": 2321547264
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030161616161616164,
      "loss": 2.6987,
      "theoretical_loss": 3.3886189805574833,
      "tokens_seen": 2321612800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015959595959596,
      "loss": 2.6412,
      "theoretical_loss": 3.3886112369400623,
      "tokens_seen": 2321678336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030157575757575755,
      "loss": 2.5613,
      "theoretical_loss": 3.3886034936024263,
      "tokens_seen": 2321743872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030155555555555556,
      "loss": 2.5951,
      "theoretical_loss": 3.3885957505445568,
      "tokens_seen": 2321809408
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1280928,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7444186210632324,
      "objective/train/theoretical_loss": 3.388588007766436,
      "objective/train/tokens_used": 680734176,
      "theoretical_loss": 3.388588007766436,
      "tokens_seen": 2321874944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030153535353535357,
      "loss": 2.5264,
      "theoretical_loss": 3.388588007766436,
      "tokens_seen": 2321874944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030151515151515153,
      "loss": 2.6344,
      "theoretical_loss": 3.3885802652680455,
      "tokens_seen": 2321940480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014949494949495,
      "loss": 2.6376,
      "theoretical_loss": 3.3885725230493677,
      "tokens_seen": 2322006016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030147474747474744,
      "loss": 2.5501,
      "theoretical_loss": 3.388564781110385,
      "tokens_seen": 2322071552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014545454545455,
      "loss": 2.6154,
      "theoretical_loss": 3.3885570394510784,
      "tokens_seen": 2322137088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030143434343434346,
      "loss": 2.7452,
      "theoretical_loss": 3.388549298071431,
      "tokens_seen": 2322202624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014141414141414,
      "loss": 2.6932,
      "theoretical_loss": 3.3885415569714237,
      "tokens_seen": 2322268160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013939393939394,
      "loss": 2.6463,
      "theoretical_loss": 3.3885338161510394,
      "tokens_seen": 2322333696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030137373737373734,
      "loss": 2.7288,
      "theoretical_loss": 3.3885260756102595,
      "tokens_seen": 2322399232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013535353535354,
      "loss": 2.5697,
      "theoretical_loss": 3.3885183353490667,
      "tokens_seen": 2322464768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030133333333333336,
      "loss": 2.5758,
      "theoretical_loss": 3.388510595367442,
      "tokens_seen": 2322530304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013131313131313,
      "loss": 2.4679,
      "theoretical_loss": 3.3885028556653687,
      "tokens_seen": 2322595840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030129292929292927,
      "loss": 2.5451,
      "theoretical_loss": 3.3884951162428276,
      "tokens_seen": 2322661376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030127272727272723,
      "loss": 2.5937,
      "theoretical_loss": 3.3884873770998016,
      "tokens_seen": 2322726912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012525252525253,
      "loss": 2.549,
      "theoretical_loss": 3.388479638236272,
      "tokens_seen": 2322792448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030123232323232325,
      "loss": 2.672,
      "theoretical_loss": 3.3884718996522216,
      "tokens_seen": 2322857984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012121212121212,
      "loss": 2.483,
      "theoretical_loss": 3.388464161347632,
      "tokens_seen": 2322923520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030119191919191916,
      "loss": 2.3783,
      "theoretical_loss": 3.3884564233224848,
      "tokens_seen": 2322989056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011717171717172,
      "loss": 2.6028,
      "theoretical_loss": 3.388448685576763,
      "tokens_seen": 2323054592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011515151515152,
      "loss": 2.5197,
      "theoretical_loss": 3.3884409481104476,
      "tokens_seen": 2323120128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030113131313131314,
      "loss": 2.7962,
      "theoretical_loss": 3.3884332109235213,
      "tokens_seen": 2323185664
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011111111111111,
      "loss": 2.619,
      "theoretical_loss": 3.388425474015966,
      "tokens_seen": 2323251200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030109090909090906,
      "loss": 2.5069,
      "theoretical_loss": 3.3884177373877633,
      "tokens_seen": 2323316736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030107070707070707,
      "loss": 2.7151,
      "theoretical_loss": 3.388410001038896,
      "tokens_seen": 2323382272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010505050505051,
      "loss": 2.7642,
      "theoretical_loss": 3.3884022649693453,
      "tokens_seen": 2323447808
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1282083,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3946757316589355,
      "objective/train/theoretical_loss": 3.3883945291790942,
      "objective/train/tokens_used": 682372576,
      "theoretical_loss": 3.3883945291790942,
      "tokens_seen": 2323513344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030103030303030304,
      "loss": 2.5281,
      "theoretical_loss": 3.3883945291790942,
      "tokens_seen": 2323513344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301010101010101,
      "loss": 2.4667,
      "theoretical_loss": 3.3883867936681242,
      "tokens_seen": 2323578880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000300989898989899,
      "loss": 2.551,
      "theoretical_loss": 3.388379058436417,
      "tokens_seen": 2323644416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030096969696969696,
      "loss": 2.7176,
      "theoretical_loss": 3.3883713234839554,
      "tokens_seen": 2323709952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030094949494949497,
      "loss": 2.2652,
      "theoretical_loss": 3.3883635888107206,
      "tokens_seen": 2323775488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030092929292929293,
      "loss": 2.5805,
      "theoretical_loss": 3.3883558544166954,
      "tokens_seen": 2323841024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003009090909090909,
      "loss": 2.433,
      "theoretical_loss": 3.3883481203018615,
      "tokens_seen": 2323906560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008888888888889,
      "loss": 2.4721,
      "theoretical_loss": 3.388340386466201,
      "tokens_seen": 2323972096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030086868686868685,
      "loss": 2.4395,
      "theoretical_loss": 3.3883326529096958,
      "tokens_seen": 2324037632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030084848484848486,
      "loss": 2.4886,
      "theoretical_loss": 3.3883249196323284,
      "tokens_seen": 2324103168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008282828282828,
      "loss": 2.6109,
      "theoretical_loss": 3.3883171866340804,
      "tokens_seen": 2324168704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030080808080808083,
      "loss": 2.8457,
      "theoretical_loss": 3.3883094539149337,
      "tokens_seen": 2324234240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007878787878788,
      "loss": 2.6223,
      "theoretical_loss": 3.3883017214748707,
      "tokens_seen": 2324299776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007676767676768,
      "loss": 2.4087,
      "theoretical_loss": 3.388293989313874,
      "tokens_seen": 2324365312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030074747474747476,
      "loss": 2.6035,
      "theoretical_loss": 3.3882862574319246,
      "tokens_seen": 2324430848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007272727272727,
      "loss": 2.507,
      "theoretical_loss": 3.388278525829005,
      "tokens_seen": 2324496384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007070707070707,
      "loss": 2.5625,
      "theoretical_loss": 3.3882707945050976,
      "tokens_seen": 2324561920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006868686868687,
      "loss": 2.3879,
      "theoretical_loss": 3.388263063460184,
      "tokens_seen": 2324627456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006666666666667,
      "loss": 2.4293,
      "theoretical_loss": 3.3882553326942464,
      "tokens_seen": 2324692992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030064646464646465,
      "loss": 2.5838,
      "theoretical_loss": 3.388247602207267,
      "tokens_seen": 2324758528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030062626262626266,
      "loss": 2.5775,
      "theoretical_loss": 3.3882398719992275,
      "tokens_seen": 2324824064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006060606060606,
      "loss": 2.73,
      "theoretical_loss": 3.3882321420701103,
      "tokens_seen": 2324889600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003005858585858586,
      "loss": 2.6222,
      "theoretical_loss": 3.3882244124198975,
      "tokens_seen": 2324955136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003005656565656566,
      "loss": 2.586,
      "theoretical_loss": 3.3882166830485714,
      "tokens_seen": 2325020672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030054545454545454,
      "loss": 2.7063,
      "theoretical_loss": 3.3882089539561133,
      "tokens_seen": 2325086208
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 1282722,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1986050605773926,
      "objective/train/theoretical_loss": 3.388201225142506,
      "objective/train/tokens_used": 684010976,
      "theoretical_loss": 3.388201225142506,
      "tokens_seen": 2325151744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030052525252525255,
      "loss": 2.4966,
      "theoretical_loss": 3.388201225142506,
      "tokens_seen": 2325151744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003005050505050505,
      "loss": 2.4373,
      "theoretical_loss": 3.388193496607731,
      "tokens_seen": 2325217280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030048484848484847,
      "loss": 2.512,
      "theoretical_loss": 3.3881857683517715,
      "tokens_seen": 2325282816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004646464646465,
      "loss": 2.6173,
      "theoretical_loss": 3.3881780403746085,
      "tokens_seen": 2325348352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004444444444445,
      "loss": 2.4411,
      "theoretical_loss": 3.388170312676224,
      "tokens_seen": 2325413888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030042424242424245,
      "loss": 2.6006,
      "theoretical_loss": 3.388162585256601,
      "tokens_seen": 2325479424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004040404040404,
      "loss": 2.5983,
      "theoretical_loss": 3.3881548581157204,
      "tokens_seen": 2325544960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030038383838383836,
      "loss": 2.447,
      "theoretical_loss": 3.3881471312535654,
      "tokens_seen": 2325610496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030036363636363637,
      "loss": 2.7848,
      "theoretical_loss": 3.3881394046701176,
      "tokens_seen": 2325676032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003434343434344,
      "loss": 2.5488,
      "theoretical_loss": 3.388131678365359,
      "tokens_seen": 2325741568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030032323232323234,
      "loss": 2.7041,
      "theoretical_loss": 3.3881239523392717,
      "tokens_seen": 2325807104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003030303030303,
      "loss": 2.3782,
      "theoretical_loss": 3.3881162265918383,
      "tokens_seen": 2325872640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030028282828282825,
      "loss": 2.5706,
      "theoretical_loss": 3.3881085011230403,
      "tokens_seen": 2325938176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002626262626263,
      "loss": 2.6527,
      "theoretical_loss": 3.38810077593286,
      "tokens_seen": 2326003712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002424242424243,
      "loss": 2.3732,
      "theoretical_loss": 3.3880930510212797,
      "tokens_seen": 2326069248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030022222222222223,
      "loss": 2.4778,
      "theoretical_loss": 3.388085326388281,
      "tokens_seen": 2326134784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002020202020202,
      "loss": 2.3248,
      "theoretical_loss": 3.388077602033847,
      "tokens_seen": 2326200320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030018181818181815,
      "loss": 2.7419,
      "theoretical_loss": 3.3880698779579586,
      "tokens_seen": 2326265856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001616161616162,
      "loss": 2.725,
      "theoretical_loss": 3.3880621541605986,
      "tokens_seen": 2326331392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030014141414141417,
      "loss": 2.6013,
      "theoretical_loss": 3.388054430641749,
      "tokens_seen": 2326396928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001212121212121,
      "loss": 2.4874,
      "theoretical_loss": 3.388046707401392,
      "tokens_seen": 2326462464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001010101010101,
      "loss": 2.3981,
      "theoretical_loss": 3.3880389844395093,
      "tokens_seen": 2326528000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000808080808081,
      "loss": 2.637,
      "theoretical_loss": 3.3880312617560833,
      "tokens_seen": 2326593536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000606060606061,
      "loss": 2.4284,
      "theoretical_loss": 3.3880235393510962,
      "tokens_seen": 2326659072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030004040404040406,
      "loss": 2.5925,
      "theoretical_loss": 3.38801581722453,
      "tokens_seen": 2326724608
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1283581,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.384247064590454,
      "objective/train/theoretical_loss": 3.388008095376367,
      "objective/train/tokens_used": 685649376,
      "theoretical_loss": 3.388008095376367,
      "tokens_seen": 2326790144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300020202020202,
      "loss": 2.4081,
      "theoretical_loss": 3.388008095376367,
      "tokens_seen": 2326790144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003,
      "loss": 2.6352,
      "theoretical_loss": 3.388000373806589,
      "tokens_seen": 2326855680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299979797979798,
      "loss": 2.6886,
      "theoretical_loss": 3.3879926525151784,
      "tokens_seen": 2326921216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299959595959596,
      "loss": 2.717,
      "theoretical_loss": 3.3879849315021175,
      "tokens_seen": 2326986752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029993939393939395,
      "loss": 2.768,
      "theoretical_loss": 3.387977210767388,
      "tokens_seen": 2327052288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999191919191919,
      "loss": 2.6944,
      "theoretical_loss": 3.387969490310972,
      "tokens_seen": 2327117824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029989898989898987,
      "loss": 2.6222,
      "theoretical_loss": 3.3879617701328515,
      "tokens_seen": 2327183360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998787878787879,
      "loss": 2.5626,
      "theoretical_loss": 3.3879540502330094,
      "tokens_seen": 2327248896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998585858585859,
      "loss": 2.5228,
      "theoretical_loss": 3.3879463306114275,
      "tokens_seen": 2327314432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029983838383838385,
      "loss": 2.6241,
      "theoretical_loss": 3.387938611268088,
      "tokens_seen": 2327379968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998181818181818,
      "loss": 2.4605,
      "theoretical_loss": 3.3879308922029723,
      "tokens_seen": 2327445504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997979797979798,
      "loss": 2.5005,
      "theoretical_loss": 3.387923173416063,
      "tokens_seen": 2327511040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029977777777777777,
      "loss": 2.7255,
      "theoretical_loss": 3.387915454907343,
      "tokens_seen": 2327576576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997575757575758,
      "loss": 2.5728,
      "theoretical_loss": 3.387907736676793,
      "tokens_seen": 2327642112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029973737373737374,
      "loss": 2.5169,
      "theoretical_loss": 3.3879000187243964,
      "tokens_seen": 2327707648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997171717171717,
      "loss": 2.6249,
      "theoretical_loss": 3.387892301050135,
      "tokens_seen": 2327773184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996969696969697,
      "loss": 2.5634,
      "theoretical_loss": 3.3878845836539906,
      "tokens_seen": 2327838720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029967676767676766,
      "loss": 2.5164,
      "theoretical_loss": 3.3878768665359456,
      "tokens_seen": 2327904256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996565656565657,
      "loss": 2.5191,
      "theoretical_loss": 3.387869149695982,
      "tokens_seen": 2327969792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029963636363636363,
      "loss": 2.3382,
      "theoretical_loss": 3.387861433134082,
      "tokens_seen": 2328035328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029961616161616164,
      "loss": 2.497,
      "theoretical_loss": 3.387853716850228,
      "tokens_seen": 2328100864
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995959595959596,
      "loss": 2.7257,
      "theoretical_loss": 3.3878460008444016,
      "tokens_seen": 2328166400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029957575757575756,
      "loss": 2.579,
      "theoretical_loss": 3.387838285116586,
      "tokens_seen": 2328231936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029955555555555557,
      "loss": 2.7085,
      "theoretical_loss": 3.387830569666762,
      "tokens_seen": 2328297472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995353535353535,
      "loss": 2.5562,
      "theoretical_loss": 3.3878228544949125,
      "tokens_seen": 2328363008
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1284315,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.855656147003174,
      "objective/train/theoretical_loss": 3.38781513960102,
      "objective/train/tokens_used": 687287776,
      "theoretical_loss": 3.38781513960102,
      "tokens_seen": 2328428544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029951515151515153,
      "loss": 2.5938,
      "theoretical_loss": 3.38781513960102,
      "tokens_seen": 2328428544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994949494949495,
      "loss": 2.4312,
      "theoretical_loss": 3.3878074249850663,
      "tokens_seen": 2328494080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029947474747474745,
      "loss": 2.6684,
      "theoretical_loss": 3.387799710647033,
      "tokens_seen": 2328559616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029945454545454546,
      "loss": 2.6161,
      "theoretical_loss": 3.387791996586903,
      "tokens_seen": 2328625152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029943434343434347,
      "loss": 2.8061,
      "theoretical_loss": 3.3877842828046587,
      "tokens_seen": 2328690688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029941414141414143,
      "loss": 2.3686,
      "theoretical_loss": 3.3877765693002813,
      "tokens_seen": 2328756224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993939393939394,
      "loss": 2.5241,
      "theoretical_loss": 3.3877688560737536,
      "tokens_seen": 2328821760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029937373737373734,
      "loss": 2.5534,
      "theoretical_loss": 3.387761143125058,
      "tokens_seen": 2328887296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029935353535353535,
      "loss": 2.4191,
      "theoretical_loss": 3.387753430454176,
      "tokens_seen": 2328952832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029933333333333336,
      "loss": 2.4657,
      "theoretical_loss": 3.3877457180610904,
      "tokens_seen": 2329018368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993131313131313,
      "loss": 2.4365,
      "theoretical_loss": 3.3877380059457827,
      "tokens_seen": 2329083904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992929292929293,
      "loss": 2.5397,
      "theoretical_loss": 3.387730294108236,
      "tokens_seen": 2329149440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029927272727272723,
      "loss": 2.5297,
      "theoretical_loss": 3.3877225825484314,
      "tokens_seen": 2329214976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992525252525253,
      "loss": 2.6503,
      "theoretical_loss": 3.387714871266352,
      "tokens_seen": 2329280512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029923232323232326,
      "loss": 2.5606,
      "theoretical_loss": 3.3877071602619795,
      "tokens_seen": 2329346048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992121212121212,
      "loss": 2.6814,
      "theoretical_loss": 3.3876994495352966,
      "tokens_seen": 2329411584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029919191919191917,
      "loss": 2.6162,
      "theoretical_loss": 3.3876917390862844,
      "tokens_seen": 2329477120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991717171717171,
      "loss": 2.5459,
      "theoretical_loss": 3.387684028914926,
      "tokens_seen": 2329542656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991515151515152,
      "loss": 2.5083,
      "theoretical_loss": 3.387676319021204,
      "tokens_seen": 2329608192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029913131313131315,
      "loss": 2.4128,
      "theoretical_loss": 3.3876686094050994,
      "tokens_seen": 2329673728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991111111111111,
      "loss": 2.6386,
      "theoretical_loss": 3.387660900066595,
      "tokens_seen": 2329739264
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029909090909090906,
      "loss": 2.641,
      "theoretical_loss": 3.3876531910056733,
      "tokens_seen": 2329804800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029907070707070713,
      "loss": 2.2965,
      "theoretical_loss": 3.387645482222316,
      "tokens_seen": 2329870336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002990505050505051,
      "loss": 2.2496,
      "theoretical_loss": 3.3876377737165053,
      "tokens_seen": 2329935872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029903030303030304,
      "loss": 2.446,
      "theoretical_loss": 3.387630065488224,
      "tokens_seen": 2330001408
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1285456,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5356266498565674,
      "objective/train/theoretical_loss": 3.3876223575374533,
      "objective/train/tokens_used": 688926176,
      "theoretical_loss": 3.3876223575374533,
      "tokens_seen": 2330066944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299010101010101,
      "loss": 2.743,
      "theoretical_loss": 3.3876223575374533,
      "tokens_seen": 2330066944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029898989898989895,
      "loss": 2.5193,
      "theoretical_loss": 3.387614649864176,
      "tokens_seen": 2330132480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298969696969697,
      "loss": 2.4647,
      "theoretical_loss": 3.387606942468375,
      "tokens_seen": 2330198016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298949494949495,
      "loss": 2.553,
      "theoretical_loss": 3.387599235350031,
      "tokens_seen": 2330263552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029892929292929293,
      "loss": 2.4722,
      "theoretical_loss": 3.387591528509127,
      "tokens_seen": 2330329088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989090909090909,
      "loss": 2.4273,
      "theoretical_loss": 3.387583821945646,
      "tokens_seen": 2330394624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988888888888889,
      "loss": 2.6327,
      "theoretical_loss": 3.3875761156595687,
      "tokens_seen": 2330460160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988686868686869,
      "loss": 2.7975,
      "theoretical_loss": 3.3875684096508785,
      "tokens_seen": 2330525696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029884848484848487,
      "loss": 2.4514,
      "theoretical_loss": 3.3875607039195565,
      "tokens_seen": 2330591232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988282828282828,
      "loss": 2.6605,
      "theoretical_loss": 3.387552998465586,
      "tokens_seen": 2330656768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988080808080808,
      "loss": 2.5183,
      "theoretical_loss": 3.387545293288949,
      "tokens_seen": 2330722304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987878787878788,
      "loss": 2.3649,
      "theoretical_loss": 3.3875375883896273,
      "tokens_seen": 2330787840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987676767676768,
      "loss": 2.402,
      "theoretical_loss": 3.387529883767603,
      "tokens_seen": 2330853376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029874747474747476,
      "loss": 2.8542,
      "theoretical_loss": 3.3875221794228594,
      "tokens_seen": 2330918912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987272727272727,
      "loss": 2.5957,
      "theoretical_loss": 3.387514475355377,
      "tokens_seen": 2330984448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029870707070707073,
      "loss": 2.7468,
      "theoretical_loss": 3.38750677156514,
      "tokens_seen": 2331049984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986868686868687,
      "loss": 2.5161,
      "theoretical_loss": 3.387499068052129,
      "tokens_seen": 2331115520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986666666666667,
      "loss": 2.5359,
      "theoretical_loss": 3.3874913648163267,
      "tokens_seen": 2331181056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029864646464646466,
      "loss": 2.6,
      "theoretical_loss": 3.3874836618577158,
      "tokens_seen": 2331246592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986262626262626,
      "loss": 2.761,
      "theoretical_loss": 3.387475959176278,
      "tokens_seen": 2331312128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986060606060606,
      "loss": 2.4458,
      "theoretical_loss": 3.3874682567719963,
      "tokens_seen": 2331377664
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985858585858586,
      "loss": 2.5724,
      "theoretical_loss": 3.387460554644852,
      "tokens_seen": 2331443200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985656565656566,
      "loss": 2.5471,
      "theoretical_loss": 3.387452852794828,
      "tokens_seen": 2331508736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029854545454545455,
      "loss": 2.3454,
      "theoretical_loss": 3.387445151221906,
      "tokens_seen": 2331574272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029852525252525256,
      "loss": 2.6932,
      "theoretical_loss": 3.3874374499260687,
      "tokens_seen": 2331639808
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1286600,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6554930210113525,
      "objective/train/theoretical_loss": 3.387429748907298,
      "objective/train/tokens_used": 690564576,
      "theoretical_loss": 3.387429748907298,
      "tokens_seen": 2331705344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985050505050505,
      "loss": 2.5126,
      "theoretical_loss": 3.387429748907298,
      "tokens_seen": 2331705344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029848484848484847,
      "loss": 2.7253,
      "theoretical_loss": 3.3874220481655763,
      "tokens_seen": 2331770880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984646464646465,
      "loss": 2.4796,
      "theoretical_loss": 3.3874143477008865,
      "tokens_seen": 2331836416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029844444444444444,
      "loss": 2.6368,
      "theoretical_loss": 3.3874066475132096,
      "tokens_seen": 2331901952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029842424242424245,
      "loss": 2.6352,
      "theoretical_loss": 3.387398947602528,
      "tokens_seen": 2331967488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984040404040404,
      "loss": 2.6589,
      "theoretical_loss": 3.387391247968825,
      "tokens_seen": 2332033024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029838383838383837,
      "loss": 2.4235,
      "theoretical_loss": 3.3873835486120822,
      "tokens_seen": 2332098560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983636363636364,
      "loss": 2.5454,
      "theoretical_loss": 3.3873758495322823,
      "tokens_seen": 2332164096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983434343434344,
      "loss": 2.7885,
      "theoretical_loss": 3.3873681507294067,
      "tokens_seen": 2332229632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029832323232323234,
      "loss": 2.6011,
      "theoretical_loss": 3.387360452203438,
      "tokens_seen": 2332295168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983030303030303,
      "loss": 2.6292,
      "theoretical_loss": 3.387352753954359,
      "tokens_seen": 2332360704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029828282828282826,
      "loss": 2.7666,
      "theoretical_loss": 3.387345055982151,
      "tokens_seen": 2332426240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029826262626262627,
      "loss": 2.6206,
      "theoretical_loss": 3.387337358286797,
      "tokens_seen": 2332491776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982424242424243,
      "loss": 2.5381,
      "theoretical_loss": 3.3873296608682795,
      "tokens_seen": 2332557312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029822222222222224,
      "loss": 2.6186,
      "theoretical_loss": 3.3873219637265803,
      "tokens_seen": 2332622848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982020202020202,
      "loss": 2.4217,
      "theoretical_loss": 3.387314266861681,
      "tokens_seen": 2332688384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029818181818181815,
      "loss": 2.7029,
      "theoretical_loss": 3.3873065702735654,
      "tokens_seen": 2332753920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981616161616162,
      "loss": 2.662,
      "theoretical_loss": 3.387298873962215,
      "tokens_seen": 2332819456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981414141414142,
      "loss": 2.6625,
      "theoretical_loss": 3.3872911779276116,
      "tokens_seen": 2332884992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029812121212121213,
      "loss": 2.3996,
      "theoretical_loss": 3.387283482169738,
      "tokens_seen": 2332950528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981010101010101,
      "loss": 2.4591,
      "theoretical_loss": 3.3872757866885763,
      "tokens_seen": 2333016064
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029808080808080804,
      "loss": 2.5562,
      "theoretical_loss": 3.387268091484109,
      "tokens_seen": 2333081600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980606060606061,
      "loss": 2.3632,
      "theoretical_loss": 3.387260396556318,
      "tokens_seen": 2333147136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029804040404040407,
      "loss": 2.4928,
      "theoretical_loss": 3.3872527019051857,
      "tokens_seen": 2333212672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298020202020202,
      "loss": 2.6419,
      "theoretical_loss": 3.387245007530695,
      "tokens_seen": 2333278208
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1287008,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1645236015319824,
      "objective/train/theoretical_loss": 3.387237313432827,
      "objective/train/tokens_used": 692202976,
      "theoretical_loss": 3.387237313432827,
      "tokens_seen": 2333343744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298,
      "loss": 2.4941,
      "theoretical_loss": 3.387237313432827,
      "tokens_seen": 2333343744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029797979797979794,
      "loss": 2.8065,
      "theoretical_loss": 3.387229619611565,
      "tokens_seen": 2333409280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297959595959596,
      "loss": 2.7294,
      "theoretical_loss": 3.3872219260668914,
      "tokens_seen": 2333474816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029793939393939396,
      "loss": 2.4559,
      "theoretical_loss": 3.3872142327987875,
      "tokens_seen": 2333540352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979191919191919,
      "loss": 2.5859,
      "theoretical_loss": 3.3872065398072366,
      "tokens_seen": 2333605888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029789898989898987,
      "loss": 2.6923,
      "theoretical_loss": 3.38719884709222,
      "tokens_seen": 2333671424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978787878787879,
      "loss": 2.5763,
      "theoretical_loss": 3.387191154653721,
      "tokens_seen": 2333736960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978585858585859,
      "loss": 2.6255,
      "theoretical_loss": 3.387183462491721,
      "tokens_seen": 2333802496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029783838383838385,
      "loss": 2.5329,
      "theoretical_loss": 3.387175770606203,
      "tokens_seen": 2333868032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978181818181818,
      "loss": 2.6269,
      "theoretical_loss": 3.3871680789971492,
      "tokens_seen": 2333933568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029779797979797976,
      "loss": 2.6989,
      "theoretical_loss": 3.387160387664541,
      "tokens_seen": 2333999104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977777777777778,
      "loss": 2.8185,
      "theoretical_loss": 3.3871526966083625,
      "tokens_seen": 2334064640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977575757575758,
      "loss": 2.4832,
      "theoretical_loss": 3.387145005828594,
      "tokens_seen": 2334130176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029773737373737374,
      "loss": 2.6361,
      "theoretical_loss": 3.3871373153252193,
      "tokens_seen": 2334195712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977171717171717,
      "loss": 2.3801,
      "theoretical_loss": 3.3871296250982197,
      "tokens_seen": 2334261248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976969696969697,
      "loss": 2.6512,
      "theoretical_loss": 3.387121935147578,
      "tokens_seen": 2334326784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029767676767676767,
      "loss": 2.5558,
      "theoretical_loss": 3.3871142454732768,
      "tokens_seen": 2334392320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976565656565657,
      "loss": 2.4155,
      "theoretical_loss": 3.3871065560752975,
      "tokens_seen": 2334457856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029763636363636364,
      "loss": 2.6089,
      "theoretical_loss": 3.3870988669536235,
      "tokens_seen": 2334523392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976161616161616,
      "loss": 2.4402,
      "theoretical_loss": 3.3870911781082365,
      "tokens_seen": 2334588928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975959595959596,
      "loss": 2.5136,
      "theoretical_loss": 3.3870834895391186,
      "tokens_seen": 2334654464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029757575757575756,
      "loss": 2.5871,
      "theoretical_loss": 3.3870758012462527,
      "tokens_seen": 2334720000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029755555555555557,
      "loss": 2.5719,
      "theoretical_loss": 3.387068113229621,
      "tokens_seen": 2334785536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029753535353535353,
      "loss": 2.5404,
      "theoretical_loss": 3.387060425489205,
      "tokens_seen": 2334851072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029751515151515154,
      "loss": 2.2632,
      "theoretical_loss": 3.3870527380249884,
      "tokens_seen": 2334916608
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1288435,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5754096508026123,
      "objective/train/theoretical_loss": 3.3870450508369525,
      "objective/train/tokens_used": 693841376,
      "theoretical_loss": 3.3870450508369525,
      "tokens_seen": 2334982144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974949494949495,
      "loss": 2.5484,
      "theoretical_loss": 3.3870450508369525,
      "tokens_seen": 2334982144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029747474747474745,
      "loss": 2.3657,
      "theoretical_loss": 3.38703736392508,
      "tokens_seen": 2335047680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029745454545454546,
      "loss": 2.688,
      "theoretical_loss": 3.3870296772893536,
      "tokens_seen": 2335113216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974343434343434,
      "loss": 2.7214,
      "theoretical_loss": 3.3870219909297545,
      "tokens_seen": 2335178752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029741414141414143,
      "loss": 2.6295,
      "theoretical_loss": 3.387014304846266,
      "tokens_seen": 2335244288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973939393939394,
      "loss": 2.5466,
      "theoretical_loss": 3.3870066190388703,
      "tokens_seen": 2335309824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973737373737374,
      "loss": 2.4963,
      "theoretical_loss": 3.3869989335075497,
      "tokens_seen": 2335375360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029735353535353536,
      "loss": 2.5459,
      "theoretical_loss": 3.3869912482522864,
      "tokens_seen": 2335440896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029733333333333337,
      "loss": 2.6287,
      "theoretical_loss": 3.3869835632730627,
      "tokens_seen": 2335506432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973131313131313,
      "loss": 2.4409,
      "theoretical_loss": 3.386975878569861,
      "tokens_seen": 2335571968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972929292929293,
      "loss": 2.5834,
      "theoretical_loss": 3.3869681941426633,
      "tokens_seen": 2335637504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972727272727273,
      "loss": 2.6672,
      "theoretical_loss": 3.386960509991453,
      "tokens_seen": 2335703040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029725252525252525,
      "loss": 2.585,
      "theoretical_loss": 3.386952826116211,
      "tokens_seen": 2335768576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029723232323232326,
      "loss": 2.6306,
      "theoretical_loss": 3.386945142516921,
      "tokens_seen": 2335834112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972121212121212,
      "loss": 2.5344,
      "theoretical_loss": 3.3869374591935646,
      "tokens_seen": 2335899648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971919191919192,
      "loss": 2.5429,
      "theoretical_loss": 3.386929776146124,
      "tokens_seen": 2335965184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971717171717172,
      "loss": 2.5843,
      "theoretical_loss": 3.386922093374582,
      "tokens_seen": 2336030720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971515151515152,
      "loss": 2.568,
      "theoretical_loss": 3.386914410878921,
      "tokens_seen": 2336096256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029713131313131315,
      "loss": 2.637,
      "theoretical_loss": 3.386906728659123,
      "tokens_seen": 2336161792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971111111111111,
      "loss": 2.3562,
      "theoretical_loss": 3.3868990467151705,
      "tokens_seen": 2336227328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029709090909090907,
      "loss": 2.6356,
      "theoretical_loss": 3.3868913650470462,
      "tokens_seen": 2336292864
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970707070707071,
      "loss": 2.6018,
      "theoretical_loss": 3.386883683654732,
      "tokens_seen": 2336358400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970505050505051,
      "loss": 2.5028,
      "theoretical_loss": 3.3868760025382096,
      "tokens_seen": 2336423936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029703030303030305,
      "loss": 2.4544,
      "theoretical_loss": 3.386868321697463,
      "tokens_seen": 2336489472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297010101010101,
      "loss": 2.4403,
      "theoretical_loss": 3.3868606411324738,
      "tokens_seen": 2336555008
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1288948,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5773682594299316,
      "objective/train/theoretical_loss": 3.386852960843224,
      "objective/train/tokens_used": 695479776,
      "theoretical_loss": 3.386852960843224,
      "tokens_seen": 2336620544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029698989898989896,
      "loss": 2.4461,
      "theoretical_loss": 3.386852960843224,
      "tokens_seen": 2336620544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296969696969697,
      "loss": 2.4447,
      "theoretical_loss": 3.386845280829696,
      "tokens_seen": 2336686080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296949494949495,
      "loss": 2.6317,
      "theoretical_loss": 3.3868376010918726,
      "tokens_seen": 2336751616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029692929292929294,
      "loss": 2.6922,
      "theoretical_loss": 3.3868299216297357,
      "tokens_seen": 2336817152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969090909090909,
      "loss": 2.5182,
      "theoretical_loss": 3.3868222424432686,
      "tokens_seen": 2336882688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029688888888888885,
      "loss": 2.689,
      "theoretical_loss": 3.3868145635324525,
      "tokens_seen": 2336948224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968686868686869,
      "loss": 2.5282,
      "theoretical_loss": 3.3868068848972706,
      "tokens_seen": 2337013760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968484848484849,
      "loss": 2.5075,
      "theoretical_loss": 3.3867992065377046,
      "tokens_seen": 2337079296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029682828282828283,
      "loss": 2.7941,
      "theoretical_loss": 3.3867915284537378,
      "tokens_seen": 2337144832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968080808080808,
      "loss": 2.8759,
      "theoretical_loss": 3.386783850645352,
      "tokens_seen": 2337210368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967878787878788,
      "loss": 2.5257,
      "theoretical_loss": 3.386776173112529,
      "tokens_seen": 2337275904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967676767676768,
      "loss": 2.5613,
      "theoretical_loss": 3.386768495855252,
      "tokens_seen": 2337341440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029674747474747477,
      "loss": 2.5094,
      "theoretical_loss": 3.386760818873504,
      "tokens_seen": 2337406976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967272727272727,
      "loss": 2.6228,
      "theoretical_loss": 3.3867531421672656,
      "tokens_seen": 2337472512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967070707070707,
      "loss": 2.5084,
      "theoretical_loss": 3.386745465736521,
      "tokens_seen": 2337538048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966868686868687,
      "loss": 2.4286,
      "theoretical_loss": 3.3867377895812507,
      "tokens_seen": 2337603584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966666666666667,
      "loss": 2.5172,
      "theoretical_loss": 3.386730113701439,
      "tokens_seen": 2337669120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029664646464646466,
      "loss": 2.5309,
      "theoretical_loss": 3.386722438097067,
      "tokens_seen": 2337734656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966262626262626,
      "loss": 2.5358,
      "theoretical_loss": 3.3867147627681176,
      "tokens_seen": 2337800192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966060606060606,
      "loss": 2.5944,
      "theoretical_loss": 3.3867070877145733,
      "tokens_seen": 2337865728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965858585858586,
      "loss": 2.6264,
      "theoretical_loss": 3.3866994129364163,
      "tokens_seen": 2337931264
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965656565656566,
      "loss": 2.609,
      "theoretical_loss": 3.3866917384336293,
      "tokens_seen": 2337996800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029654545454545455,
      "loss": 2.7484,
      "theoretical_loss": 3.386684064206194,
      "tokens_seen": 2338062336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965252525252525,
      "loss": 2.5472,
      "theoretical_loss": 3.3866763902540935,
      "tokens_seen": 2338127872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965050505050505,
      "loss": 2.788,
      "theoretical_loss": 3.3866687165773097,
      "tokens_seen": 2338193408
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1290000,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6659116744995117,
      "objective/train/theoretical_loss": 3.3866610431758253,
      "objective/train/tokens_used": 697118176,
      "theoretical_loss": 3.3866610431758253,
      "tokens_seen": 2338258944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964848484848485,
      "loss": 2.5281,
      "theoretical_loss": 3.3866610431758253,
      "tokens_seen": 2338258944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964646464646465,
      "loss": 2.613,
      "theoretical_loss": 3.386653370049623,
      "tokens_seen": 2338324480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029644444444444445,
      "loss": 2.597,
      "theoretical_loss": 3.3866456971986842,
      "tokens_seen": 2338390016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964242424242424,
      "loss": 2.4275,
      "theoretical_loss": 3.3866380246229926,
      "tokens_seen": 2338455552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964040404040404,
      "loss": 2.4164,
      "theoretical_loss": 3.3866303523225296,
      "tokens_seen": 2338521088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029638383838383837,
      "loss": 2.5275,
      "theoretical_loss": 3.386622680297278,
      "tokens_seen": 2338586624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963636363636364,
      "loss": 2.5986,
      "theoretical_loss": 3.3866150085472206,
      "tokens_seen": 2338652160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029634343434343434,
      "loss": 2.5817,
      "theoretical_loss": 3.3866073370723395,
      "tokens_seen": 2338717696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029632323232323235,
      "loss": 2.5196,
      "theoretical_loss": 3.3865996658726165,
      "tokens_seen": 2338783232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963030303030303,
      "loss": 2.3676,
      "theoretical_loss": 3.3865919949480348,
      "tokens_seen": 2338848768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029628282828282826,
      "loss": 2.3948,
      "theoretical_loss": 3.3865843242985765,
      "tokens_seen": 2338914304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962626262626263,
      "loss": 2.5673,
      "theoretical_loss": 3.3865766539242244,
      "tokens_seen": 2338979840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029624242424242423,
      "loss": 2.3985,
      "theoretical_loss": 3.3865689838249606,
      "tokens_seen": 2339045376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029622222222222224,
      "loss": 2.7088,
      "theoretical_loss": 3.3865613140007675,
      "tokens_seen": 2339110912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962020202020202,
      "loss": 2.6563,
      "theoretical_loss": 3.3865536444516273,
      "tokens_seen": 2339176448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029618181818181816,
      "loss": 2.8682,
      "theoretical_loss": 3.386545975177523,
      "tokens_seen": 2339241984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029616161616161617,
      "loss": 2.385,
      "theoretical_loss": 3.386538306178437,
      "tokens_seen": 2339307520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961414141414142,
      "loss": 2.4802,
      "theoretical_loss": 3.386530637454351,
      "tokens_seen": 2339373056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029612121212121214,
      "loss": 2.764,
      "theoretical_loss": 3.3865229690052487,
      "tokens_seen": 2339438592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961010101010101,
      "loss": 2.6332,
      "theoretical_loss": 3.3865153008311113,
      "tokens_seen": 2339504128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029608080808080805,
      "loss": 2.3376,
      "theoretical_loss": 3.3865076329319215,
      "tokens_seen": 2339569664
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029606060606060606,
      "loss": 2.5313,
      "theoretical_loss": 3.386499965307662,
      "tokens_seen": 2339635200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029604040404040407,
      "loss": 2.5455,
      "theoretical_loss": 3.3864922979583154,
      "tokens_seen": 2339700736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029602020202020203,
      "loss": 2.4528,
      "theoretical_loss": 3.386484630883864,
      "tokens_seen": 2339766272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296,
      "loss": 2.5559,
      "theoretical_loss": 3.3864769640842898,
      "tokens_seen": 2339831808
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1290520,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1252095699310303,
      "objective/train/theoretical_loss": 3.386469297559576,
      "objective/train/tokens_used": 698756576,
      "theoretical_loss": 3.386469297559576,
      "tokens_seen": 2339897344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029597979797979794,
      "loss": 2.5348,
      "theoretical_loss": 3.386469297559576,
      "tokens_seen": 2339897344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000295959595959596,
      "loss": 2.4725,
      "theoretical_loss": 3.3864616313097047,
      "tokens_seen": 2339962880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029593939393939396,
      "loss": 2.5402,
      "theoretical_loss": 3.3864539653346584,
      "tokens_seen": 2340028416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002959191919191919,
      "loss": 2.617,
      "theoretical_loss": 3.386446299634419,
      "tokens_seen": 2340093952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958989898989899,
      "loss": 2.4474,
      "theoretical_loss": 3.38643863420897,
      "tokens_seen": 2340159488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029587878787878783,
      "loss": 2.614,
      "theoretical_loss": 3.386430969058293,
      "tokens_seen": 2340225024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958585858585859,
      "loss": 2.5714,
      "theoretical_loss": 3.3864233041823706,
      "tokens_seen": 2340290560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029583838383838386,
      "loss": 2.5186,
      "theoretical_loss": 3.3864156395811853,
      "tokens_seen": 2340356096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958181818181818,
      "loss": 2.6369,
      "theoretical_loss": 3.38640797525472,
      "tokens_seen": 2340421632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029579797979797977,
      "loss": 2.3658,
      "theoretical_loss": 3.386400311202957,
      "tokens_seen": 2340487168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957777777777778,
      "loss": 2.6491,
      "theoretical_loss": 3.386392647425878,
      "tokens_seen": 2340552704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957575757575758,
      "loss": 2.6185,
      "theoretical_loss": 3.3863849839234663,
      "tokens_seen": 2340618240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029573737373737375,
      "loss": 2.4718,
      "theoretical_loss": 3.386377320695704,
      "tokens_seen": 2340683776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957171717171717,
      "loss": 2.4129,
      "theoretical_loss": 3.386369657742574,
      "tokens_seen": 2340749312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029569696969696966,
      "loss": 2.5709,
      "theoretical_loss": 3.386361995064058,
      "tokens_seen": 2340814848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029567676767676773,
      "loss": 2.7302,
      "theoretical_loss": 3.3863543326601393,
      "tokens_seen": 2340880384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956565656565657,
      "loss": 2.8111,
      "theoretical_loss": 3.3863466705308,
      "tokens_seen": 2340945920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029563636363636364,
      "loss": 2.5205,
      "theoretical_loss": 3.386339008676022,
      "tokens_seen": 2341011456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956161616161616,
      "loss": 2.6271,
      "theoretical_loss": 3.386331347095789,
      "tokens_seen": 2341076992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955959595959596,
      "loss": 2.5173,
      "theoretical_loss": 3.3863236857900825,
      "tokens_seen": 2341142528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955757575757576,
      "loss": 2.766,
      "theoretical_loss": 3.3863160247588855,
      "tokens_seen": 2341208064
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955555555555556,
      "loss": 2.6761,
      "theoretical_loss": 3.38630836400218,
      "tokens_seen": 2341273600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029553535353535353,
      "loss": 2.6146,
      "theoretical_loss": 3.3863007035199493,
      "tokens_seen": 2341339136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955151515151515,
      "loss": 2.7249,
      "theoretical_loss": 3.386293043312175,
      "tokens_seen": 2341404672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002954949494949495,
      "loss": 2.5817,
      "theoretical_loss": 3.3862853833788398,
      "tokens_seen": 2341470208
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 1291680,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.936016082763672,
      "objective/train/theoretical_loss": 3.3862777237199264,
      "objective/train/tokens_used": 700394976,
      "theoretical_loss": 3.3862777237199264,
      "tokens_seen": 2341535744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002954747474747475,
      "loss": 2.749,
      "theoretical_loss": 3.3862777237199264,
      "tokens_seen": 2341535744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029545454545454547,
      "loss": 2.4722,
      "theoretical_loss": 3.386270064335417,
      "tokens_seen": 2341601280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029543434343434343,
      "loss": 2.6946,
      "theoretical_loss": 3.3862624052252945,
      "tokens_seen": 2341666816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029541414141414144,
      "loss": 2.4421,
      "theoretical_loss": 3.3862547463895414,
      "tokens_seen": 2341732352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953939393939394,
      "loss": 2.3572,
      "theoretical_loss": 3.38624708782814,
      "tokens_seen": 2341797888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953737373737374,
      "loss": 2.7433,
      "theoretical_loss": 3.3862394295410723,
      "tokens_seen": 2341863424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029535353535353536,
      "loss": 2.5928,
      "theoretical_loss": 3.3862317715283217,
      "tokens_seen": 2341928960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953333333333333,
      "loss": 2.4816,
      "theoretical_loss": 3.38622411378987,
      "tokens_seen": 2341994496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029531313131313133,
      "loss": 2.6014,
      "theoretical_loss": 3.3862164563257,
      "tokens_seen": 2342060032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952929292929293,
      "loss": 2.7071,
      "theoretical_loss": 3.3862087991357943,
      "tokens_seen": 2342125568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952727272727273,
      "loss": 2.5742,
      "theoretical_loss": 3.386201142220135,
      "tokens_seen": 2342191104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029525252525252526,
      "loss": 2.4848,
      "theoretical_loss": 3.3861934855787053,
      "tokens_seen": 2342256640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029523232323232327,
      "loss": 2.3532,
      "theoretical_loss": 3.386185829211487,
      "tokens_seen": 2342322176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952121212121212,
      "loss": 2.6362,
      "theoretical_loss": 3.386178173118463,
      "tokens_seen": 2342387712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951919191919192,
      "loss": 2.875,
      "theoretical_loss": 3.3861705172996155,
      "tokens_seen": 2342453248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951717171717172,
      "loss": 2.305,
      "theoretical_loss": 3.3861628617549275,
      "tokens_seen": 2342518784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029515151515151515,
      "loss": 2.5675,
      "theoretical_loss": 3.386155206484381,
      "tokens_seen": 2342584320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029513131313131316,
      "loss": 2.4903,
      "theoretical_loss": 3.3861475514879587,
      "tokens_seen": 2342649856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951111111111111,
      "loss": 2.7167,
      "theoretical_loss": 3.3861398967656435,
      "tokens_seen": 2342715392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950909090909091,
      "loss": 2.6,
      "theoretical_loss": 3.386132242317417,
      "tokens_seen": 2342780928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950707070707071,
      "loss": 2.5007,
      "theoretical_loss": 3.3861245881432627,
      "tokens_seen": 2342846464
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950505050505051,
      "loss": 2.6889,
      "theoretical_loss": 3.3861169342431627,
      "tokens_seen": 2342912000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029503030303030305,
      "loss": 2.5251,
      "theoretical_loss": 3.3861092806170996,
      "tokens_seen": 2342977536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000295010101010101,
      "loss": 2.6591,
      "theoretical_loss": 3.386101627265056,
      "tokens_seen": 2343043072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029498989898989897,
      "loss": 2.8258,
      "theoretical_loss": 3.3860939741870135,
      "tokens_seen": 2343108608
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1292123,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4943063259124756,
      "objective/train/theoretical_loss": 3.386086321382956,
      "objective/train/tokens_used": 702033376,
      "theoretical_loss": 3.386086321382956,
      "tokens_seen": 2343174144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294969696969697,
      "loss": 2.5336,
      "theoretical_loss": 3.386086321382956,
      "tokens_seen": 2343174144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294949494949495,
      "loss": 2.6416,
      "theoretical_loss": 3.3860786688528655,
      "tokens_seen": 2343239680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029492929292929295,
      "loss": 2.9048,
      "theoretical_loss": 3.3860710165967243,
      "tokens_seen": 2343305216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949090909090909,
      "loss": 2.4854,
      "theoretical_loss": 3.386063364614515,
      "tokens_seen": 2343370752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029488888888888886,
      "loss": 2.59,
      "theoretical_loss": 3.3860557129062205,
      "tokens_seen": 2343436288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948686868686869,
      "loss": 2.6525,
      "theoretical_loss": 3.386048061471823,
      "tokens_seen": 2343501824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948484848484849,
      "loss": 2.88,
      "theoretical_loss": 3.386040410311305,
      "tokens_seen": 2343567360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029482828282828284,
      "loss": 2.6506,
      "theoretical_loss": 3.386032759424649,
      "tokens_seen": 2343632896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948080808080808,
      "loss": 2.4685,
      "theoretical_loss": 3.386025108811838,
      "tokens_seen": 2343698432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029478787878787875,
      "loss": 2.5132,
      "theoretical_loss": 3.386017458472854,
      "tokens_seen": 2343763968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947676767676768,
      "loss": 2.7642,
      "theoretical_loss": 3.38600980840768,
      "tokens_seen": 2343829504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947474747474748,
      "loss": 2.7564,
      "theoretical_loss": 3.386002158616298,
      "tokens_seen": 2343895040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029472727272727273,
      "loss": 2.5789,
      "theoretical_loss": 3.385994509098691,
      "tokens_seen": 2343960576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947070707070707,
      "loss": 2.6892,
      "theoretical_loss": 3.3859868598548415,
      "tokens_seen": 2344026112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029468686868686864,
      "loss": 2.6213,
      "theoretical_loss": 3.3859792108847317,
      "tokens_seen": 2344091648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946666666666667,
      "loss": 2.693,
      "theoretical_loss": 3.3859715621883444,
      "tokens_seen": 2344157184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029464646464646467,
      "loss": 2.7018,
      "theoretical_loss": 3.3859639137656625,
      "tokens_seen": 2344222720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946262626262626,
      "loss": 2.5896,
      "theoretical_loss": 3.385956265616668,
      "tokens_seen": 2344288256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946060606060606,
      "loss": 2.4672,
      "theoretical_loss": 3.385948617741344,
      "tokens_seen": 2344353792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945858585858586,
      "loss": 2.7143,
      "theoretical_loss": 3.3859409701396723,
      "tokens_seen": 2344419328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945656565656566,
      "loss": 2.5054,
      "theoretical_loss": 3.3859333228116357,
      "tokens_seen": 2344484864
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029454545454545456,
      "loss": 2.6772,
      "theoretical_loss": 3.3859256757572176,
      "tokens_seen": 2344550400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945252525252525,
      "loss": 2.7907,
      "theoretical_loss": 3.3859180289763993,
      "tokens_seen": 2344615936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029450505050505047,
      "loss": 2.5474,
      "theoretical_loss": 3.3859103824691643,
      "tokens_seen": 2344681472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944848484848485,
      "loss": 2.6514,
      "theoretical_loss": 3.385902736235495,
      "tokens_seen": 2344747008
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1293031,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.416686534881592,
      "objective/train/theoretical_loss": 3.3858950902753735,
      "objective/train/tokens_used": 703671776,
      "theoretical_loss": 3.3858950902753735,
      "tokens_seen": 2344812544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944646464646465,
      "loss": 2.6185,
      "theoretical_loss": 3.3858950902753735,
      "tokens_seen": 2344812544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029444444444444445,
      "loss": 2.6763,
      "theoretical_loss": 3.3858874445887825,
      "tokens_seen": 2344878080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944242424242424,
      "loss": 2.4738,
      "theoretical_loss": 3.3858797991757053,
      "tokens_seen": 2344943616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944040404040404,
      "loss": 2.6323,
      "theoretical_loss": 3.3858721540361234,
      "tokens_seen": 2345009152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943838383838384,
      "loss": 2.6377,
      "theoretical_loss": 3.38586450917002,
      "tokens_seen": 2345074688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943636363636364,
      "loss": 2.4647,
      "theoretical_loss": 3.3858568645773777,
      "tokens_seen": 2345140224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029434343434343434,
      "loss": 2.5484,
      "theoretical_loss": 3.385849220258179,
      "tokens_seen": 2345205760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943232323232323,
      "loss": 2.6288,
      "theoretical_loss": 3.3858415762124063,
      "tokens_seen": 2345271296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943030303030303,
      "loss": 2.5396,
      "theoretical_loss": 3.3858339324400424,
      "tokens_seen": 2345336832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029428282828282827,
      "loss": 2.4864,
      "theoretical_loss": 3.3858262889410695,
      "tokens_seen": 2345402368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942626262626263,
      "loss": 2.5486,
      "theoretical_loss": 3.3858186457154704,
      "tokens_seen": 2345467904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029424242424242424,
      "loss": 2.6082,
      "theoretical_loss": 3.385811002763228,
      "tokens_seen": 2345533440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029422222222222225,
      "loss": 2.3866,
      "theoretical_loss": 3.3858033600843247,
      "tokens_seen": 2345598976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942020202020202,
      "loss": 2.429,
      "theoretical_loss": 3.385795717678743,
      "tokens_seen": 2345664512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029418181818181816,
      "loss": 2.7159,
      "theoretical_loss": 3.3857880755464658,
      "tokens_seen": 2345730048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941616161616162,
      "loss": 2.7147,
      "theoretical_loss": 3.385780433687475,
      "tokens_seen": 2345795584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029414141414141413,
      "loss": 2.5166,
      "theoretical_loss": 3.3857727921017533,
      "tokens_seen": 2345861120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029412121212121214,
      "loss": 2.3101,
      "theoretical_loss": 3.385765150789284,
      "tokens_seen": 2345926656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941010101010101,
      "loss": 2.6126,
      "theoretical_loss": 3.385757509750049,
      "tokens_seen": 2345992192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029408080808080805,
      "loss": 2.3296,
      "theoretical_loss": 3.3857498689840315,
      "tokens_seen": 2346057728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029406060606060607,
      "loss": 2.4155,
      "theoretical_loss": 3.385742228491214,
      "tokens_seen": 2346123264
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002940404040404041,
      "loss": 2.655,
      "theoretical_loss": 3.3857345882715784,
      "tokens_seen": 2346188800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029402020202020203,
      "loss": 2.5982,
      "theoretical_loss": 3.385726948325108,
      "tokens_seen": 2346254336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294,
      "loss": 2.7136,
      "theoretical_loss": 3.385719308651785,
      "tokens_seen": 2346319872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029397979797979795,
      "loss": 2.3272,
      "theoretical_loss": 3.3857116692515925,
      "tokens_seen": 2346385408
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1293973,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.691797971725464,
      "objective/train/theoretical_loss": 3.3857040301245123,
      "objective/train/tokens_used": 705310176,
      "theoretical_loss": 3.3857040301245123,
      "tokens_seen": 2346450944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029395959595959596,
      "loss": 2.5661,
      "theoretical_loss": 3.3857040301245123,
      "tokens_seen": 2346450944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029393939393939397,
      "loss": 2.5383,
      "theoretical_loss": 3.385696391270528,
      "tokens_seen": 2346516480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002939191919191919,
      "loss": 2.3794,
      "theoretical_loss": 3.3856887526896213,
      "tokens_seen": 2346582016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938989898989899,
      "loss": 2.6422,
      "theoretical_loss": 3.3856811143817755,
      "tokens_seen": 2346647552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938787878787879,
      "loss": 2.5329,
      "theoretical_loss": 3.3856734763469727,
      "tokens_seen": 2346713088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938585858585859,
      "loss": 2.5702,
      "theoretical_loss": 3.385665838585196,
      "tokens_seen": 2346778624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029383838383838386,
      "loss": 2.3762,
      "theoretical_loss": 3.385658201096428,
      "tokens_seen": 2346844160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938181818181818,
      "loss": 2.4848,
      "theoretical_loss": 3.3856505638806507,
      "tokens_seen": 2346909696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937979797979798,
      "loss": 2.4811,
      "theoretical_loss": 3.385642926937847,
      "tokens_seen": 2346975232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937777777777778,
      "loss": 2.5428,
      "theoretical_loss": 3.3856352902679996,
      "tokens_seen": 2347040768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937575757575758,
      "loss": 2.5246,
      "theoretical_loss": 3.385627653871092,
      "tokens_seen": 2347106304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029373737373737375,
      "loss": 2.4987,
      "theoretical_loss": 3.385620017747105,
      "tokens_seen": 2347171840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937171717171717,
      "loss": 2.4716,
      "theoretical_loss": 3.3856123818960224,
      "tokens_seen": 2347237376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029369696969696967,
      "loss": 2.6963,
      "theoretical_loss": 3.385604746317827,
      "tokens_seen": 2347302912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029367676767676773,
      "loss": 2.5826,
      "theoretical_loss": 3.3855971110125007,
      "tokens_seen": 2347368448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936565656565657,
      "loss": 2.5608,
      "theoretical_loss": 3.3855894759800265,
      "tokens_seen": 2347433984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029363636363636365,
      "loss": 2.7137,
      "theoretical_loss": 3.3855818412203873,
      "tokens_seen": 2347499520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936161616161616,
      "loss": 2.3285,
      "theoretical_loss": 3.3855742067335655,
      "tokens_seen": 2347565056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029359595959595956,
      "loss": 2.7452,
      "theoretical_loss": 3.3855665725195436,
      "tokens_seen": 2347630592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935757575757576,
      "loss": 2.5853,
      "theoretical_loss": 3.385558938578304,
      "tokens_seen": 2347696128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935555555555556,
      "loss": 2.5034,
      "theoretical_loss": 3.3855513049098303,
      "tokens_seen": 2347761664
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029353535353535354,
      "loss": 2.543,
      "theoretical_loss": 3.385543671514104,
      "tokens_seen": 2347827200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935151515151515,
      "loss": 2.3646,
      "theoretical_loss": 3.385536038391108,
      "tokens_seen": 2347892736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934949494949495,
      "loss": 2.4485,
      "theoretical_loss": 3.3855284055408257,
      "tokens_seen": 2347958272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934747474747475,
      "loss": 2.7168,
      "theoretical_loss": 3.3855207729632393,
      "tokens_seen": 2348023808
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1294468,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1802330017089844,
      "objective/train/theoretical_loss": 3.385513140658331,
      "objective/train/tokens_used": 706948576,
      "theoretical_loss": 3.385513140658331,
      "tokens_seen": 2348089344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934545454545455,
      "loss": 2.588,
      "theoretical_loss": 3.385513140658331,
      "tokens_seen": 2348089344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029343434343434343,
      "loss": 2.6582,
      "theoretical_loss": 3.385505508626084,
      "tokens_seen": 2348154880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934141414141414,
      "loss": 2.5386,
      "theoretical_loss": 3.385497876866481,
      "tokens_seen": 2348220416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933939393939394,
      "loss": 2.5273,
      "theoretical_loss": 3.385490245379504,
      "tokens_seen": 2348285952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933737373737374,
      "loss": 2.475,
      "theoretical_loss": 3.3854826141651366,
      "tokens_seen": 2348351488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029335353535353537,
      "loss": 2.6544,
      "theoretical_loss": 3.385474983223361,
      "tokens_seen": 2348417024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933333333333333,
      "loss": 2.6035,
      "theoretical_loss": 3.385467352554159,
      "tokens_seen": 2348482560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933131313131313,
      "loss": 2.782,
      "theoretical_loss": 3.3854597221575147,
      "tokens_seen": 2348548096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932929292929293,
      "loss": 2.4634,
      "theoretical_loss": 3.38545209203341,
      "tokens_seen": 2348613632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932727272727273,
      "loss": 2.6999,
      "theoretical_loss": 3.3854444621818276,
      "tokens_seen": 2348679168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029325252525252526,
      "loss": 2.557,
      "theoretical_loss": 3.38543683260275,
      "tokens_seen": 2348744704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932323232323232,
      "loss": 2.5582,
      "theoretical_loss": 3.3854292032961606,
      "tokens_seen": 2348810240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029321212121212123,
      "loss": 2.3718,
      "theoretical_loss": 3.3854215742620415,
      "tokens_seen": 2348875776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931919191919192,
      "loss": 2.6344,
      "theoretical_loss": 3.385413945500375,
      "tokens_seen": 2348941312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931717171717172,
      "loss": 2.8614,
      "theoretical_loss": 3.3854063170111446,
      "tokens_seen": 2349006848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029315151515151515,
      "loss": 2.6338,
      "theoretical_loss": 3.3853986887943326,
      "tokens_seen": 2349072384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931313131313131,
      "loss": 2.6859,
      "theoretical_loss": 3.3853910608499214,
      "tokens_seen": 2349137920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931111111111111,
      "loss": 2.6705,
      "theoretical_loss": 3.385383433177894,
      "tokens_seen": 2349203456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930909090909091,
      "loss": 2.5417,
      "theoretical_loss": 3.3853758057782333,
      "tokens_seen": 2349268992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930707070707071,
      "loss": 2.5322,
      "theoretical_loss": 3.3853681786509213,
      "tokens_seen": 2349334528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029305050505050505,
      "loss": 2.6303,
      "theoretical_loss": 3.385360551795941,
      "tokens_seen": 2349400064
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029303030303030306,
      "loss": 2.734,
      "theoretical_loss": 3.3853529252132755,
      "tokens_seen": 2349465600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000293010101010101,
      "loss": 2.5497,
      "theoretical_loss": 3.385345298902907,
      "tokens_seen": 2349531136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029298989898989897,
      "loss": 2.5689,
      "theoretical_loss": 3.385337672864818,
      "tokens_seen": 2349596672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292969696969697,
      "loss": 2.6474,
      "theoretical_loss": 3.385330047098992,
      "tokens_seen": 2349662208
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1295772,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4741108417510986,
      "objective/train/theoretical_loss": 3.3853224216054105,
      "objective/train/tokens_used": 708586976,
      "theoretical_loss": 3.3853224216054105,
      "tokens_seen": 2349727744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029294949494949494,
      "loss": 2.5149,
      "theoretical_loss": 3.3853224216054105,
      "tokens_seen": 2349727744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029292929292929295,
      "loss": 2.544,
      "theoretical_loss": 3.3853147963840575,
      "tokens_seen": 2349793280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929090909090909,
      "loss": 2.6175,
      "theoretical_loss": 3.3853071714349143,
      "tokens_seen": 2349858816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029288888888888886,
      "loss": 2.7578,
      "theoretical_loss": 3.385299546757965,
      "tokens_seen": 2349924352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928686868686869,
      "loss": 2.6745,
      "theoretical_loss": 3.3852919223531917,
      "tokens_seen": 2349989888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928484848484849,
      "loss": 2.3819,
      "theoretical_loss": 3.3852842982205766,
      "tokens_seen": 2350055424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029282828282828284,
      "loss": 2.3935,
      "theoretical_loss": 3.385276674360103,
      "tokens_seen": 2350120960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928080808080808,
      "loss": 2.5262,
      "theoretical_loss": 3.385269050771753,
      "tokens_seen": 2350186496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029278787878787876,
      "loss": 2.5877,
      "theoretical_loss": 3.3852614274555104,
      "tokens_seen": 2350252032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029276767676767677,
      "loss": 2.5712,
      "theoretical_loss": 3.3852538044113567,
      "tokens_seen": 2350317568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927474747474748,
      "loss": 2.8019,
      "theoretical_loss": 3.3852461816392756,
      "tokens_seen": 2350383104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029272727272727274,
      "loss": 2.6033,
      "theoretical_loss": 3.385238559139249,
      "tokens_seen": 2350448640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927070707070707,
      "loss": 2.5216,
      "theoretical_loss": 3.3852309369112596,
      "tokens_seen": 2350514176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029268686868686865,
      "loss": 2.5175,
      "theoretical_loss": 3.385223314955291,
      "tokens_seen": 2350579712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926666666666667,
      "loss": 2.4967,
      "theoretical_loss": 3.3852156932713253,
      "tokens_seen": 2350645248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029264646464646467,
      "loss": 2.5049,
      "theoretical_loss": 3.385208071859345,
      "tokens_seen": 2350710784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029262626262626263,
      "loss": 2.6128,
      "theoretical_loss": 3.385200450719333,
      "tokens_seen": 2350776320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926060606060606,
      "loss": 2.784,
      "theoretical_loss": 3.3851928298512726,
      "tokens_seen": 2350841856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029258585858585854,
      "loss": 2.6949,
      "theoretical_loss": 3.3851852092551455,
      "tokens_seen": 2350907392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925656565656566,
      "loss": 2.6908,
      "theoretical_loss": 3.3851775889309352,
      "tokens_seen": 2350972928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029254545454545456,
      "loss": 2.6029,
      "theoretical_loss": 3.385169968878624,
      "tokens_seen": 2351038464
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925252525252525,
      "loss": 2.6881,
      "theoretical_loss": 3.385162349098195,
      "tokens_seen": 2351104000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925050505050505,
      "loss": 2.5663,
      "theoretical_loss": 3.38515472958963,
      "tokens_seen": 2351169536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924848484848485,
      "loss": 2.5862,
      "theoretical_loss": 3.385147110352913,
      "tokens_seen": 2351235072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924646464646465,
      "loss": 2.7534,
      "theoretical_loss": 3.385139491388026,
      "tokens_seen": 2351300608
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1296037,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.399495840072632,
      "objective/train/theoretical_loss": 3.3851318726949513,
      "objective/train/tokens_used": 710225376,
      "theoretical_loss": 3.3851318726949513,
      "tokens_seen": 2351366144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029244444444444446,
      "loss": 2.6963,
      "theoretical_loss": 3.3851318726949513,
      "tokens_seen": 2351366144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924242424242424,
      "loss": 2.66,
      "theoretical_loss": 3.385124254273673,
      "tokens_seen": 2351431680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029240404040404037,
      "loss": 2.7219,
      "theoretical_loss": 3.3851166361241725,
      "tokens_seen": 2351497216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923838383838384,
      "loss": 2.7103,
      "theoretical_loss": 3.385109018246433,
      "tokens_seen": 2351562752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923636363636364,
      "loss": 2.6919,
      "theoretical_loss": 3.385101400640438,
      "tokens_seen": 2351628288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029234343434343435,
      "loss": 2.6424,
      "theoretical_loss": 3.3850937833061687,
      "tokens_seen": 2351693824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923232323232323,
      "loss": 2.7473,
      "theoretical_loss": 3.3850861662436085,
      "tokens_seen": 2351759360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923030303030303,
      "loss": 2.3818,
      "theoretical_loss": 3.3850785494527407,
      "tokens_seen": 2351824896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922828282828283,
      "loss": 2.4927,
      "theoretical_loss": 3.3850709329335475,
      "tokens_seen": 2351890432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922626262626263,
      "loss": 2.5495,
      "theoretical_loss": 3.385063316686012,
      "tokens_seen": 2351955968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029224242424242424,
      "loss": 2.6525,
      "theoretical_loss": 3.385055700710116,
      "tokens_seen": 2352021504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922222222222222,
      "loss": 2.6369,
      "theoretical_loss": 3.3850480850058435,
      "tokens_seen": 2352087040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922020202020202,
      "loss": 2.8024,
      "theoretical_loss": 3.3850404695731764,
      "tokens_seen": 2352152576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921818181818182,
      "loss": 2.5016,
      "theoretical_loss": 3.3850328544120982,
      "tokens_seen": 2352218112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921616161616162,
      "loss": 2.933,
      "theoretical_loss": 3.385025239522591,
      "tokens_seen": 2352283648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029214141414141414,
      "loss": 2.5866,
      "theoretical_loss": 3.385017624904637,
      "tokens_seen": 2352349184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029212121212121215,
      "loss": 2.5993,
      "theoretical_loss": 3.3850100105582204,
      "tokens_seen": 2352414720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921010101010101,
      "loss": 2.7311,
      "theoretical_loss": 3.3850023964833227,
      "tokens_seen": 2352480256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920808080808081,
      "loss": 2.6685,
      "theoretical_loss": 3.384994782679928,
      "tokens_seen": 2352545792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029206060606060607,
      "loss": 2.787,
      "theoretical_loss": 3.3849871691480176,
      "tokens_seen": 2352611328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029204040404040403,
      "loss": 2.5299,
      "theoretical_loss": 3.384979555887575,
      "tokens_seen": 2352676864
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029202020202020204,
      "loss": 2.7781,
      "theoretical_loss": 3.384971942898583,
      "tokens_seen": 2352742400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292,
      "loss": 2.4366,
      "theoretical_loss": 3.384964330181024,
      "tokens_seen": 2352807936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291979797979798,
      "loss": 2.5819,
      "theoretical_loss": 3.3849567177348807,
      "tokens_seen": 2352873472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029195959595959596,
      "loss": 2.4894,
      "theoretical_loss": 3.3849491055601364,
      "tokens_seen": 2352939008
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.863433837890625,
      "objective/train/theoretical_loss": 3.3849414936567737,
      "objective/train/tokens_used": 711863776,
      "theoretical_loss": 3.3849414936567737,
      "tokens_seen": 2353004544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291939393939394,
      "loss": 2.4362,
      "theoretical_loss": 3.3849414936567737,
      "tokens_seen": 2353004544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029191919191919193,
      "loss": 2.6936,
      "theoretical_loss": 3.3849338820247756,
      "tokens_seen": 2353070080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918989898989899,
      "loss": 2.624,
      "theoretical_loss": 3.384926270664124,
      "tokens_seen": 2353135616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918787878787879,
      "loss": 2.7503,
      "theoretical_loss": 3.3849186595748026,
      "tokens_seen": 2353201152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029185858585858586,
      "loss": 2.6667,
      "theoretical_loss": 3.3849110487567935,
      "tokens_seen": 2353266688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029183838383838387,
      "loss": 2.565,
      "theoretical_loss": 3.3849034382100798,
      "tokens_seen": 2353332224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918181818181818,
      "loss": 2.5301,
      "theoretical_loss": 3.384895827934644,
      "tokens_seen": 2353397760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917979797979798,
      "loss": 2.7628,
      "theoretical_loss": 3.3848882179304693,
      "tokens_seen": 2353463296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917777777777778,
      "loss": 2.7541,
      "theoretical_loss": 3.3848806081975384,
      "tokens_seen": 2353528832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917575757575758,
      "loss": 2.7078,
      "theoretical_loss": 3.384872998735834,
      "tokens_seen": 2353594368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029173737373737376,
      "loss": 2.6661,
      "theoretical_loss": 3.3848653895453387,
      "tokens_seen": 2353659904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917171717171717,
      "loss": 2.5376,
      "theoretical_loss": 3.3848577806260356,
      "tokens_seen": 2353725440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916969696969697,
      "loss": 2.6752,
      "theoretical_loss": 3.384850171977907,
      "tokens_seen": 2353790976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916767676767677,
      "loss": 2.4622,
      "theoretical_loss": 3.3848425636009365,
      "tokens_seen": 2353856512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916565656565657,
      "loss": 2.4389,
      "theoretical_loss": 3.384834955495106,
      "tokens_seen": 2353922048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029163636363636365,
      "loss": 2.777,
      "theoretical_loss": 3.3848273476603987,
      "tokens_seen": 2353987584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916161616161616,
      "loss": 2.8391,
      "theoretical_loss": 3.3848197400967974,
      "tokens_seen": 2354053120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029159595959595957,
      "loss": 2.8029,
      "theoretical_loss": 3.3848121328042846,
      "tokens_seen": 2354118656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029157575757575763,
      "loss": 2.8413,
      "theoretical_loss": 3.3848045257828434,
      "tokens_seen": 2354184192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915555555555556,
      "loss": 2.7559,
      "theoretical_loss": 3.384796919032457,
      "tokens_seen": 2354249728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029153535353535355,
      "loss": 2.6729,
      "theoretical_loss": 3.384789312553107,
      "tokens_seen": 2354315264
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915151515151515,
      "loss": 2.6606,
      "theoretical_loss": 3.3847817063447776,
      "tokens_seen": 2354380800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029149494949494946,
      "loss": 2.5145,
      "theoretical_loss": 3.3847741004074505,
      "tokens_seen": 2354446336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914747474747475,
      "loss": 2.667,
      "theoretical_loss": 3.3847664947411094,
      "tokens_seen": 2354511872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914545454545455,
      "loss": 2.6851,
      "theoretical_loss": 3.3847588893457363,
      "tokens_seen": 2354577408
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.831690549850464,
      "objective/train/theoretical_loss": 3.3847512842213145,
      "objective/train/tokens_used": 713502176,
      "theoretical_loss": 3.3847512842213145,
      "tokens_seen": 2354642944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029143434343434344,
      "loss": 2.9613,
      "theoretical_loss": 3.3847512842213145,
      "tokens_seen": 2354642944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914141414141414,
      "loss": 2.83,
      "theoretical_loss": 3.3847436793678263,
      "tokens_seen": 2354708480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029139393939393935,
      "loss": 2.8386,
      "theoretical_loss": 3.384736074785255,
      "tokens_seen": 2354774016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913737373737374,
      "loss": 2.7046,
      "theoretical_loss": 3.3847284704735836,
      "tokens_seen": 2354839552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913535353535354,
      "loss": 2.8229,
      "theoretical_loss": 3.384720866432794,
      "tokens_seen": 2354905088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029133333333333333,
      "loss": 2.6687,
      "theoretical_loss": 3.38471326266287,
      "tokens_seen": 2354970624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913131313131313,
      "loss": 2.781,
      "theoretical_loss": 3.384705659163794,
      "tokens_seen": 2355036160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912929292929293,
      "loss": 2.7317,
      "theoretical_loss": 3.384698055935549,
      "tokens_seen": 2355101696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912727272727273,
      "loss": 2.7084,
      "theoretical_loss": 3.384690452978117,
      "tokens_seen": 2355167232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029125252525252527,
      "loss": 2.6948,
      "theoretical_loss": 3.3846828502914823,
      "tokens_seen": 2355232768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912323232323232,
      "loss": 2.7676,
      "theoretical_loss": 3.3846752478756263,
      "tokens_seen": 2355298304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912121212121212,
      "loss": 2.6817,
      "theoretical_loss": 3.3846676457305325,
      "tokens_seen": 2355363840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911919191919192,
      "loss": 2.6406,
      "theoretical_loss": 3.384660043856184,
      "tokens_seen": 2355429376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911717171717172,
      "loss": 2.6755,
      "theoretical_loss": 3.3846524422525626,
      "tokens_seen": 2355494912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029115151515151516,
      "loss": 2.5943,
      "theoretical_loss": 3.384644840919652,
      "tokens_seen": 2355560448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911313131313131,
      "loss": 2.7349,
      "theoretical_loss": 3.384637239857435,
      "tokens_seen": 2355625984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029111111111111113,
      "loss": 2.8644,
      "theoretical_loss": 3.384629639065894,
      "tokens_seen": 2355691520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910909090909091,
      "loss": 2.6618,
      "theoretical_loss": 3.384622038545012,
      "tokens_seen": 2355757056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910707070707071,
      "loss": 2.7833,
      "theoretical_loss": 3.384614438294772,
      "tokens_seen": 2355822592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029105050505050505,
      "loss": 2.6828,
      "theoretical_loss": 3.384606838315157,
      "tokens_seen": 2355888128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291030303030303,
      "loss": 2.8611,
      "theoretical_loss": 3.384599238606149,
      "tokens_seen": 2355953664
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291010101010101,
      "loss": 2.6833,
      "theoretical_loss": 3.384591639167732,
      "tokens_seen": 2356019200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290989898989899,
      "loss": 2.7436,
      "theoretical_loss": 3.384584039999888,
      "tokens_seen": 2356084736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290969696969697,
      "loss": 2.6357,
      "theoretical_loss": 3.3845764411026,
      "tokens_seen": 2356150272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029094949494949495,
      "loss": 2.7971,
      "theoretical_loss": 3.384568842475851,
      "tokens_seen": 2356215808
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7116146087646484,
      "objective/train/theoretical_loss": 3.384561244119624,
      "objective/train/tokens_used": 715140576,
      "theoretical_loss": 3.384561244119624,
      "tokens_seen": 2356281344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029092929292929296,
      "loss": 2.6878,
      "theoretical_loss": 3.384561244119624,
      "tokens_seen": 2356281344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909090909090909,
      "loss": 2.7204,
      "theoretical_loss": 3.3845536460339014,
      "tokens_seen": 2356346880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029088888888888887,
      "loss": 2.9062,
      "theoretical_loss": 3.384546048218666,
      "tokens_seen": 2356412416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908686868686869,
      "loss": 2.7743,
      "theoretical_loss": 3.384538450673901,
      "tokens_seen": 2356477952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029084848484848484,
      "loss": 2.8154,
      "theoretical_loss": 3.3845308533995895,
      "tokens_seen": 2356543488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029082828282828285,
      "loss": 2.828,
      "theoretical_loss": 3.3845232563957137,
      "tokens_seen": 2356609024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908080808080808,
      "loss": 2.8674,
      "theoretical_loss": 3.3845156596622568,
      "tokens_seen": 2356674560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029078787878787876,
      "loss": 2.9225,
      "theoretical_loss": 3.3845080631992017,
      "tokens_seen": 2356740096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907676767676768,
      "loss": 2.7783,
      "theoretical_loss": 3.3845004670065313,
      "tokens_seen": 2356805632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907474747474748,
      "loss": 2.7414,
      "theoretical_loss": 3.384492871084228,
      "tokens_seen": 2356871168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029072727272727274,
      "loss": 2.7643,
      "theoretical_loss": 3.3844852754322754,
      "tokens_seen": 2356936704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907070707070707,
      "loss": 2.8731,
      "theoretical_loss": 3.3844776800506557,
      "tokens_seen": 2357002240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029068686868686866,
      "loss": 2.4969,
      "theoretical_loss": 3.384470084939352,
      "tokens_seen": 2357067776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029066666666666667,
      "loss": 2.6736,
      "theoretical_loss": 3.3844624900983473,
      "tokens_seen": 2357133312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906464646464647,
      "loss": 2.802,
      "theoretical_loss": 3.3844548955276244,
      "tokens_seen": 2357198848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029062626262626263,
      "loss": 2.6056,
      "theoretical_loss": 3.3844473012271656,
      "tokens_seen": 2357264384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906060606060606,
      "loss": 2.7787,
      "theoretical_loss": 3.3844397071969547,
      "tokens_seen": 2357329920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029058585858585855,
      "loss": 2.8631,
      "theoretical_loss": 3.3844321134369744,
      "tokens_seen": 2357395456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905656565656566,
      "loss": 2.7115,
      "theoretical_loss": 3.384424519947207,
      "tokens_seen": 2357460992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029054545454545457,
      "loss": 2.7696,
      "theoretical_loss": 3.384416926727636,
      "tokens_seen": 2357526528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029052525252525253,
      "loss": 2.5302,
      "theoretical_loss": 3.3844093337782435,
      "tokens_seen": 2357592064
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905050505050505,
      "loss": 2.7376,
      "theoretical_loss": 3.384401741099013,
      "tokens_seen": 2357657600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904848484848485,
      "loss": 2.8724,
      "theoretical_loss": 3.3843941486899274,
      "tokens_seen": 2357723136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904646464646465,
      "loss": 2.9169,
      "theoretical_loss": 3.384386556550969,
      "tokens_seen": 2357788672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029044444444444446,
      "loss": 2.7314,
      "theoretical_loss": 3.3843789646821216,
      "tokens_seen": 2357854208
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0510120391845703,
      "objective/train/theoretical_loss": 3.384371373083367,
      "objective/train/tokens_used": 716778976,
      "theoretical_loss": 3.384371373083367,
      "tokens_seen": 2357919744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904242424242424,
      "loss": 2.6395,
      "theoretical_loss": 3.384371373083367,
      "tokens_seen": 2357919744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002904040404040404,
      "loss": 2.6134,
      "theoretical_loss": 3.3843637817546894,
      "tokens_seen": 2357985280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029038383838383844,
      "loss": 2.5383,
      "theoretical_loss": 3.3843561906960704,
      "tokens_seen": 2358050816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903636363636364,
      "loss": 2.6797,
      "theoretical_loss": 3.3843485999074936,
      "tokens_seen": 2358116352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029034343434343436,
      "loss": 2.7628,
      "theoretical_loss": 3.3843410093889417,
      "tokens_seen": 2358181888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903232323232323,
      "loss": 2.7176,
      "theoretical_loss": 3.3843334191403978,
      "tokens_seen": 2358247424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029030303030303027,
      "loss": 2.4503,
      "theoretical_loss": 3.384325829161844,
      "tokens_seen": 2358312960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029028282828282833,
      "loss": 2.7259,
      "theoretical_loss": 3.3843182394532643,
      "tokens_seen": 2358378496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902626262626263,
      "loss": 2.8343,
      "theoretical_loss": 3.384310650014641,
      "tokens_seen": 2358444032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029024242424242425,
      "loss": 3.0389,
      "theoretical_loss": 3.3843030608459572,
      "tokens_seen": 2358509568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902222222222222,
      "loss": 2.7251,
      "theoretical_loss": 3.3842954719471954,
      "tokens_seen": 2358575104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902020202020202,
      "loss": 2.8515,
      "theoretical_loss": 3.384287883318339,
      "tokens_seen": 2358640640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029018181818181823,
      "loss": 2.7634,
      "theoretical_loss": 3.3842802949593707,
      "tokens_seen": 2358706176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901616161616162,
      "loss": 2.7505,
      "theoretical_loss": 3.384272706870273,
      "tokens_seen": 2358771712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029014141414141414,
      "loss": 3.0544,
      "theoretical_loss": 3.3842651190510296,
      "tokens_seen": 2358837248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901212121212121,
      "loss": 2.9072,
      "theoretical_loss": 3.384257531501623,
      "tokens_seen": 2358902784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901010101010101,
      "loss": 2.648,
      "theoretical_loss": 3.384249944222036,
      "tokens_seen": 2358968320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900808080808081,
      "loss": 2.7514,
      "theoretical_loss": 3.3842423572122513,
      "tokens_seen": 2359033856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900606060606061,
      "loss": 2.8266,
      "theoretical_loss": 3.384234770472253,
      "tokens_seen": 2359099392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029004040404040403,
      "loss": 2.9063,
      "theoretical_loss": 3.384227184002022,
      "tokens_seen": 2359164928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000290020202020202,
      "loss": 3.0149,
      "theoretical_loss": 3.3842195978015432,
      "tokens_seen": 2359230464
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029,
      "loss": 2.9477,
      "theoretical_loss": 3.3842120118707983,
      "tokens_seen": 2359296000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289979797979798,
      "loss": 2.8882,
      "theoretical_loss": 3.384204426209771,
      "tokens_seen": 2359361536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028995959595959597,
      "loss": 2.7126,
      "theoretical_loss": 3.3841968408184435,
      "tokens_seen": 2359427072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899393939393939,
      "loss": 2.9074,
      "theoretical_loss": 3.384189255696799,
      "tokens_seen": 2359492608
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9561614990234375,
      "objective/train/theoretical_loss": 3.3841816708448205,
      "objective/train/tokens_used": 718417376,
      "theoretical_loss": 3.3841816708448205,
      "tokens_seen": 2359558144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028991919191919194,
      "loss": 2.8289,
      "theoretical_loss": 3.3841816708448205,
      "tokens_seen": 2359558144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898989898989899,
      "loss": 2.7987,
      "theoretical_loss": 3.3841740862624907,
      "tokens_seen": 2359623680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898787878787879,
      "loss": 2.9507,
      "theoretical_loss": 3.384166501949793,
      "tokens_seen": 2359689216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028985858585858586,
      "loss": 2.8454,
      "theoretical_loss": 3.38415891790671,
      "tokens_seen": 2359754752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898383838383838,
      "loss": 2.9303,
      "theoretical_loss": 3.3841513341332243,
      "tokens_seen": 2359820288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028981818181818183,
      "loss": 2.6474,
      "theoretical_loss": 3.3841437506293195,
      "tokens_seen": 2359885824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897979797979798,
      "loss": 2.8907,
      "theoretical_loss": 3.384136167394978,
      "tokens_seen": 2359951360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897777777777778,
      "loss": 2.8254,
      "theoretical_loss": 3.384128584430183,
      "tokens_seen": 2360016896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028975757575757576,
      "loss": 2.7634,
      "theoretical_loss": 3.3841210017349175,
      "tokens_seen": 2360082432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028973737373737377,
      "loss": 2.5879,
      "theoretical_loss": 3.3841134193091644,
      "tokens_seen": 2360147968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897171717171717,
      "loss": 2.8118,
      "theoretical_loss": 3.3841058371529065,
      "tokens_seen": 2360213504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896969696969697,
      "loss": 2.7495,
      "theoretical_loss": 3.3840982552661263,
      "tokens_seen": 2360279040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896767676767677,
      "loss": 2.8714,
      "theoretical_loss": 3.384090673648808,
      "tokens_seen": 2360344576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028965656565656565,
      "loss": 2.8073,
      "theoretical_loss": 3.3840830923009335,
      "tokens_seen": 2360410112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028963636363636366,
      "loss": 2.7274,
      "theoretical_loss": 3.384075511222486,
      "tokens_seen": 2360475648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896161616161616,
      "loss": 2.7599,
      "theoretical_loss": 3.384067930413448,
      "tokens_seen": 2360541184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028959595959595957,
      "loss": 2.9951,
      "theoretical_loss": 3.3840603498738036,
      "tokens_seen": 2360606720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895757575757576,
      "loss": 2.8374,
      "theoretical_loss": 3.3840527696035343,
      "tokens_seen": 2360672256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895555555555556,
      "loss": 2.7211,
      "theoretical_loss": 3.3840451896026242,
      "tokens_seen": 2360737792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028953535353535355,
      "loss": 2.9183,
      "theoretical_loss": 3.384037609871056,
      "tokens_seen": 2360803328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895151515151515,
      "loss": 2.7325,
      "theoretical_loss": 3.3840300304088125,
      "tokens_seen": 2360868864
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028949494949494947,
      "loss": 2.7306,
      "theoretical_loss": 3.3840224512158765,
      "tokens_seen": 2360934400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894747474747475,
      "loss": 2.7594,
      "theoretical_loss": 3.3840148722922314,
      "tokens_seen": 2360999936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894545454545455,
      "loss": 2.7942,
      "theoretical_loss": 3.3840072936378593,
      "tokens_seen": 2361065472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028943434343434344,
      "loss": 2.8832,
      "theoretical_loss": 3.3839997152527443,
      "tokens_seen": 2361131008
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9772825241088867,
      "objective/train/theoretical_loss": 3.3839921371368686,
      "objective/train/tokens_used": 720055776,
      "theoretical_loss": 3.3839921371368686,
      "tokens_seen": 2361196544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894141414141414,
      "loss": 2.8571,
      "theoretical_loss": 3.3839921371368686,
      "tokens_seen": 2361196544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028939393939393936,
      "loss": 2.9166,
      "theoretical_loss": 3.3839845592902154,
      "tokens_seen": 2361262080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893737373737374,
      "loss": 2.814,
      "theoretical_loss": 3.3839769817127676,
      "tokens_seen": 2361327616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893535353535354,
      "loss": 2.7937,
      "theoretical_loss": 3.383969404404508,
      "tokens_seen": 2361393152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028933333333333334,
      "loss": 2.8665,
      "theoretical_loss": 3.38396182736542,
      "tokens_seen": 2361458688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893131313131313,
      "loss": 2.7818,
      "theoretical_loss": 3.3839542505954863,
      "tokens_seen": 2361524224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028929292929292925,
      "loss": 2.5287,
      "theoretical_loss": 3.38394667409469,
      "tokens_seen": 2361589760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892727272727273,
      "loss": 2.6725,
      "theoretical_loss": 3.3839390978630135,
      "tokens_seen": 2361655296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028925252525252527,
      "loss": 2.7712,
      "theoretical_loss": 3.3839315219004407,
      "tokens_seen": 2361720832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028923232323232323,
      "loss": 2.8722,
      "theoretical_loss": 3.383923946206954,
      "tokens_seen": 2361786368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892121212121212,
      "loss": 2.8434,
      "theoretical_loss": 3.3839163707825364,
      "tokens_seen": 2361851904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891919191919192,
      "loss": 2.8021,
      "theoretical_loss": 3.383908795627171,
      "tokens_seen": 2361917440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891717171717172,
      "loss": 2.8349,
      "theoretical_loss": 3.383901220740841,
      "tokens_seen": 2361982976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028915151515151517,
      "loss": 2.7123,
      "theoretical_loss": 3.383893646123529,
      "tokens_seen": 2362048512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891313131313131,
      "loss": 2.9454,
      "theoretical_loss": 3.383886071775218,
      "tokens_seen": 2362114048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891111111111111,
      "loss": 2.8406,
      "theoretical_loss": 3.383878497695891,
      "tokens_seen": 2362179584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890909090909091,
      "loss": 2.7326,
      "theoretical_loss": 3.383870923885531,
      "tokens_seen": 2362245120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890707070707071,
      "loss": 2.854,
      "theoretical_loss": 3.3838633503441216,
      "tokens_seen": 2362310656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028905050505050506,
      "loss": 2.9163,
      "theoretical_loss": 3.383855777071645,
      "tokens_seen": 2362376192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289030303030303,
      "loss": 2.7203,
      "theoretical_loss": 3.383848204068084,
      "tokens_seen": 2362441728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289010101010101,
      "loss": 2.844,
      "theoretical_loss": 3.3838406313334226,
      "tokens_seen": 2362507264
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288989898989899,
      "loss": 2.8456,
      "theoretical_loss": 3.383833058867643,
      "tokens_seen": 2362572800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288969696969697,
      "loss": 2.7197,
      "theoretical_loss": 3.3838254866707285,
      "tokens_seen": 2362638336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028894949494949495,
      "loss": 2.9643,
      "theoretical_loss": 3.3838179147426617,
      "tokens_seen": 2362703872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889292929292929,
      "loss": 2.722,
      "theoretical_loss": 3.383810343083426,
      "tokens_seen": 2362769408
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1296887,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9292616844177246,
      "objective/train/theoretical_loss": 3.383802771693005,
      "objective/train/tokens_used": 721694176,
      "theoretical_loss": 3.383802771693005,
      "tokens_seen": 2362834944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889090909090909,
      "loss": 2.9646,
      "theoretical_loss": 3.383802771693005,
      "tokens_seen": 2362834944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888888888888889,
      "loss": 2.825,
      "theoretical_loss": 3.3837952005713805,
      "tokens_seen": 2362900480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888686868686869,
      "loss": 2.8537,
      "theoretical_loss": 3.383787629718536,
      "tokens_seen": 2362966016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028884848484848484,
      "loss": 3.0183,
      "theoretical_loss": 3.3837800591344545,
      "tokens_seen": 2363031552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028882828282828285,
      "loss": 2.6299,
      "theoretical_loss": 3.383772488819119,
      "tokens_seen": 2363097088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888080808080808,
      "loss": 2.886,
      "theoretical_loss": 3.3837649187725125,
      "tokens_seen": 2363162624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887878787878788,
      "loss": 2.7028,
      "theoretical_loss": 3.383757348994618,
      "tokens_seen": 2363228160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887676767676768,
      "loss": 2.8194,
      "theoretical_loss": 3.383749779485419,
      "tokens_seen": 2363293696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028874747474747474,
      "loss": 2.7882,
      "theoretical_loss": 3.3837422102448977,
      "tokens_seen": 2363359232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028872727272727275,
      "loss": 2.8096,
      "theoretical_loss": 3.383734641273038,
      "tokens_seen": 2363424768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887070707070707,
      "loss": 2.756,
      "theoretical_loss": 3.3837270725698216,
      "tokens_seen": 2363490304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886868686868687,
      "loss": 2.6432,
      "theoretical_loss": 3.383719504135233,
      "tokens_seen": 2363555840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028866666666666667,
      "loss": 2.9209,
      "theoretical_loss": 3.383711935969254,
      "tokens_seen": 2363621376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886464646464647,
      "loss": 2.6537,
      "theoretical_loss": 3.383704368071869,
      "tokens_seen": 2363686912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028862626262626264,
      "loss": 2.9342,
      "theoretical_loss": 3.383696800443059,
      "tokens_seen": 2363752448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886060606060606,
      "loss": 2.9508,
      "theoretical_loss": 3.383689233082809,
      "tokens_seen": 2363817984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885858585858586,
      "loss": 2.9502,
      "theoretical_loss": 3.3836816659911007,
      "tokens_seen": 2363883520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028856565656565656,
      "loss": 2.7681,
      "theoretical_loss": 3.383674099167918,
      "tokens_seen": 2363949056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885454545454546,
      "loss": 2.8636,
      "theoretical_loss": 3.3836665326132436,
      "tokens_seen": 2364014592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028852525252525253,
      "loss": 2.8493,
      "theoretical_loss": 3.38365896632706,
      "tokens_seen": 2364080128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885050505050505,
      "loss": 2.6714,
      "theoretical_loss": 3.3836514003093514,
      "tokens_seen": 2364145664
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884848484848485,
      "loss": 2.8766,
      "theoretical_loss": 3.3836438345601,
      "tokens_seen": 2364211200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884646464646465,
      "loss": 2.7063,
      "theoretical_loss": 3.3836362690792887,
      "tokens_seen": 2364276736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028844444444444447,
      "loss": 3.0039,
      "theoretical_loss": 3.383628703866901,
      "tokens_seen": 2364342272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884242424242424,
      "loss": 2.6418,
      "theoretical_loss": 3.38362113892292,
      "tokens_seen": 2364407808
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1297770,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.729642629623413,
      "objective/train/theoretical_loss": 3.3836135742473283,
      "objective/train/tokens_used": 723332576,
      "theoretical_loss": 3.3836135742473283,
      "tokens_seen": 2364473344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884040404040404,
      "loss": 2.6027,
      "theoretical_loss": 3.3836135742473283,
      "tokens_seen": 2364473344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883838383838384,
      "loss": 2.8451,
      "theoretical_loss": 3.3836060098401086,
      "tokens_seen": 2364538880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883636363636364,
      "loss": 2.7172,
      "theoretical_loss": 3.3835984457012453,
      "tokens_seen": 2364604416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028834343434343436,
      "loss": 2.9995,
      "theoretical_loss": 3.3835908818307203,
      "tokens_seen": 2364669952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883232323232323,
      "loss": 2.7271,
      "theoretical_loss": 3.383583318228517,
      "tokens_seen": 2364735488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883030303030303,
      "loss": 2.701,
      "theoretical_loss": 3.3835757548946184,
      "tokens_seen": 2364801024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028828282828282834,
      "loss": 2.955,
      "theoretical_loss": 3.3835681918290077,
      "tokens_seen": 2364866560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882626262626263,
      "loss": 2.8081,
      "theoretical_loss": 3.3835606290316678,
      "tokens_seen": 2364932096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028824242424242425,
      "loss": 2.6179,
      "theoretical_loss": 3.3835530665025813,
      "tokens_seen": 2364997632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882222222222222,
      "loss": 2.6764,
      "theoretical_loss": 3.383545504241732,
      "tokens_seen": 2365063168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028820202020202017,
      "loss": 2.9193,
      "theoretical_loss": 3.3835379422491028,
      "tokens_seen": 2365128704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028818181818181823,
      "loss": 2.6585,
      "theoretical_loss": 3.3835303805246766,
      "tokens_seen": 2365194240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881616161616162,
      "loss": 2.7959,
      "theoretical_loss": 3.3835228190684363,
      "tokens_seen": 2365259776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028814141414141415,
      "loss": 2.9671,
      "theoretical_loss": 3.383515257880365,
      "tokens_seen": 2365325312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881212121212121,
      "loss": 2.6223,
      "theoretical_loss": 3.3835076969604465,
      "tokens_seen": 2365390848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028810101010101006,
      "loss": 2.7847,
      "theoretical_loss": 3.3835001363086628,
      "tokens_seen": 2365456384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002880808080808081,
      "loss": 2.6494,
      "theoretical_loss": 3.383492575924998,
      "tokens_seen": 2365521920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002880606060606061,
      "loss": 2.6848,
      "theoretical_loss": 3.383485015809434,
      "tokens_seen": 2365587456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028804040404040404,
      "loss": 3.0481,
      "theoretical_loss": 3.3834774559619545,
      "tokens_seen": 2365652992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288020202020202,
      "loss": 2.7109,
      "theoretical_loss": 3.3834698963825427,
      "tokens_seen": 2365718528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288,
      "loss": 2.9154,
      "theoretical_loss": 3.3834623370711814,
      "tokens_seen": 2365784064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287979797979798,
      "loss": 2.8094,
      "theoretical_loss": 3.3834547780278537,
      "tokens_seen": 2365849600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287959595959596,
      "loss": 2.8296,
      "theoretical_loss": 3.3834472192525427,
      "tokens_seen": 2365915136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028793939393939393,
      "loss": 2.6589,
      "theoretical_loss": 3.3834396607452315,
      "tokens_seen": 2365980672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879191919191919,
      "loss": 2.7813,
      "theoretical_loss": 3.3834321025059033,
      "tokens_seen": 2366046208
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1298991,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9180285930633545,
      "objective/train/theoretical_loss": 3.383424544534541,
      "objective/train/tokens_used": 724970976,
      "theoretical_loss": 3.383424544534541,
      "tokens_seen": 2366111744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878989898989899,
      "loss": 2.6105,
      "theoretical_loss": 3.383424544534541,
      "tokens_seen": 2366111744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878787878787879,
      "loss": 2.8354,
      "theoretical_loss": 3.383416986831128,
      "tokens_seen": 2366177280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028785858585858587,
      "loss": 2.6593,
      "theoretical_loss": 3.383409429395647,
      "tokens_seen": 2366242816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878383838383838,
      "loss": 2.6004,
      "theoretical_loss": 3.383401872228081,
      "tokens_seen": 2366308352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028781818181818184,
      "loss": 2.8588,
      "theoretical_loss": 3.383394315328413,
      "tokens_seen": 2366373888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877979797979798,
      "loss": 2.8283,
      "theoretical_loss": 3.3833867586966266,
      "tokens_seen": 2366439424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877777777777778,
      "loss": 2.8935,
      "theoretical_loss": 3.383379202332705,
      "tokens_seen": 2366504960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028775757575757576,
      "loss": 2.8037,
      "theoretical_loss": 3.383371646236631,
      "tokens_seen": 2366570496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877373737373737,
      "loss": 2.7439,
      "theoretical_loss": 3.383364090408387,
      "tokens_seen": 2366636032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028771717171717173,
      "loss": 2.7442,
      "theoretical_loss": 3.383356534847957,
      "tokens_seen": 2366701568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876969696969697,
      "loss": 2.7006,
      "theoretical_loss": 3.3833489795553238,
      "tokens_seen": 2366767104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876767676767677,
      "loss": 2.7843,
      "theoretical_loss": 3.3833414245304705,
      "tokens_seen": 2366832640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028765656565656565,
      "loss": 2.7358,
      "theoretical_loss": 3.38333386977338,
      "tokens_seen": 2366898176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028763636363636366,
      "loss": 2.8552,
      "theoretical_loss": 3.3833263152840356,
      "tokens_seen": 2366963712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876161616161616,
      "loss": 2.833,
      "theoretical_loss": 3.3833187610624207,
      "tokens_seen": 2367029248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875959595959596,
      "loss": 2.8562,
      "theoretical_loss": 3.383311207108518,
      "tokens_seen": 2367094784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875757575757576,
      "loss": 2.6936,
      "theoretical_loss": 3.3833036534223107,
      "tokens_seen": 2367160320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028755555555555555,
      "loss": 2.8159,
      "theoretical_loss": 3.383296100003782,
      "tokens_seen": 2367225856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028753535353535356,
      "loss": 2.8299,
      "theoretical_loss": 3.3832885468529144,
      "tokens_seen": 2367291392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875151515151515,
      "loss": 2.669,
      "theoretical_loss": 3.383280993969692,
      "tokens_seen": 2367356928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028749494949494947,
      "loss": 2.7604,
      "theoretical_loss": 3.3832734413540972,
      "tokens_seen": 2367422464
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874747474747475,
      "loss": 2.6439,
      "theoretical_loss": 3.3832658890061134,
      "tokens_seen": 2367488000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874545454545455,
      "loss": 2.8217,
      "theoretical_loss": 3.3832583369257234,
      "tokens_seen": 2367553536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028743434343434345,
      "loss": 2.6788,
      "theoretical_loss": 3.383250785112911,
      "tokens_seen": 2367619072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874141414141414,
      "loss": 2.7312,
      "theoretical_loss": 3.3832432335676583,
      "tokens_seen": 2367684608
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1299407,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.548238515853882,
      "objective/train/theoretical_loss": 3.383235682289949,
      "objective/train/tokens_used": 726609376,
      "theoretical_loss": 3.383235682289949,
      "tokens_seen": 2367750144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028739393939393936,
      "loss": 2.7809,
      "theoretical_loss": 3.383235682289949,
      "tokens_seen": 2367750144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873737373737374,
      "loss": 2.8353,
      "theoretical_loss": 3.3832281312797665,
      "tokens_seen": 2367815680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873535353535354,
      "loss": 2.7917,
      "theoretical_loss": 3.3832205805370936,
      "tokens_seen": 2367881216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028733333333333334,
      "loss": 2.59,
      "theoretical_loss": 3.3832130300619134,
      "tokens_seen": 2367946752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873131313131313,
      "loss": 2.7401,
      "theoretical_loss": 3.383205479854209,
      "tokens_seen": 2368012288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028729292929292926,
      "loss": 2.7757,
      "theoretical_loss": 3.3831979299139636,
      "tokens_seen": 2368077824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872727272727273,
      "loss": 2.6085,
      "theoretical_loss": 3.3831903802411603,
      "tokens_seen": 2368143360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872525252525253,
      "loss": 2.8826,
      "theoretical_loss": 3.3831828308357816,
      "tokens_seen": 2368208896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028723232323232324,
      "loss": 2.6337,
      "theoretical_loss": 3.3831752816978122,
      "tokens_seen": 2368274432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872121212121212,
      "loss": 2.7368,
      "theoretical_loss": 3.383167732827234,
      "tokens_seen": 2368339968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028719191919191915,
      "loss": 2.7714,
      "theoretical_loss": 3.38316018422403,
      "tokens_seen": 2368405504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871717171717172,
      "loss": 2.6164,
      "theoretical_loss": 3.383152635888184,
      "tokens_seen": 2368471040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028715151515151517,
      "loss": 2.5985,
      "theoretical_loss": 3.3831450878196785,
      "tokens_seen": 2368536576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028713131313131313,
      "loss": 2.9362,
      "theoretical_loss": 3.3831375400184975,
      "tokens_seen": 2368602112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871111111111111,
      "loss": 2.7855,
      "theoretical_loss": 3.383129992484623,
      "tokens_seen": 2368667648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870909090909091,
      "loss": 2.7867,
      "theoretical_loss": 3.3831224452180395,
      "tokens_seen": 2368733184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870707070707071,
      "loss": 2.6093,
      "theoretical_loss": 3.383114898218729,
      "tokens_seen": 2368798720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028705050505050506,
      "loss": 2.6574,
      "theoretical_loss": 3.383107351486675,
      "tokens_seen": 2368864256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287030303030303,
      "loss": 2.7298,
      "theoretical_loss": 3.3830998050218613,
      "tokens_seen": 2368929792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287010101010101,
      "loss": 2.7609,
      "theoretical_loss": 3.38309225882427,
      "tokens_seen": 2368995328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028698989898989904,
      "loss": 2.935,
      "theoretical_loss": 3.3830847128938846,
      "tokens_seen": 2369060864
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286969696969697,
      "loss": 2.5602,
      "theoretical_loss": 3.3830771672306885,
      "tokens_seen": 2369126400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028694949494949496,
      "loss": 2.8036,
      "theoretical_loss": 3.3830696218346645,
      "tokens_seen": 2369191936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869292929292929,
      "loss": 2.8502,
      "theoretical_loss": 3.383062076705796,
      "tokens_seen": 2369257472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869090909090909,
      "loss": 2.8216,
      "theoretical_loss": 3.383054531844066,
      "tokens_seen": 2369323008
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1300592,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9446256160736084,
      "objective/train/theoretical_loss": 3.383046987249458,
      "objective/train/tokens_used": 728247776,
      "theoretical_loss": 3.383046987249458,
      "tokens_seen": 2369388544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028688888888888894,
      "loss": 2.8198,
      "theoretical_loss": 3.383046987249458,
      "tokens_seen": 2369388544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868686868686869,
      "loss": 2.6907,
      "theoretical_loss": 3.383039442921955,
      "tokens_seen": 2369454080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028684848484848485,
      "loss": 2.8761,
      "theoretical_loss": 3.3830318988615398,
      "tokens_seen": 2369519616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868282828282828,
      "loss": 2.7659,
      "theoretical_loss": 3.3830243550681955,
      "tokens_seen": 2369585152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868080808080808,
      "loss": 2.8049,
      "theoretical_loss": 3.383016811541906,
      "tokens_seen": 2369650688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028678787878787883,
      "loss": 2.8566,
      "theoretical_loss": 3.383009268282654,
      "tokens_seen": 2369716224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867676767676768,
      "loss": 2.7565,
      "theoretical_loss": 3.3830017252904225,
      "tokens_seen": 2369781760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028674747474747474,
      "loss": 2.7113,
      "theoretical_loss": 3.382994182565195,
      "tokens_seen": 2369847296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867272727272727,
      "loss": 2.7538,
      "theoretical_loss": 3.382986640106954,
      "tokens_seen": 2369912832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867070707070707,
      "loss": 2.7925,
      "theoretical_loss": 3.3829790979156837,
      "tokens_seen": 2369978368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866868686868687,
      "loss": 2.7005,
      "theoretical_loss": 3.3829715559913667,
      "tokens_seen": 2370043904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866666666666667,
      "loss": 2.7673,
      "theoretical_loss": 3.3829640143339863,
      "tokens_seen": 2370109440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028664646464646463,
      "loss": 2.6366,
      "theoretical_loss": 3.3829564729435253,
      "tokens_seen": 2370174976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028662626262626265,
      "loss": 2.6965,
      "theoretical_loss": 3.3829489318199673,
      "tokens_seen": 2370240512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866060606060606,
      "loss": 2.7347,
      "theoretical_loss": 3.3829413909632953,
      "tokens_seen": 2370306048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865858585858586,
      "loss": 2.6849,
      "theoretical_loss": 3.382933850373493,
      "tokens_seen": 2370371584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028656565656565657,
      "loss": 2.6003,
      "theoretical_loss": 3.3829263100505425,
      "tokens_seen": 2370437120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028654545454545453,
      "loss": 2.6349,
      "theoretical_loss": 3.3829187699944274,
      "tokens_seen": 2370502656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028652525252525254,
      "loss": 2.6287,
      "theoretical_loss": 3.3829112302051314,
      "tokens_seen": 2370568192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865050505050505,
      "loss": 2.8806,
      "theoretical_loss": 3.3829036906826375,
      "tokens_seen": 2370633728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864848484848485,
      "loss": 2.7613,
      "theoretical_loss": 3.3828961514269285,
      "tokens_seen": 2370699264
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028646464646464646,
      "loss": 2.7364,
      "theoretical_loss": 3.3828886124379878,
      "tokens_seen": 2370764800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864444444444445,
      "loss": 2.653,
      "theoretical_loss": 3.3828810737157986,
      "tokens_seen": 2370830336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028642424242424243,
      "loss": 2.7884,
      "theoretical_loss": 3.3828735352603445,
      "tokens_seen": 2370895872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864040404040404,
      "loss": 2.7444,
      "theoretical_loss": 3.3828659970716077,
      "tokens_seen": 2370961408
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1301610,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.708845376968384,
      "objective/train/theoretical_loss": 3.3828584591495723,
      "objective/train/tokens_used": 729886176,
      "theoretical_loss": 3.3828584591495723,
      "tokens_seen": 2371026944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863838383838384,
      "loss": 2.6055,
      "theoretical_loss": 3.3828584591495723,
      "tokens_seen": 2371026944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028636363636363636,
      "loss": 2.6848,
      "theoretical_loss": 3.382850921494221,
      "tokens_seen": 2371092480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028634343434343437,
      "loss": 2.9651,
      "theoretical_loss": 3.382843384105537,
      "tokens_seen": 2371158016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863232323232323,
      "loss": 2.7766,
      "theoretical_loss": 3.382835846983504,
      "tokens_seen": 2371223552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863030303030303,
      "loss": 2.647,
      "theoretical_loss": 3.3828283101281045,
      "tokens_seen": 2371289088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862828282828283,
      "loss": 2.7359,
      "theoretical_loss": 3.382820773539322,
      "tokens_seen": 2371354624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862626262626263,
      "loss": 2.6404,
      "theoretical_loss": 3.38281323721714,
      "tokens_seen": 2371420160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028624242424242426,
      "loss": 2.7411,
      "theoretical_loss": 3.382805701161541,
      "tokens_seen": 2371485696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862222222222222,
      "loss": 2.5558,
      "theoretical_loss": 3.382798165372509,
      "tokens_seen": 2371551232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862020202020202,
      "loss": 2.7604,
      "theoretical_loss": 3.382790629850027,
      "tokens_seen": 2371616768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861818181818182,
      "loss": 2.8462,
      "theoretical_loss": 3.382783094594078,
      "tokens_seen": 2371682304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861616161616162,
      "loss": 2.6763,
      "theoretical_loss": 3.382775559604645,
      "tokens_seen": 2371747840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028614141414141415,
      "loss": 2.7553,
      "theoretical_loss": 3.382768024881712,
      "tokens_seen": 2371813376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861212121212121,
      "loss": 2.623,
      "theoretical_loss": 3.382760490425261,
      "tokens_seen": 2371878912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028610101010101007,
      "loss": 2.6844,
      "theoretical_loss": 3.3827529562352763,
      "tokens_seen": 2371944448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028608080808080813,
      "loss": 2.804,
      "theoretical_loss": 3.3827454223117406,
      "tokens_seen": 2372009984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860606060606061,
      "loss": 2.7884,
      "theoretical_loss": 3.382737888654637,
      "tokens_seen": 2372075520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028604040404040404,
      "loss": 2.6557,
      "theoretical_loss": 3.3827303552639494,
      "tokens_seen": 2372141056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286020202020202,
      "loss": 2.851,
      "theoretical_loss": 3.38272282213966,
      "tokens_seen": 2372206592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028599999999999996,
      "loss": 2.8257,
      "theoretical_loss": 3.3827152892817534,
      "tokens_seen": 2372272128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285979797979798,
      "loss": 2.8114,
      "theoretical_loss": 3.3827077566902113,
      "tokens_seen": 2372337664
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285959595959596,
      "loss": 2.6481,
      "theoretical_loss": 3.382700224365018,
      "tokens_seen": 2372403200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028593939393939394,
      "loss": 2.6321,
      "theoretical_loss": 3.382692692306156,
      "tokens_seen": 2372468736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859191919191919,
      "loss": 2.8074,
      "theoretical_loss": 3.382685160513609,
      "tokens_seen": 2372534272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858989898989899,
      "loss": 2.7955,
      "theoretical_loss": 3.38267762898736,
      "tokens_seen": 2372599808
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1302157,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.228212356567383,
      "objective/train/theoretical_loss": 3.382670097727393,
      "objective/train/tokens_used": 731524576,
      "theoretical_loss": 3.382670097727393,
      "tokens_seen": 2372665344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858787878787879,
      "loss": 2.5333,
      "theoretical_loss": 3.382670097727393,
      "tokens_seen": 2372665344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858585858585859,
      "loss": 2.7062,
      "theoretical_loss": 3.38266256673369,
      "tokens_seen": 2372730880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028583838383838383,
      "loss": 2.8846,
      "theoretical_loss": 3.3826550360062346,
      "tokens_seen": 2372796416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858181818181818,
      "loss": 2.8929,
      "theoretical_loss": 3.3826475055450107,
      "tokens_seen": 2372861952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857979797979798,
      "loss": 2.6605,
      "theoretical_loss": 3.382639975350001,
      "tokens_seen": 2372927488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857777777777778,
      "loss": 2.7888,
      "theoretical_loss": 3.3826324454211885,
      "tokens_seen": 2372993024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028575757575757577,
      "loss": 2.6401,
      "theoretical_loss": 3.382624915758557,
      "tokens_seen": 2373058560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857373737373737,
      "loss": 2.5738,
      "theoretical_loss": 3.3826173863620896,
      "tokens_seen": 2373124096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028571717171717173,
      "loss": 2.5644,
      "theoretical_loss": 3.382609857231769,
      "tokens_seen": 2373189632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856969696969697,
      "loss": 2.9077,
      "theoretical_loss": 3.3826023283675792,
      "tokens_seen": 2373255168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856767676767677,
      "loss": 2.3755,
      "theoretical_loss": 3.382594799769503,
      "tokens_seen": 2373320704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028565656565656566,
      "loss": 2.7103,
      "theoretical_loss": 3.382587271437524,
      "tokens_seen": 2373386240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856363636363636,
      "loss": 2.6764,
      "theoretical_loss": 3.382579743371625,
      "tokens_seen": 2373451776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856161616161616,
      "loss": 2.6704,
      "theoretical_loss": 3.3825722155717894,
      "tokens_seen": 2373517312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855959595959596,
      "loss": 2.4917,
      "theoretical_loss": 3.3825646880380003,
      "tokens_seen": 2373582848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855757575757576,
      "loss": 2.8492,
      "theoretical_loss": 3.3825571607702414,
      "tokens_seen": 2373648384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028555555555555555,
      "loss": 2.8155,
      "theoretical_loss": 3.3825496337684955,
      "tokens_seen": 2373713920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028553535353535356,
      "loss": 2.6375,
      "theoretical_loss": 3.3825421070327466,
      "tokens_seen": 2373779456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855151515151515,
      "loss": 2.6984,
      "theoretical_loss": 3.382534580562977,
      "tokens_seen": 2373844992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854949494949495,
      "loss": 2.7302,
      "theoretical_loss": 3.3825270543591706,
      "tokens_seen": 2373910528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854747474747475,
      "loss": 2.6061,
      "theoretical_loss": 3.3825195284213105,
      "tokens_seen": 2373976064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028545454545454544,
      "loss": 2.6342,
      "theoretical_loss": 3.3825120027493796,
      "tokens_seen": 2374041600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028543434343434346,
      "loss": 2.5622,
      "theoretical_loss": 3.3825044773433617,
      "tokens_seen": 2374107136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854141414141414,
      "loss": 2.8012,
      "theoretical_loss": 3.3824969522032395,
      "tokens_seen": 2374172672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028539393939393937,
      "loss": 2.8441,
      "theoretical_loss": 3.3824894273289967,
      "tokens_seen": 2374238208
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 1303451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7277791500091553,
      "objective/train/theoretical_loss": 3.3824819027206168,
      "objective/train/tokens_used": 733162976,
      "theoretical_loss": 3.3824819027206168,
      "tokens_seen": 2374303744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002853737373737374,
      "loss": 2.7293,
      "theoretical_loss": 3.3824819027206168,
      "tokens_seen": 2374303744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853535353535354,
      "loss": 2.7882,
      "theoretical_loss": 3.3824743783780824,
      "tokens_seen": 2374369280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028533333333333335,
      "loss": 2.5895,
      "theoretical_loss": 3.382466854301377,
      "tokens_seen": 2374434816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853131313131313,
      "loss": 2.6554,
      "theoretical_loss": 3.3824593304904846,
      "tokens_seen": 2374500352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852929292929293,
      "loss": 2.5905,
      "theoretical_loss": 3.382451806945387,
      "tokens_seen": 2374565888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028527272727272727,
      "loss": 2.8454,
      "theoretical_loss": 3.382444283666069,
      "tokens_seen": 2374631424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852525252525253,
      "loss": 2.6722,
      "theoretical_loss": 3.3824367606525128,
      "tokens_seen": 2374696960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028523232323232324,
      "loss": 2.6536,
      "theoretical_loss": 3.3824292379047023,
      "tokens_seen": 2374762496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852121212121212,
      "loss": 2.7028,
      "theoretical_loss": 3.3824217154226206,
      "tokens_seen": 2374828032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851919191919192,
      "loss": 2.7209,
      "theoretical_loss": 3.382414193206251,
      "tokens_seen": 2374893568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851717171717172,
      "loss": 2.5259,
      "theoretical_loss": 3.382406671255576,
      "tokens_seen": 2374959104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851515151515152,
      "loss": 2.5073,
      "theoretical_loss": 3.3823991495705803,
      "tokens_seen": 2375024640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028513131313131313,
      "loss": 2.6183,
      "theoretical_loss": 3.3823916281512463,
      "tokens_seen": 2375090176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851111111111111,
      "loss": 2.9363,
      "theoretical_loss": 3.3823841069975575,
      "tokens_seen": 2375155712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850909090909091,
      "loss": 2.5514,
      "theoretical_loss": 3.382376586109497,
      "tokens_seen": 2375221248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850707070707071,
      "loss": 2.7707,
      "theoretical_loss": 3.3823690654870484,
      "tokens_seen": 2375286784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028505050505050507,
      "loss": 2.6421,
      "theoretical_loss": 3.3823615451301947,
      "tokens_seen": 2375352320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285030303030303,
      "loss": 2.7776,
      "theoretical_loss": 3.3823540250389197,
      "tokens_seen": 2375417856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285010101010101,
      "loss": 2.6807,
      "theoretical_loss": 3.382346505213206,
      "tokens_seen": 2375483392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028498989898989905,
      "loss": 2.7964,
      "theoretical_loss": 3.3823389856530373,
      "tokens_seen": 2375548928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284969696969697,
      "loss": 2.7716,
      "theoretical_loss": 3.382331466358397,
      "tokens_seen": 2375614464
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028494949494949496,
      "loss": 2.5232,
      "theoretical_loss": 3.382323947329268,
      "tokens_seen": 2375680000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849292929292929,
      "loss": 2.6095,
      "theoretical_loss": 3.382316428565634,
      "tokens_seen": 2375745536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849090909090909,
      "loss": 2.7417,
      "theoretical_loss": 3.382308910067478,
      "tokens_seen": 2375811072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028488888888888894,
      "loss": 2.5701,
      "theoretical_loss": 3.3823013918347833,
      "tokens_seen": 2375876608
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1303936,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6463329792022705,
      "objective/train/theoretical_loss": 3.3822938738675337,
      "objective/train/tokens_used": 734801376,
      "theoretical_loss": 3.3822938738675337,
      "tokens_seen": 2375942144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848686868686869,
      "loss": 2.6626,
      "theoretical_loss": 3.3822938738675337,
      "tokens_seen": 2375942144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028484848484848485,
      "loss": 2.6632,
      "theoretical_loss": 3.3822863561657117,
      "tokens_seen": 2376007680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848282828282828,
      "loss": 2.6169,
      "theoretical_loss": 3.3822788387293015,
      "tokens_seen": 2376073216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028480808080808077,
      "loss": 2.8378,
      "theoretical_loss": 3.382271321558286,
      "tokens_seen": 2376138752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028478787878787883,
      "loss": 2.7283,
      "theoretical_loss": 3.382263804652648,
      "tokens_seen": 2376204288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847676767676768,
      "loss": 2.7262,
      "theoretical_loss": 3.3822562880123717,
      "tokens_seen": 2376269824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028474747474747475,
      "loss": 2.6373,
      "theoretical_loss": 3.3822487716374394,
      "tokens_seen": 2376335360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847272727272727,
      "loss": 2.4912,
      "theoretical_loss": 3.382241255527836,
      "tokens_seen": 2376400896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847070707070707,
      "loss": 2.743,
      "theoretical_loss": 3.3822337396835427,
      "tokens_seen": 2376466432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846868686868687,
      "loss": 2.855,
      "theoretical_loss": 3.3822262241045444,
      "tokens_seen": 2376531968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846666666666667,
      "loss": 2.7858,
      "theoretical_loss": 3.382218708790824,
      "tokens_seen": 2376597504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028464646464646464,
      "loss": 2.4187,
      "theoretical_loss": 3.382211193742365,
      "tokens_seen": 2376663040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846262626262626,
      "loss": 2.8033,
      "theoretical_loss": 3.3822036789591503,
      "tokens_seen": 2376728576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846060606060606,
      "loss": 2.76,
      "theoretical_loss": 3.3821961644411633,
      "tokens_seen": 2376794112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845858585858586,
      "loss": 2.7136,
      "theoretical_loss": 3.3821886501883878,
      "tokens_seen": 2376859648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845656565656566,
      "loss": 2.6643,
      "theoretical_loss": 3.3821811362008067,
      "tokens_seen": 2376925184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028454545454545453,
      "loss": 2.569,
      "theoretical_loss": 3.3821736224784034,
      "tokens_seen": 2376990720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028452525252525254,
      "loss": 2.8341,
      "theoretical_loss": 3.382166109021161,
      "tokens_seen": 2377056256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845050505050505,
      "loss": 2.7549,
      "theoretical_loss": 3.382158595829064,
      "tokens_seen": 2377121792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844848484848485,
      "loss": 2.6837,
      "theoretical_loss": 3.3821510829020935,
      "tokens_seen": 2377187328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028446464646464647,
      "loss": 2.6413,
      "theoretical_loss": 3.3821435702402347,
      "tokens_seen": 2377252864
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844444444444444,
      "loss": 2.5526,
      "theoretical_loss": 3.3821360578434705,
      "tokens_seen": 2377318400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028442424242424244,
      "loss": 2.4521,
      "theoretical_loss": 3.3821285457117845,
      "tokens_seen": 2377383936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844040404040404,
      "loss": 2.4732,
      "theoretical_loss": 3.3821210338451593,
      "tokens_seen": 2377449472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843838383838384,
      "loss": 2.6388,
      "theoretical_loss": 3.3821135222435785,
      "tokens_seen": 2377515008
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1305174,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.373270273208618,
      "objective/train/theoretical_loss": 3.382106010907026,
      "objective/train/tokens_used": 736439776,
      "theoretical_loss": 3.382106010907026,
      "tokens_seen": 2377580544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028436363636363636,
      "loss": 2.7005,
      "theoretical_loss": 3.382106010907026,
      "tokens_seen": 2377580544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028434343434343437,
      "loss": 2.7245,
      "theoretical_loss": 3.3820984998354837,
      "tokens_seen": 2377646080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028432323232323233,
      "loss": 2.7895,
      "theoretical_loss": 3.382090989028937,
      "tokens_seen": 2377711616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843030303030303,
      "loss": 2.8304,
      "theoretical_loss": 3.3820834784873677,
      "tokens_seen": 2377777152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842828282828283,
      "loss": 2.7219,
      "theoretical_loss": 3.3820759682107595,
      "tokens_seen": 2377842688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028426262626262625,
      "loss": 2.6387,
      "theoretical_loss": 3.382068458199096,
      "tokens_seen": 2377908224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028424242424242427,
      "loss": 2.7945,
      "theoretical_loss": 3.3820609484523603,
      "tokens_seen": 2377973760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842222222222222,
      "loss": 2.6654,
      "theoretical_loss": 3.3820534389705363,
      "tokens_seen": 2378039296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842020202020202,
      "loss": 2.7555,
      "theoretical_loss": 3.3820459297536067,
      "tokens_seen": 2378104832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841818181818182,
      "loss": 2.726,
      "theoretical_loss": 3.382038420801555,
      "tokens_seen": 2378170368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841616161616162,
      "loss": 2.8324,
      "theoretical_loss": 3.3820309121143652,
      "tokens_seen": 2378235904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028414141414141416,
      "loss": 2.5481,
      "theoretical_loss": 3.3820234036920196,
      "tokens_seen": 2378301440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841212121212121,
      "loss": 2.657,
      "theoretical_loss": 3.3820158955345025,
      "tokens_seen": 2378366976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028410101010101007,
      "loss": 2.848,
      "theoretical_loss": 3.3820083876417963,
      "tokens_seen": 2378432512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840808080808081,
      "loss": 2.5993,
      "theoretical_loss": 3.3820008800138854,
      "tokens_seen": 2378498048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840606060606061,
      "loss": 2.7084,
      "theoretical_loss": 3.3819933726507525,
      "tokens_seen": 2378563584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028404040404040405,
      "loss": 2.4528,
      "theoretical_loss": 3.381985865552381,
      "tokens_seen": 2378629120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284020202020202,
      "loss": 2.5714,
      "theoretical_loss": 3.3819783587187544,
      "tokens_seen": 2378694656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028399999999999996,
      "loss": 2.7692,
      "theoretical_loss": 3.3819708521498564,
      "tokens_seen": 2378760192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028397979797979803,
      "loss": 2.4897,
      "theoretical_loss": 3.3819633458456697,
      "tokens_seen": 2378825728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283959595959596,
      "loss": 2.7357,
      "theoretical_loss": 3.381955839806178,
      "tokens_seen": 2378891264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028393939393939394,
      "loss": 2.7141,
      "theoretical_loss": 3.3819483340313647,
      "tokens_seen": 2378956800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002839191919191919,
      "loss": 2.6596,
      "theoretical_loss": 3.3819408285212136,
      "tokens_seen": 2379022336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028389898989898986,
      "loss": 2.6599,
      "theoretical_loss": 3.3819333232757076,
      "tokens_seen": 2379087872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838787878787879,
      "loss": 2.6759,
      "theoretical_loss": 3.3819258182948295,
      "tokens_seen": 2379153408
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1305886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3820316791534424,
      "objective/train/theoretical_loss": 3.3819183135785638,
      "objective/train/tokens_used": 738078176,
      "theoretical_loss": 3.3819183135785638,
      "tokens_seen": 2379218944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838585858585859,
      "loss": 2.289,
      "theoretical_loss": 3.3819183135785638,
      "tokens_seen": 2379218944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028383838383838384,
      "loss": 2.8343,
      "theoretical_loss": 3.3819108091268935,
      "tokens_seen": 2379284480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838181818181818,
      "loss": 2.7488,
      "theoretical_loss": 3.3819033049398017,
      "tokens_seen": 2379350016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837979797979798,
      "loss": 2.7085,
      "theoretical_loss": 3.3818958010172717,
      "tokens_seen": 2379415552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837777777777778,
      "loss": 2.7585,
      "theoretical_loss": 3.3818882973592874,
      "tokens_seen": 2379481088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028375757575757577,
      "loss": 2.7206,
      "theoretical_loss": 3.381880793965832,
      "tokens_seen": 2379546624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028373737373737373,
      "loss": 2.7653,
      "theoretical_loss": 3.381873290836889,
      "tokens_seen": 2379612160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837171717171717,
      "loss": 2.6582,
      "theoretical_loss": 3.3818657879724414,
      "tokens_seen": 2379677696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836969696969697,
      "loss": 2.658,
      "theoretical_loss": 3.3818582853724726,
      "tokens_seen": 2379743232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836767676767677,
      "loss": 2.6887,
      "theoretical_loss": 3.3818507830369664,
      "tokens_seen": 2379808768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028365656565656566,
      "loss": 2.5464,
      "theoretical_loss": 3.381843280965906,
      "tokens_seen": 2379874304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836363636363636,
      "loss": 2.6371,
      "theoretical_loss": 3.3818357791592746,
      "tokens_seen": 2379939840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028361616161616163,
      "loss": 2.5925,
      "theoretical_loss": 3.381828277617056,
      "tokens_seen": 2380005376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028359595959595964,
      "loss": 2.7838,
      "theoretical_loss": 3.381820776339233,
      "tokens_seen": 2380070912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835757575757576,
      "loss": 2.6511,
      "theoretical_loss": 3.38181327532579,
      "tokens_seen": 2380136448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028355555555555556,
      "loss": 2.6926,
      "theoretical_loss": 3.38180577457671,
      "tokens_seen": 2380201984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835353535353535,
      "loss": 2.6903,
      "theoretical_loss": 3.381798274091975,
      "tokens_seen": 2380267520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835151515151515,
      "loss": 2.6797,
      "theoretical_loss": 3.381790773871571,
      "tokens_seen": 2380333056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028349494949494954,
      "loss": 2.6454,
      "theoretical_loss": 3.381783273915479,
      "tokens_seen": 2380398592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834747474747475,
      "loss": 2.5699,
      "theoretical_loss": 3.381775774223684,
      "tokens_seen": 2380464128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028345454545454545,
      "loss": 2.5086,
      "theoretical_loss": 3.3817682747961686,
      "tokens_seen": 2380529664
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834343434343434,
      "loss": 2.3691,
      "theoretical_loss": 3.3817607756329164,
      "tokens_seen": 2380595200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834141414141414,
      "loss": 2.6804,
      "theoretical_loss": 3.3817532767339107,
      "tokens_seen": 2380660736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028339393939393943,
      "loss": 2.7132,
      "theoretical_loss": 3.3817457780991353,
      "tokens_seen": 2380726272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833737373737374,
      "loss": 2.856,
      "theoretical_loss": 3.3817382797285735,
      "tokens_seen": 2380791808
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1307234,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6940321922302246,
      "objective/train/theoretical_loss": 3.3817307816222084,
      "objective/train/tokens_used": 739716576,
      "theoretical_loss": 3.3817307816222084,
      "tokens_seen": 2380857344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028335353535353534,
      "loss": 2.6585,
      "theoretical_loss": 3.3817307816222084,
      "tokens_seen": 2380857344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028333333333333335,
      "loss": 2.661,
      "theoretical_loss": 3.381723283780024,
      "tokens_seen": 2380922880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833131313131313,
      "loss": 2.7875,
      "theoretical_loss": 3.381715786202003,
      "tokens_seen": 2380988416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832929292929293,
      "loss": 2.7196,
      "theoretical_loss": 3.381708288888129,
      "tokens_seen": 2381053952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832727272727273,
      "loss": 2.5298,
      "theoretical_loss": 3.381700791838386,
      "tokens_seen": 2381119488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028325252525252524,
      "loss": 2.7766,
      "theoretical_loss": 3.3816932950527567,
      "tokens_seen": 2381185024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028323232323232325,
      "loss": 2.7571,
      "theoretical_loss": 3.381685798531225,
      "tokens_seen": 2381250560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832121212121212,
      "loss": 2.7308,
      "theoretical_loss": 3.381678302273774,
      "tokens_seen": 2381316096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831919191919192,
      "loss": 2.4646,
      "theoretical_loss": 3.3816708062803875,
      "tokens_seen": 2381381632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028317171717171717,
      "loss": 2.6481,
      "theoretical_loss": 3.381663310551049,
      "tokens_seen": 2381447168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831515151515152,
      "loss": 2.7243,
      "theoretical_loss": 3.3816558150857414,
      "tokens_seen": 2381512704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028313131313131314,
      "loss": 2.7129,
      "theoretical_loss": 3.381648319884448,
      "tokens_seen": 2381578240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831111111111111,
      "loss": 2.7205,
      "theoretical_loss": 3.3816408249471532,
      "tokens_seen": 2381643776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830909090909091,
      "loss": 2.7871,
      "theoretical_loss": 3.3816333302738397,
      "tokens_seen": 2381709312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028307070707070706,
      "loss": 2.6776,
      "theoretical_loss": 3.381625835864491,
      "tokens_seen": 2381774848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830505050505051,
      "loss": 2.7531,
      "theoretical_loss": 3.381618341719091,
      "tokens_seen": 2381840384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028303030303030303,
      "loss": 2.6221,
      "theoretical_loss": 3.3816108478376226,
      "tokens_seen": 2381905920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283010101010101,
      "loss": 2.8476,
      "theoretical_loss": 3.3816033542200694,
      "tokens_seen": 2381971456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282989898989899,
      "loss": 2.5985,
      "theoretical_loss": 3.3815958608664145,
      "tokens_seen": 2382036992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282969696969697,
      "loss": 2.5843,
      "theoretical_loss": 3.3815883677766423,
      "tokens_seen": 2382102528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028294949494949497,
      "loss": 2.5626,
      "theoretical_loss": 3.3815808749507354,
      "tokens_seen": 2382168064
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829292929292929,
      "loss": 2.611,
      "theoretical_loss": 3.3815733823886776,
      "tokens_seen": 2382233600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829090909090909,
      "loss": 2.5989,
      "theoretical_loss": 3.381565890090452,
      "tokens_seen": 2382299136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828888888888889,
      "loss": 2.6684,
      "theoretical_loss": 3.381558398056043,
      "tokens_seen": 2382364672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828686868686869,
      "loss": 2.688,
      "theoretical_loss": 3.3815509062854328,
      "tokens_seen": 2382430208
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1308524,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5440561771392822,
      "objective/train/theoretical_loss": 3.381543414778606,
      "objective/train/tokens_used": 741354976,
      "theoretical_loss": 3.381543414778606,
      "tokens_seen": 2382495744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028284848484848486,
      "loss": 2.5642,
      "theoretical_loss": 3.381543414778606,
      "tokens_seen": 2382495744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828282828282828,
      "loss": 2.655,
      "theoretical_loss": 3.381535923535545,
      "tokens_seen": 2382561280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828080808080808,
      "loss": 2.5311,
      "theoretical_loss": 3.381528432556234,
      "tokens_seen": 2382626816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028278787878787884,
      "loss": 2.3585,
      "theoretical_loss": 3.3815209418406558,
      "tokens_seen": 2382692352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827676767676768,
      "loss": 2.6624,
      "theoretical_loss": 3.3815134513887948,
      "tokens_seen": 2382757888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028274747474747475,
      "loss": 2.7171,
      "theoretical_loss": 3.3815059612006335,
      "tokens_seen": 2382823424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827272727272727,
      "loss": 2.6547,
      "theoretical_loss": 3.381498471276156,
      "tokens_seen": 2382888960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028270707070707067,
      "loss": 2.6185,
      "theoretical_loss": 3.381490981615346,
      "tokens_seen": 2382954496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028268686868686873,
      "loss": 2.588,
      "theoretical_loss": 3.3814834922181856,
      "tokens_seen": 2383020032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826666666666667,
      "loss": 2.7493,
      "theoretical_loss": 3.38147600308466,
      "tokens_seen": 2383085568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028264646464646465,
      "loss": 2.6777,
      "theoretical_loss": 3.3814685142147516,
      "tokens_seen": 2383151104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826262626262626,
      "loss": 2.7726,
      "theoretical_loss": 3.381461025608444,
      "tokens_seen": 2383216640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826060606060606,
      "loss": 2.5976,
      "theoretical_loss": 3.381453537265721,
      "tokens_seen": 2383282176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825858585858586,
      "loss": 2.7502,
      "theoretical_loss": 3.381446049186566,
      "tokens_seen": 2383347712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825656565656566,
      "loss": 2.7912,
      "theoretical_loss": 3.381438561370962,
      "tokens_seen": 2383413248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028254545454545454,
      "loss": 2.4623,
      "theoretical_loss": 3.3814310738188933,
      "tokens_seen": 2383478784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825252525252525,
      "loss": 2.4759,
      "theoretical_loss": 3.3814235865303424,
      "tokens_seen": 2383544320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825050505050505,
      "loss": 2.722,
      "theoretical_loss": 3.3814160995052935,
      "tokens_seen": 2383609856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824848484848485,
      "loss": 2.6941,
      "theoretical_loss": 3.38140861274373,
      "tokens_seen": 2383675392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824646464646465,
      "loss": 2.6101,
      "theoretical_loss": 3.3814011262456356,
      "tokens_seen": 2383740928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028244444444444443,
      "loss": 2.5728,
      "theoretical_loss": 3.3813936400109927,
      "tokens_seen": 2383806464
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028242424242424244,
      "loss": 2.6235,
      "theoretical_loss": 3.3813861540397863,
      "tokens_seen": 2383872000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824040404040404,
      "loss": 2.5543,
      "theoretical_loss": 3.381378668331999,
      "tokens_seen": 2383937536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823838383838384,
      "loss": 2.6172,
      "theoretical_loss": 3.381371182887614,
      "tokens_seen": 2384003072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028236363636363637,
      "loss": 2.5012,
      "theoretical_loss": 3.3813636977066155,
      "tokens_seen": 2384068608
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1309072,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.579392433166504,
      "objective/train/theoretical_loss": 3.3813562127889867,
      "objective/train/tokens_used": 742993376,
      "theoretical_loss": 3.3813562127889867,
      "tokens_seen": 2384134144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823434343434343,
      "loss": 2.5352,
      "theoretical_loss": 3.3813562127889867,
      "tokens_seen": 2384134144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028232323232323233,
      "loss": 2.5313,
      "theoretical_loss": 3.3813487281347108,
      "tokens_seen": 2384199680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823030303030303,
      "loss": 2.5842,
      "theoretical_loss": 3.3813412437437718,
      "tokens_seen": 2384265216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822828282828283,
      "loss": 2.4019,
      "theoretical_loss": 3.381333759616153,
      "tokens_seen": 2384330752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028226262626262626,
      "loss": 2.5113,
      "theoretical_loss": 3.381326275751838,
      "tokens_seen": 2384396288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028224242424242427,
      "loss": 2.7688,
      "theoretical_loss": 3.38131879215081,
      "tokens_seen": 2384461824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028222222222222223,
      "loss": 2.5801,
      "theoretical_loss": 3.3813113088130526,
      "tokens_seen": 2384527360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822020202020202,
      "loss": 2.5308,
      "theoretical_loss": 3.38130382573855,
      "tokens_seen": 2384592896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821818181818182,
      "loss": 2.4669,
      "theoretical_loss": 3.3812963429272846,
      "tokens_seen": 2384658432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028216161616161615,
      "loss": 2.6314,
      "theoretical_loss": 3.3812888603792404,
      "tokens_seen": 2384723968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028214141414141416,
      "loss": 2.7238,
      "theoretical_loss": 3.381281378094401,
      "tokens_seen": 2384789504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821212121212121,
      "loss": 2.5219,
      "theoretical_loss": 3.3812738960727495,
      "tokens_seen": 2384855040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821010101010101,
      "loss": 2.562,
      "theoretical_loss": 3.38126641431427,
      "tokens_seen": 2384920576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820808080808081,
      "loss": 2.6081,
      "theoretical_loss": 3.381258932818946,
      "tokens_seen": 2384986112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820606060606061,
      "loss": 2.565,
      "theoretical_loss": 3.3812514515867607,
      "tokens_seen": 2385051648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028204040404040406,
      "loss": 2.67,
      "theoretical_loss": 3.3812439706176978,
      "tokens_seen": 2385117184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282020202020202,
      "loss": 2.6214,
      "theoretical_loss": 3.38123648991174,
      "tokens_seen": 2385182720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028199999999999997,
      "loss": 2.3948,
      "theoretical_loss": 3.381229009468872,
      "tokens_seen": 2385248256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281979797979798,
      "loss": 2.6144,
      "theoretical_loss": 3.3812215292890766,
      "tokens_seen": 2385313792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281959595959596,
      "loss": 2.2897,
      "theoretical_loss": 3.3812140493723377,
      "tokens_seen": 2385379328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028193939393939395,
      "loss": 2.7042,
      "theoretical_loss": 3.3812065697186386,
      "tokens_seen": 2385444864
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002819191919191919,
      "loss": 2.4648,
      "theoretical_loss": 3.381199090327963,
      "tokens_seen": 2385510400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028189898989898986,
      "loss": 2.6589,
      "theoretical_loss": 3.381191611200294,
      "tokens_seen": 2385575936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028187878787878793,
      "loss": 2.6317,
      "theoretical_loss": 3.381184132335616,
      "tokens_seen": 2385641472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818585858585859,
      "loss": 2.7684,
      "theoretical_loss": 3.381176653733912,
      "tokens_seen": 2385707008
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1310224,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8640713691711426,
      "objective/train/theoretical_loss": 3.381169175395165,
      "objective/train/tokens_used": 744631776,
      "theoretical_loss": 3.381169175395165,
      "tokens_seen": 2385772544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028183838383838384,
      "loss": 2.7092,
      "theoretical_loss": 3.381169175395165,
      "tokens_seen": 2385772544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818181818181818,
      "loss": 2.6205,
      "theoretical_loss": 3.3811616973193592,
      "tokens_seen": 2385838080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817979797979798,
      "loss": 2.4941,
      "theoretical_loss": 3.381154219506478,
      "tokens_seen": 2385903616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817777777777778,
      "loss": 2.6066,
      "theoretical_loss": 3.3811467419565044,
      "tokens_seen": 2385969152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817575757575758,
      "loss": 2.3554,
      "theoretical_loss": 3.3811392646694234,
      "tokens_seen": 2386034688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028173737373737373,
      "loss": 2.7169,
      "theoretical_loss": 3.381131787645217,
      "tokens_seen": 2386100224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817171717171717,
      "loss": 2.6337,
      "theoretical_loss": 3.381124310883869,
      "tokens_seen": 2386165760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028169696969696976,
      "loss": 2.6387,
      "theoretical_loss": 3.381116834385364,
      "tokens_seen": 2386231296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816767676767677,
      "loss": 2.6259,
      "theoretical_loss": 3.3811093581496845,
      "tokens_seen": 2386296832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028165656565656567,
      "loss": 2.4768,
      "theoretical_loss": 3.381101882176814,
      "tokens_seen": 2386362368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816363636363636,
      "loss": 2.6671,
      "theoretical_loss": 3.3810944064667368,
      "tokens_seen": 2386427904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816161616161616,
      "loss": 2.784,
      "theoretical_loss": 3.3810869310194356,
      "tokens_seen": 2386493440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028159595959595965,
      "loss": 2.5278,
      "theoretical_loss": 3.381079455834895,
      "tokens_seen": 2386558976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815757575757576,
      "loss": 2.8345,
      "theoretical_loss": 3.3810719809130974,
      "tokens_seen": 2386624512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028155555555555556,
      "loss": 2.738,
      "theoretical_loss": 3.381064506254027,
      "tokens_seen": 2386690048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815353535353535,
      "loss": 2.8047,
      "theoretical_loss": 3.381057031857667,
      "tokens_seen": 2386755584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815151515151515,
      "loss": 2.7453,
      "theoretical_loss": 3.3810495577240016,
      "tokens_seen": 2386821120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028149494949494954,
      "loss": 2.7913,
      "theoretical_loss": 3.3810420838530137,
      "tokens_seen": 2386886656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814747474747475,
      "loss": 2.6015,
      "theoretical_loss": 3.3810346102446873,
      "tokens_seen": 2386952192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028145454545454546,
      "loss": 2.4982,
      "theoretical_loss": 3.3810271368990055,
      "tokens_seen": 2387017728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814343434343434,
      "loss": 2.8202,
      "theoretical_loss": 3.381019663815952,
      "tokens_seen": 2387083264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814141414141414,
      "loss": 2.6795,
      "theoretical_loss": 3.3810121909955106,
      "tokens_seen": 2387148800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028139393939393943,
      "loss": 2.3849,
      "theoretical_loss": 3.381004718437665,
      "tokens_seen": 2387214336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813737373737374,
      "loss": 2.5284,
      "theoretical_loss": 3.3809972461423983,
      "tokens_seen": 2387279872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028135353535353535,
      "loss": 2.857,
      "theoretical_loss": 3.380989774109694,
      "tokens_seen": 2387345408
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1310846,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.386476755142212,
      "objective/train/theoretical_loss": 3.3809823023395364,
      "objective/train/tokens_used": 746270176,
      "theoretical_loss": 3.3809823023395364,
      "tokens_seen": 2387410944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813333333333333,
      "loss": 2.4214,
      "theoretical_loss": 3.3809823023395364,
      "tokens_seen": 2387410944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813131313131313,
      "loss": 2.6282,
      "theoretical_loss": 3.3809748308319083,
      "tokens_seen": 2387476480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028129292929292933,
      "loss": 2.579,
      "theoretical_loss": 3.3809673595867933,
      "tokens_seen": 2387542016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812727272727273,
      "loss": 2.7025,
      "theoretical_loss": 3.380959888604176,
      "tokens_seen": 2387607552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028125252525252524,
      "loss": 2.6733,
      "theoretical_loss": 3.3809524178840387,
      "tokens_seen": 2387673088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028123232323232325,
      "loss": 2.4146,
      "theoretical_loss": 3.3809449474263653,
      "tokens_seen": 2387738624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812121212121212,
      "loss": 2.6842,
      "theoretical_loss": 3.38093747723114,
      "tokens_seen": 2387804160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811919191919192,
      "loss": 2.4407,
      "theoretical_loss": 3.3809300072983453,
      "tokens_seen": 2387869696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811717171717172,
      "loss": 2.4228,
      "theoretical_loss": 3.380922537627966,
      "tokens_seen": 2387935232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028115151515151513,
      "loss": 2.7888,
      "theoretical_loss": 3.380915068219985,
      "tokens_seen": 2388000768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028113131313131314,
      "loss": 2.4777,
      "theoretical_loss": 3.3809075990743858,
      "tokens_seen": 2388066304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811111111111111,
      "loss": 2.5646,
      "theoretical_loss": 3.380900130191152,
      "tokens_seen": 2388131840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810909090909091,
      "loss": 2.5683,
      "theoretical_loss": 3.3808926615702677,
      "tokens_seen": 2388197376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028107070707070707,
      "loss": 2.4524,
      "theoretical_loss": 3.380885193211716,
      "tokens_seen": 2388262912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810505050505051,
      "loss": 2.6967,
      "theoretical_loss": 3.380877725115481,
      "tokens_seen": 2388328448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028103030303030304,
      "loss": 2.6038,
      "theoretical_loss": 3.380870257281545,
      "tokens_seen": 2388393984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281010101010101,
      "loss": 2.4525,
      "theoretical_loss": 3.380862789709893,
      "tokens_seen": 2388459520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280989898989899,
      "loss": 2.5881,
      "theoretical_loss": 3.3808553224005085,
      "tokens_seen": 2388525056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028096969696969696,
      "loss": 2.5306,
      "theoretical_loss": 3.380847855353374,
      "tokens_seen": 2388590592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280949494949495,
      "loss": 2.7317,
      "theoretical_loss": 3.3808403885684744,
      "tokens_seen": 2388656128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028092929292929293,
      "loss": 2.4498,
      "theoretical_loss": 3.380832922045792,
      "tokens_seen": 2388721664
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002809090909090909,
      "loss": 2.6253,
      "theoretical_loss": 3.3808254557853115,
      "tokens_seen": 2388787200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808888888888889,
      "loss": 2.6037,
      "theoretical_loss": 3.380817989787016,
      "tokens_seen": 2388852736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808686868686869,
      "loss": 2.8293,
      "theoretical_loss": 3.3808105240508888,
      "tokens_seen": 2388918272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028084848484848487,
      "loss": 2.4457,
      "theoretical_loss": 3.380803058576914,
      "tokens_seen": 2388983808
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1312174,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4064464569091797,
      "objective/train/theoretical_loss": 3.3807955933650753,
      "objective/train/tokens_used": 747908576,
      "theoretical_loss": 3.3807955933650753,
      "tokens_seen": 2389049344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808282828282828,
      "loss": 2.626,
      "theoretical_loss": 3.3807955933650753,
      "tokens_seen": 2389049344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808080808080808,
      "loss": 2.7644,
      "theoretical_loss": 3.3807881284153565,
      "tokens_seen": 2389114880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807878787878788,
      "loss": 2.7973,
      "theoretical_loss": 3.38078066372774,
      "tokens_seen": 2389180416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807676767676768,
      "loss": 2.7315,
      "theoretical_loss": 3.3807731993022108,
      "tokens_seen": 2389245952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028074747474747476,
      "loss": 2.5144,
      "theoretical_loss": 3.3807657351387514,
      "tokens_seen": 2389311488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807272727272727,
      "loss": 2.5448,
      "theoretical_loss": 3.380758271237346,
      "tokens_seen": 2389377024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028070707070707067,
      "loss": 2.5791,
      "theoretical_loss": 3.380750807597978,
      "tokens_seen": 2389442560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028068686868686874,
      "loss": 2.8976,
      "theoretical_loss": 3.3807433442206314,
      "tokens_seen": 2389508096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806666666666667,
      "loss": 2.4212,
      "theoretical_loss": 3.3807358811052897,
      "tokens_seen": 2389573632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028064646464646465,
      "loss": 2.5482,
      "theoretical_loss": 3.380728418251936,
      "tokens_seen": 2389639168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806262626262626,
      "loss": 2.4117,
      "theoretical_loss": 3.380720955660555,
      "tokens_seen": 2389704704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028060606060606057,
      "loss": 2.6246,
      "theoretical_loss": 3.380713493331129,
      "tokens_seen": 2389770240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028058585858585863,
      "loss": 2.4588,
      "theoretical_loss": 3.3807060312636423,
      "tokens_seen": 2389835776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805656565656566,
      "loss": 2.6998,
      "theoretical_loss": 3.3806985694580782,
      "tokens_seen": 2389901312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028054545454545454,
      "loss": 2.4716,
      "theoretical_loss": 3.380691107914421,
      "tokens_seen": 2389966848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805252525252525,
      "loss": 2.5811,
      "theoretical_loss": 3.3806836466326535,
      "tokens_seen": 2390032384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805050505050505,
      "loss": 2.7345,
      "theoretical_loss": 3.38067618561276,
      "tokens_seen": 2390097920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804848484848485,
      "loss": 2.4443,
      "theoretical_loss": 3.3806687248547242,
      "tokens_seen": 2390163456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804646464646465,
      "loss": 2.6973,
      "theoretical_loss": 3.3806612643585288,
      "tokens_seen": 2390228992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028044444444444444,
      "loss": 2.5089,
      "theoretical_loss": 3.380653804124158,
      "tokens_seen": 2390294528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804242424242424,
      "loss": 2.5099,
      "theoretical_loss": 3.380646344151596,
      "tokens_seen": 2390360064
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804040404040404,
      "loss": 2.7341,
      "theoretical_loss": 3.3806388844408257,
      "tokens_seen": 2390425600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002803838383838384,
      "loss": 2.3642,
      "theoretical_loss": 3.380631424991831,
      "tokens_seen": 2390491136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028036363636363637,
      "loss": 2.7166,
      "theoretical_loss": 3.380623965804595,
      "tokens_seen": 2390556672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028034343434343433,
      "loss": 2.5862,
      "theoretical_loss": 3.380616506879102,
      "tokens_seen": 2390622208
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 1312726,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7473642826080322,
      "objective/train/theoretical_loss": 3.3806090482153355,
      "objective/train/tokens_used": 749546976,
      "theoretical_loss": 3.3806090482153355,
      "tokens_seen": 2390687744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803232323232323,
      "loss": 2.7195,
      "theoretical_loss": 3.3806090482153355,
      "tokens_seen": 2390687744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803030303030303,
      "loss": 2.6134,
      "theoretical_loss": 3.3806015898132795,
      "tokens_seen": 2390753280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802828282828283,
      "loss": 2.4971,
      "theoretical_loss": 3.3805941316729164,
      "tokens_seen": 2390818816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028026262626262627,
      "loss": 2.5792,
      "theoretical_loss": 3.380586673794231,
      "tokens_seen": 2390884352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802424242424242,
      "loss": 2.7271,
      "theoretical_loss": 3.380579216177207,
      "tokens_seen": 2390949888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028022222222222223,
      "loss": 2.3641,
      "theoretical_loss": 3.3805717588218274,
      "tokens_seen": 2391015424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802020202020202,
      "loss": 2.4436,
      "theoretical_loss": 3.380564301728076,
      "tokens_seen": 2391080960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801818181818182,
      "loss": 2.7457,
      "theoretical_loss": 3.3805568448959367,
      "tokens_seen": 2391146496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028016161616161616,
      "loss": 2.599,
      "theoretical_loss": 3.380549388325393,
      "tokens_seen": 2391212032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801414141414141,
      "loss": 2.6285,
      "theoretical_loss": 3.3805419320164285,
      "tokens_seen": 2391277568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801212121212121,
      "loss": 2.6415,
      "theoretical_loss": 3.3805344759690272,
      "tokens_seen": 2391343104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028010101010101014,
      "loss": 2.6648,
      "theoretical_loss": 3.3805270201831723,
      "tokens_seen": 2391408640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002800808080808081,
      "loss": 2.4228,
      "theoretical_loss": 3.3805195646588473,
      "tokens_seen": 2391474176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028006060606060605,
      "loss": 2.7256,
      "theoretical_loss": 3.3805121093960366,
      "tokens_seen": 2391539712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028004040404040406,
      "loss": 2.6155,
      "theoretical_loss": 3.3805046543947235,
      "tokens_seen": 2391605248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000280020202020202,
      "loss": 2.7055,
      "theoretical_loss": 3.3804971996548914,
      "tokens_seen": 2391670784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028000000000000003,
      "loss": 2.764,
      "theoretical_loss": 3.3804897451765243,
      "tokens_seen": 2391736320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279979797979798,
      "loss": 2.4448,
      "theoretical_loss": 3.3804822909596055,
      "tokens_seen": 2391801856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027995959595959594,
      "loss": 2.5938,
      "theoretical_loss": 3.3804748370041193,
      "tokens_seen": 2391867392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027993939393939395,
      "loss": 2.6107,
      "theoretical_loss": 3.380467383310049,
      "tokens_seen": 2391932928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002799191919191919,
      "loss": 2.6505,
      "theoretical_loss": 3.380459929877378,
      "tokens_seen": 2391998464
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798989898989899,
      "loss": 2.4188,
      "theoretical_loss": 3.3804524767060906,
      "tokens_seen": 2392064000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798787878787879,
      "loss": 2.6168,
      "theoretical_loss": 3.3804450237961694,
      "tokens_seen": 2392129536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798585858585859,
      "loss": 2.6171,
      "theoretical_loss": 3.3804375711475996,
      "tokens_seen": 2392195072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027983838383838385,
      "loss": 2.6437,
      "theoretical_loss": 3.380430118760364,
      "tokens_seen": 2392260608
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1313802,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4760336875915527,
      "objective/train/theoretical_loss": 3.3804226666344457,
      "objective/train/tokens_used": 751185376,
      "theoretical_loss": 3.3804226666344457,
      "tokens_seen": 2392326144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798181818181818,
      "loss": 2.6366,
      "theoretical_loss": 3.3804226666344457,
      "tokens_seen": 2392326144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797979797979798,
      "loss": 2.6789,
      "theoretical_loss": 3.380415214769829,
      "tokens_seen": 2392391680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027977777777777777,
      "loss": 2.7024,
      "theoretical_loss": 3.380407763166498,
      "tokens_seen": 2392457216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797575757575758,
      "loss": 2.6339,
      "theoretical_loss": 3.380400311824436,
      "tokens_seen": 2392522752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027973737373737374,
      "loss": 2.4198,
      "theoretical_loss": 3.3803928607436267,
      "tokens_seen": 2392588288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797171717171717,
      "loss": 2.5977,
      "theoretical_loss": 3.3803854099240533,
      "tokens_seen": 2392653824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796969696969697,
      "loss": 2.6416,
      "theoretical_loss": 3.3803779593657004,
      "tokens_seen": 2392719360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796767676767677,
      "loss": 2.7815,
      "theoretical_loss": 3.380370509068551,
      "tokens_seen": 2392784896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796565656565657,
      "loss": 2.5654,
      "theoretical_loss": 3.380363059032589,
      "tokens_seen": 2392850432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027963636363636363,
      "loss": 2.4607,
      "theoretical_loss": 3.3803556092577978,
      "tokens_seen": 2392915968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796161616161616,
      "loss": 2.4767,
      "theoretical_loss": 3.380348159744162,
      "tokens_seen": 2392981504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795959595959596,
      "loss": 2.342,
      "theoretical_loss": 3.380340710491664,
      "tokens_seen": 2393047040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795757575757576,
      "loss": 2.6491,
      "theoretical_loss": 3.380333261500289,
      "tokens_seen": 2393112576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027955555555555557,
      "loss": 2.6328,
      "theoretical_loss": 3.3803258127700193,
      "tokens_seen": 2393178112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795353535353535,
      "loss": 2.7062,
      "theoretical_loss": 3.380318364300839,
      "tokens_seen": 2393243648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795151515151515,
      "loss": 2.5557,
      "theoretical_loss": 3.380310916092732,
      "tokens_seen": 2393309184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027949494949494955,
      "loss": 2.7179,
      "theoretical_loss": 3.3803034681456827,
      "tokens_seen": 2393374720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794747474747475,
      "loss": 2.6354,
      "theoretical_loss": 3.3802960204596735,
      "tokens_seen": 2393440256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027945454545454546,
      "loss": 2.4896,
      "theoretical_loss": 3.3802885730346888,
      "tokens_seen": 2393505792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794343434343434,
      "loss": 2.6934,
      "theoretical_loss": 3.380281125870712,
      "tokens_seen": 2393571328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794141414141414,
      "loss": 2.4742,
      "theoretical_loss": 3.3802736789677272,
      "tokens_seen": 2393636864
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027939393939393944,
      "loss": 2.5481,
      "theoretical_loss": 3.3802662323257175,
      "tokens_seen": 2393702400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793737373737374,
      "loss": 2.5392,
      "theoretical_loss": 3.380258785944667,
      "tokens_seen": 2393767936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027935353535353535,
      "loss": 2.5711,
      "theoretical_loss": 3.38025133982456,
      "tokens_seen": 2393833472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793333333333333,
      "loss": 2.4893,
      "theoretical_loss": 3.380243893965379,
      "tokens_seen": 2393899008
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1314506,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.881012201309204,
      "objective/train/theoretical_loss": 3.380236448367109,
      "objective/train/tokens_used": 752823776,
      "theoretical_loss": 3.380236448367109,
      "tokens_seen": 2393964544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793131313131313,
      "loss": 2.8527,
      "theoretical_loss": 3.380236448367109,
      "tokens_seen": 2393964544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027929292929292933,
      "loss": 2.6398,
      "theoretical_loss": 3.380229003029733,
      "tokens_seen": 2394030080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792727272727273,
      "loss": 2.637,
      "theoretical_loss": 3.380221557953234,
      "tokens_seen": 2394095616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027925252525252525,
      "loss": 2.7507,
      "theoretical_loss": 3.380214113137597,
      "tokens_seen": 2394161152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792323232323232,
      "loss": 2.604,
      "theoretical_loss": 3.3802066685828054,
      "tokens_seen": 2394226688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792121212121212,
      "loss": 2.6674,
      "theoretical_loss": 3.3801992242888423,
      "tokens_seen": 2394292224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791919191919192,
      "loss": 2.6199,
      "theoretical_loss": 3.380191780255692,
      "tokens_seen": 2394357760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791717171717172,
      "loss": 2.7008,
      "theoretical_loss": 3.380184336483338,
      "tokens_seen": 2394423296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027915151515151514,
      "loss": 2.6966,
      "theoretical_loss": 3.380176892971764,
      "tokens_seen": 2394488832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027913131313131315,
      "loss": 2.7605,
      "theoretical_loss": 3.3801694497209542,
      "tokens_seen": 2394554368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791111111111111,
      "loss": 2.4733,
      "theoretical_loss": 3.3801620067308917,
      "tokens_seen": 2394619904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790909090909091,
      "loss": 2.5202,
      "theoretical_loss": 3.380154564001561,
      "tokens_seen": 2394685440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790707070707071,
      "loss": 2.455,
      "theoretical_loss": 3.3801471215329446,
      "tokens_seen": 2394750976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027905050505050503,
      "loss": 2.5015,
      "theoretical_loss": 3.380139679325027,
      "tokens_seen": 2394816512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027903030303030304,
      "loss": 2.5401,
      "theoretical_loss": 3.3801322373777927,
      "tokens_seen": 2394882048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279010101010101,
      "loss": 2.5432,
      "theoretical_loss": 3.380124795691224,
      "tokens_seen": 2394947584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278989898989899,
      "loss": 2.5095,
      "theoretical_loss": 3.3801173542653054,
      "tokens_seen": 2395013120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027896969696969697,
      "loss": 2.7084,
      "theoretical_loss": 3.38010991310002,
      "tokens_seen": 2395078656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278949494949495,
      "loss": 2.5357,
      "theoretical_loss": 3.3801024721953525,
      "tokens_seen": 2395144192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027892929292929294,
      "loss": 2.5665,
      "theoretical_loss": 3.380095031551286,
      "tokens_seen": 2395209728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789090909090909,
      "loss": 2.7815,
      "theoretical_loss": 3.3800875911678046,
      "tokens_seen": 2395275264
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788888888888889,
      "loss": 2.6284,
      "theoretical_loss": 3.3800801510448917,
      "tokens_seen": 2395340800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027886868686868686,
      "loss": 2.5418,
      "theoretical_loss": 3.3800727111825317,
      "tokens_seen": 2395406336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027884848484848487,
      "loss": 2.5,
      "theoretical_loss": 3.3800652715807074,
      "tokens_seen": 2395471872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027882828282828283,
      "loss": 2.41,
      "theoretical_loss": 3.3800578322394026,
      "tokens_seen": 2395537408
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1315670,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2574543952941895,
      "objective/train/theoretical_loss": 3.3800503931586023,
      "objective/train/tokens_used": 754462176,
      "theoretical_loss": 3.3800503931586023,
      "tokens_seen": 2395602944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788080808080808,
      "loss": 2.4554,
      "theoretical_loss": 3.3800503931586023,
      "tokens_seen": 2395602944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787878787878788,
      "loss": 2.547,
      "theoretical_loss": 3.380042954338289,
      "tokens_seen": 2395668480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787676767676768,
      "loss": 2.5429,
      "theoretical_loss": 3.380035515778447,
      "tokens_seen": 2395734016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027874747474747476,
      "loss": 2.8526,
      "theoretical_loss": 3.3800280774790594,
      "tokens_seen": 2395799552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787272727272727,
      "loss": 2.556,
      "theoretical_loss": 3.380020639440111,
      "tokens_seen": 2395865088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787070707070707,
      "loss": 2.6507,
      "theoretical_loss": 3.380013201661585,
      "tokens_seen": 2395930624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786868686868687,
      "loss": 2.5906,
      "theoretical_loss": 3.380005764143465,
      "tokens_seen": 2395996160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786666666666667,
      "loss": 2.7161,
      "theoretical_loss": 3.379998326885735,
      "tokens_seen": 2396061696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027864646464646466,
      "loss": 2.4844,
      "theoretical_loss": 3.379990889888379,
      "tokens_seen": 2396127232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786262626262626,
      "loss": 2.4491,
      "theoretical_loss": 3.3799834531513797,
      "tokens_seen": 2396192768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027860606060606057,
      "loss": 2.6456,
      "theoretical_loss": 3.3799760166747226,
      "tokens_seen": 2396258304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027858585858585864,
      "loss": 2.5085,
      "theoretical_loss": 3.37996858045839,
      "tokens_seen": 2396323840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785656565656566,
      "loss": 2.8677,
      "theoretical_loss": 3.3799611445023663,
      "tokens_seen": 2396389376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027854545454545455,
      "loss": 2.5066,
      "theoretical_loss": 3.379953708806635,
      "tokens_seen": 2396454912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785252525252525,
      "loss": 2.4542,
      "theoretical_loss": 3.3799462733711803,
      "tokens_seen": 2396520448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027850505050505046,
      "loss": 2.5422,
      "theoretical_loss": 3.379938838195985,
      "tokens_seen": 2396585984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027848484848484853,
      "loss": 2.7111,
      "theoretical_loss": 3.3799314032810344,
      "tokens_seen": 2396651520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784646464646465,
      "loss": 2.6067,
      "theoretical_loss": 3.379923968626311,
      "tokens_seen": 2396717056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027844444444444444,
      "loss": 2.6897,
      "theoretical_loss": 3.379916534231799,
      "tokens_seen": 2396782592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784242424242424,
      "loss": 2.3453,
      "theoretical_loss": 3.3799091000974824,
      "tokens_seen": 2396848128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784040404040404,
      "loss": 2.5504,
      "theoretical_loss": 3.3799016662233448,
      "tokens_seen": 2396913664
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783838383838384,
      "loss": 2.7081,
      "theoretical_loss": 3.3798942326093697,
      "tokens_seen": 2396979200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783636363636364,
      "loss": 2.5611,
      "theoretical_loss": 3.3798867992555413,
      "tokens_seen": 2397044736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027834343434343434,
      "loss": 2.7388,
      "theoretical_loss": 3.379879366161843,
      "tokens_seen": 2397110272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783232323232323,
      "loss": 2.3034,
      "theoretical_loss": 3.3798719333282587,
      "tokens_seen": 2397175808
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1316257,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.401369094848633,
      "objective/train/theoretical_loss": 3.3798645007547727,
      "objective/train/tokens_used": 756100576,
      "theoretical_loss": 3.3798645007547727,
      "tokens_seen": 2397241344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027830303030303036,
      "loss": 2.619,
      "theoretical_loss": 3.3798645007547727,
      "tokens_seen": 2397241344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782828282828283,
      "loss": 2.3024,
      "theoretical_loss": 3.3798570684413685,
      "tokens_seen": 2397306880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027826262626262627,
      "loss": 2.5966,
      "theoretical_loss": 3.3798496363880295,
      "tokens_seen": 2397372416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027824242424242423,
      "loss": 2.6577,
      "theoretical_loss": 3.3798422045947394,
      "tokens_seen": 2397437952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782222222222222,
      "loss": 2.6159,
      "theoretical_loss": 3.3798347730614826,
      "tokens_seen": 2397503488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027820202020202025,
      "loss": 2.5999,
      "theoretical_loss": 3.3798273417882427,
      "tokens_seen": 2397569024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781818181818182,
      "loss": 2.6694,
      "theoretical_loss": 3.3798199107750038,
      "tokens_seen": 2397634560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027816161616161616,
      "loss": 2.616,
      "theoretical_loss": 3.379812480021749,
      "tokens_seen": 2397700096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781414141414141,
      "loss": 2.5596,
      "theoretical_loss": 3.3798050495284624,
      "tokens_seen": 2397765632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027812121212121213,
      "loss": 2.3258,
      "theoretical_loss": 3.3797976192951276,
      "tokens_seen": 2397831168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027810101010101014,
      "loss": 2.6237,
      "theoretical_loss": 3.379790189321729,
      "tokens_seen": 2397896704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780808080808081,
      "loss": 2.5432,
      "theoretical_loss": 3.3797827596082497,
      "tokens_seen": 2397962240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027806060606060606,
      "loss": 2.4558,
      "theoretical_loss": 3.3797753301546742,
      "tokens_seen": 2398027776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278040404040404,
      "loss": 2.4353,
      "theoretical_loss": 3.379767900960986,
      "tokens_seen": 2398093312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278020202020202,
      "loss": 2.6536,
      "theoretical_loss": 3.3797604720271686,
      "tokens_seen": 2398158848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027800000000000004,
      "loss": 2.6172,
      "theoretical_loss": 3.3797530433532064,
      "tokens_seen": 2398224384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277979797979798,
      "loss": 2.5712,
      "theoretical_loss": 3.3797456149390825,
      "tokens_seen": 2398289920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027795959595959595,
      "loss": 2.4917,
      "theoretical_loss": 3.379738186784781,
      "tokens_seen": 2398355456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027793939393939396,
      "loss": 2.5065,
      "theoretical_loss": 3.379730758890286,
      "tokens_seen": 2398420992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779191919191919,
      "loss": 2.4775,
      "theoretical_loss": 3.379723331255581,
      "tokens_seen": 2398486528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027789898989898993,
      "loss": 2.6639,
      "theoretical_loss": 3.3797159038806504,
      "tokens_seen": 2398552064
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778787878787879,
      "loss": 2.4218,
      "theoretical_loss": 3.379708476765477,
      "tokens_seen": 2398617600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027785858585858584,
      "loss": 2.5206,
      "theoretical_loss": 3.3797010499100453,
      "tokens_seen": 2398683136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027783838383838385,
      "loss": 2.6598,
      "theoretical_loss": 3.379693623314339,
      "tokens_seen": 2398748672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778181818181818,
      "loss": 2.6251,
      "theoretical_loss": 3.379686196978342,
      "tokens_seen": 2398814208
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1317351,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.750336170196533,
      "objective/train/theoretical_loss": 3.379678770902038,
      "objective/train/tokens_used": 757738976,
      "theoretical_loss": 3.379678770902038,
      "tokens_seen": 2398879744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777979797979798,
      "loss": 2.6726,
      "theoretical_loss": 3.379678770902038,
      "tokens_seen": 2398879744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777777777777778,
      "loss": 2.4826,
      "theoretical_loss": 3.3796713450854106,
      "tokens_seen": 2398945280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777575757575758,
      "loss": 2.5717,
      "theoretical_loss": 3.3796639195284444,
      "tokens_seen": 2399010816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027773737373737375,
      "loss": 2.5577,
      "theoretical_loss": 3.379656494231122,
      "tokens_seen": 2399076352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777171717171717,
      "loss": 2.1633,
      "theoretical_loss": 3.3796490691934284,
      "tokens_seen": 2399141888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776969696969697,
      "loss": 2.7594,
      "theoretical_loss": 3.379641644415347,
      "tokens_seen": 2399207424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027767676767676767,
      "loss": 2.5037,
      "theoretical_loss": 3.3796342198968614,
      "tokens_seen": 2399272960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776565656565657,
      "loss": 2.4903,
      "theoretical_loss": 3.3796267956379555,
      "tokens_seen": 2399338496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027763636363636364,
      "loss": 2.472,
      "theoretical_loss": 3.3796193716386136,
      "tokens_seen": 2399404032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776161616161616,
      "loss": 2.5285,
      "theoretical_loss": 3.379611947898819,
      "tokens_seen": 2399469568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775959595959596,
      "loss": 2.6631,
      "theoretical_loss": 3.379604524418556,
      "tokens_seen": 2399535104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775757575757576,
      "loss": 2.6522,
      "theoretical_loss": 3.379597101197808,
      "tokens_seen": 2399600640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775555555555556,
      "loss": 2.3629,
      "theoretical_loss": 3.3795896782365586,
      "tokens_seen": 2399666176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027753535353535353,
      "loss": 2.5396,
      "theoretical_loss": 3.3795822555347925,
      "tokens_seen": 2399731712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775151515151515,
      "loss": 2.5843,
      "theoretical_loss": 3.3795748330924935,
      "tokens_seen": 2399797248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774949494949495,
      "loss": 2.4912,
      "theoretical_loss": 3.3795674109096443,
      "tokens_seen": 2399862784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774747474747475,
      "loss": 2.5834,
      "theoretical_loss": 3.37955998898623,
      "tokens_seen": 2399928320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027745454545454547,
      "loss": 2.5894,
      "theoretical_loss": 3.3795525673222335,
      "tokens_seen": 2399993856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774343434343434,
      "loss": 2.6344,
      "theoretical_loss": 3.3795451459176395,
      "tokens_seen": 2400059392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774141414141414,
      "loss": 2.5681,
      "theoretical_loss": 3.3795377247724314,
      "tokens_seen": 2400124928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027739393939393945,
      "loss": 2.5088,
      "theoretical_loss": 3.3795303038865927,
      "tokens_seen": 2400190464
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773737373737374,
      "loss": 2.5978,
      "theoretical_loss": 3.379522883260108,
      "tokens_seen": 2400256000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027735353535353536,
      "loss": 2.5245,
      "theoretical_loss": 3.379515462892961,
      "tokens_seen": 2400321536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773333333333333,
      "loss": 2.4467,
      "theoretical_loss": 3.379508042785135,
      "tokens_seen": 2400387072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773131313131313,
      "loss": 2.5966,
      "theoretical_loss": 3.379500622936614,
      "tokens_seen": 2400452608
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1318782,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.464829921722412,
      "objective/train/theoretical_loss": 3.3794932033473826,
      "objective/train/tokens_used": 759377376,
      "theoretical_loss": 3.3794932033473826,
      "tokens_seen": 2400518144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027729292929292934,
      "loss": 2.6287,
      "theoretical_loss": 3.3794932033473826,
      "tokens_seen": 2400518144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772727272727273,
      "loss": 2.6682,
      "theoretical_loss": 3.3794857840174233,
      "tokens_seen": 2400583680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027725252525252525,
      "loss": 2.4109,
      "theoretical_loss": 3.3794783649467215,
      "tokens_seen": 2400649216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772323232323232,
      "loss": 2.691,
      "theoretical_loss": 3.3794709461352603,
      "tokens_seen": 2400714752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772121212121212,
      "loss": 2.5769,
      "theoretical_loss": 3.379463527583024,
      "tokens_seen": 2400780288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027719191919191923,
      "loss": 2.6515,
      "theoretical_loss": 3.379456109289995,
      "tokens_seen": 2400845824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771717171717172,
      "loss": 2.5084,
      "theoretical_loss": 3.379448691256159,
      "tokens_seen": 2400911360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027715151515151514,
      "loss": 2.5864,
      "theoretical_loss": 3.379441273481499,
      "tokens_seen": 2400976896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771313131313131,
      "loss": 2.4552,
      "theoretical_loss": 3.379433855965999,
      "tokens_seen": 2401042432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771111111111111,
      "loss": 2.5868,
      "theoretical_loss": 3.3794264387096433,
      "tokens_seen": 2401107968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770909090909091,
      "loss": 2.446,
      "theoretical_loss": 3.379419021712415,
      "tokens_seen": 2401173504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770707070707071,
      "loss": 2.5476,
      "theoretical_loss": 3.379411604974298,
      "tokens_seen": 2401239040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027705050505050504,
      "loss": 2.8137,
      "theoretical_loss": 3.379404188495277,
      "tokens_seen": 2401304576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277030303030303,
      "loss": 2.6532,
      "theoretical_loss": 3.379396772275335,
      "tokens_seen": 2401370112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277010101010101,
      "loss": 2.6834,
      "theoretical_loss": 3.3793893563144564,
      "tokens_seen": 2401435648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276989898989899,
      "loss": 2.4592,
      "theoretical_loss": 3.3793819406126246,
      "tokens_seen": 2401501184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276969696969697,
      "loss": 2.4117,
      "theoretical_loss": 3.3793745251698244,
      "tokens_seen": 2401566720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027694949494949493,
      "loss": 2.5192,
      "theoretical_loss": 3.3793671099860387,
      "tokens_seen": 2401632256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027692929292929294,
      "loss": 2.6241,
      "theoretical_loss": 3.379359695061252,
      "tokens_seen": 2401697792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769090909090909,
      "loss": 2.5624,
      "theoretical_loss": 3.379352280395448,
      "tokens_seen": 2401763328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768888888888889,
      "loss": 2.6287,
      "theoretical_loss": 3.3793448659886103,
      "tokens_seen": 2401828864
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027686868686868687,
      "loss": 2.6455,
      "theoretical_loss": 3.3793374518407235,
      "tokens_seen": 2401894400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768484848484848,
      "loss": 2.5204,
      "theoretical_loss": 3.3793300379517706,
      "tokens_seen": 2401959936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027682828282828283,
      "loss": 2.5895,
      "theoretical_loss": 3.379322624321736,
      "tokens_seen": 2402025472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768080808080808,
      "loss": 2.482,
      "theoretical_loss": 3.3793152109506037,
      "tokens_seen": 2402091008
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1319376,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.870497226715088,
      "objective/train/theoretical_loss": 3.379307797838357,
      "objective/train/tokens_used": 761015776,
      "theoretical_loss": 3.379307797838357,
      "tokens_seen": 2402156544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767878787878788,
      "loss": 2.4212,
      "theoretical_loss": 3.379307797838357,
      "tokens_seen": 2402156544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027676767676767676,
      "loss": 2.8379,
      "theoretical_loss": 3.3793003849849805,
      "tokens_seen": 2402222080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027674747474747477,
      "loss": 2.6259,
      "theoretical_loss": 3.3792929723904583,
      "tokens_seen": 2402287616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767272727272727,
      "loss": 2.3961,
      "theoretical_loss": 3.379285560054773,
      "tokens_seen": 2402353152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027670707070707074,
      "loss": 2.5514,
      "theoretical_loss": 3.37927814797791,
      "tokens_seen": 2402418688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766868686868687,
      "loss": 2.4595,
      "theoretical_loss": 3.379270736159852,
      "tokens_seen": 2402484224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027666666666666665,
      "loss": 2.629,
      "theoretical_loss": 3.3792633246005837,
      "tokens_seen": 2402549760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027664646464646466,
      "loss": 2.3439,
      "theoretical_loss": 3.3792559133000886,
      "tokens_seen": 2402615296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766262626262626,
      "loss": 2.5871,
      "theoretical_loss": 3.379248502258351,
      "tokens_seen": 2402680832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027660606060606063,
      "loss": 2.5348,
      "theoretical_loss": 3.3792410914753543,
      "tokens_seen": 2402746368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765858585858586,
      "loss": 2.7453,
      "theoretical_loss": 3.379233680951083,
      "tokens_seen": 2402811904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765656565656566,
      "loss": 2.3807,
      "theoretical_loss": 3.37922627068552,
      "tokens_seen": 2402877440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027654545454545456,
      "loss": 2.3371,
      "theoretical_loss": 3.37921886067865,
      "tokens_seen": 2402942976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765252525252525,
      "loss": 2.4897,
      "theoretical_loss": 3.379211450930457,
      "tokens_seen": 2403008512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765050505050505,
      "loss": 2.6656,
      "theoretical_loss": 3.379204041440925,
      "tokens_seen": 2403074048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764848484848485,
      "loss": 2.7301,
      "theoretical_loss": 3.3791966322100375,
      "tokens_seen": 2403139584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764646464646465,
      "loss": 2.6037,
      "theoretical_loss": 3.3791892232377783,
      "tokens_seen": 2403205120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027644444444444445,
      "loss": 2.6713,
      "theoretical_loss": 3.3791818145241317,
      "tokens_seen": 2403270656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764242424242424,
      "loss": 2.5729,
      "theoretical_loss": 3.3791744060690814,
      "tokens_seen": 2403336192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764040404040404,
      "loss": 2.5159,
      "theoretical_loss": 3.3791669978726113,
      "tokens_seen": 2403401728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763838383838384,
      "loss": 2.4243,
      "theoretical_loss": 3.379159589934706,
      "tokens_seen": 2403467264
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763636363636364,
      "loss": 2.46,
      "theoretical_loss": 3.379152182255348,
      "tokens_seen": 2403532800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027634343434343434,
      "loss": 2.5554,
      "theoretical_loss": 3.379144774834523,
      "tokens_seen": 2403598336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763232323232323,
      "loss": 2.7471,
      "theoretical_loss": 3.379137367672213,
      "tokens_seen": 2403663872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763030303030303,
      "loss": 2.431,
      "theoretical_loss": 3.3791299607684038,
      "tokens_seen": 2403729408
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1320625,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8724403381347656,
      "objective/train/theoretical_loss": 3.3791225541230783,
      "objective/train/tokens_used": 762654176,
      "theoretical_loss": 3.3791225541230783,
      "tokens_seen": 2403794944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762828282828283,
      "loss": 2.8231,
      "theoretical_loss": 3.3791225541230783,
      "tokens_seen": 2403794944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762626262626263,
      "loss": 2.6772,
      "theoretical_loss": 3.3791151477362202,
      "tokens_seen": 2403860480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027624242424242423,
      "loss": 2.6169,
      "theoretical_loss": 3.379107741607814,
      "tokens_seen": 2403926016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762222222222222,
      "loss": 2.5812,
      "theoretical_loss": 3.3791003357378435,
      "tokens_seen": 2403991552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027620202020202026,
      "loss": 2.0871,
      "theoretical_loss": 3.379092930126293,
      "tokens_seen": 2404057088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761818181818182,
      "loss": 2.6355,
      "theoretical_loss": 3.3790855247731457,
      "tokens_seen": 2404122624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027616161616161617,
      "loss": 2.5493,
      "theoretical_loss": 3.379078119678386,
      "tokens_seen": 2404188160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761414141414141,
      "loss": 2.6729,
      "theoretical_loss": 3.3790707148419976,
      "tokens_seen": 2404253696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761212121212121,
      "loss": 2.4889,
      "theoretical_loss": 3.3790633102639647,
      "tokens_seen": 2404319232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027610101010101015,
      "loss": 2.5127,
      "theoretical_loss": 3.379055905944271,
      "tokens_seen": 2404384768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002760808080808081,
      "loss": 2.6958,
      "theoretical_loss": 3.3790485018829006,
      "tokens_seen": 2404450304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027606060606060606,
      "loss": 2.5552,
      "theoretical_loss": 3.3790410980798375,
      "tokens_seen": 2404515840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276040404040404,
      "loss": 2.64,
      "theoretical_loss": 3.379033694535065,
      "tokens_seen": 2404581376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027602020202020203,
      "loss": 2.6541,
      "theoretical_loss": 3.379026291248568,
      "tokens_seen": 2404646912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027600000000000004,
      "loss": 2.3945,
      "theoretical_loss": 3.3790188882203305,
      "tokens_seen": 2404712448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275979797979798,
      "loss": 2.486,
      "theoretical_loss": 3.3790114854503357,
      "tokens_seen": 2404777984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027595959595959595,
      "loss": 2.5952,
      "theoretical_loss": 3.3790040829385677,
      "tokens_seen": 2404843520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002759393939393939,
      "loss": 2.3296,
      "theoretical_loss": 3.378996680685011,
      "tokens_seen": 2404909056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002759191919191919,
      "loss": 2.6235,
      "theoretical_loss": 3.378989278689649,
      "tokens_seen": 2404974592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027589898989898993,
      "loss": 2.3744,
      "theoretical_loss": 3.3789818769524658,
      "tokens_seen": 2405040128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758787878787879,
      "loss": 2.5835,
      "theoretical_loss": 3.378974475473445,
      "tokens_seen": 2405105664
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027585858585858585,
      "loss": 2.5668,
      "theoretical_loss": 3.3789670742525715,
      "tokens_seen": 2405171200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027583838383838386,
      "loss": 2.5913,
      "theoretical_loss": 3.3789596732898284,
      "tokens_seen": 2405236736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758181818181818,
      "loss": 2.4548,
      "theoretical_loss": 3.3789522725852,
      "tokens_seen": 2405302272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757979797979798,
      "loss": 2.6723,
      "theoretical_loss": 3.3789448721386703,
      "tokens_seen": 2405367808
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1321139,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0456953048706055,
      "objective/train/theoretical_loss": 3.3789374719502234,
      "objective/train/tokens_used": 764292576,
      "theoretical_loss": 3.3789374719502234,
      "tokens_seen": 2405433344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757777777777778,
      "loss": 2.7444,
      "theoretical_loss": 3.3789374719502234,
      "tokens_seen": 2405433344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027575757575757574,
      "loss": 2.4849,
      "theoretical_loss": 3.378930072019843,
      "tokens_seen": 2405498880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027573737373737375,
      "loss": 2.5797,
      "theoretical_loss": 3.378922672347513,
      "tokens_seen": 2405564416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757171717171717,
      "loss": 2.4306,
      "theoretical_loss": 3.378915272933218,
      "tokens_seen": 2405629952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756969696969697,
      "loss": 2.3946,
      "theoretical_loss": 3.3789078737769414,
      "tokens_seen": 2405695488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756767676767677,
      "loss": 2.5311,
      "theoretical_loss": 3.378900474878667,
      "tokens_seen": 2405761024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756565656565657,
      "loss": 2.5221,
      "theoretical_loss": 3.378893076238379,
      "tokens_seen": 2405826560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027563636363636364,
      "loss": 2.5755,
      "theoretical_loss": 3.3788856778560614,
      "tokens_seen": 2405892096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756161616161616,
      "loss": 2.7056,
      "theoretical_loss": 3.3788782797316985,
      "tokens_seen": 2405957632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755959595959596,
      "loss": 2.4726,
      "theoretical_loss": 3.3788708818652737,
      "tokens_seen": 2406023168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027557575757575757,
      "loss": 2.5859,
      "theoretical_loss": 3.3788634842567715,
      "tokens_seen": 2406088704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755555555555556,
      "loss": 2.5748,
      "theoretical_loss": 3.3788560869061754,
      "tokens_seen": 2406154240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027553535353535354,
      "loss": 2.4301,
      "theoretical_loss": 3.37884868981347,
      "tokens_seen": 2406219776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755151515151515,
      "loss": 2.4314,
      "theoretical_loss": 3.3788412929786382,
      "tokens_seen": 2406285312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754949494949495,
      "loss": 2.5033,
      "theoretical_loss": 3.3788338964016655,
      "tokens_seen": 2406350848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754747474747475,
      "loss": 2.5634,
      "theoretical_loss": 3.3788265000825346,
      "tokens_seen": 2406416384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027545454545454547,
      "loss": 2.6706,
      "theoretical_loss": 3.3788191040212303,
      "tokens_seen": 2406481920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027543434343434343,
      "loss": 2.5356,
      "theoretical_loss": 3.3788117082177362,
      "tokens_seen": 2406547456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754141414141414,
      "loss": 2.4759,
      "theoretical_loss": 3.3788043126720364,
      "tokens_seen": 2406612992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753939393939394,
      "loss": 2.2995,
      "theoretical_loss": 3.378796917384115,
      "tokens_seen": 2406678528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753737373737374,
      "loss": 2.5573,
      "theoretical_loss": 3.3787895223539555,
      "tokens_seen": 2406744064
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027535353535353537,
      "loss": 2.5764,
      "theoretical_loss": 3.3787821275815424,
      "tokens_seen": 2406809600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753333333333333,
      "loss": 2.4525,
      "theoretical_loss": 3.3787747330668596,
      "tokens_seen": 2406875136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753131313131313,
      "loss": 2.7627,
      "theoretical_loss": 3.378767338809891,
      "tokens_seen": 2406940672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027529292929292934,
      "loss": 2.5667,
      "theoretical_loss": 3.378759944810621,
      "tokens_seen": 2407006208
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 1322223,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9623414278030396,
      "objective/train/theoretical_loss": 3.378752551069033,
      "objective/train/tokens_used": 765930976,
      "theoretical_loss": 3.378752551069033,
      "tokens_seen": 2407071744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752727272727273,
      "loss": 2.3891,
      "theoretical_loss": 3.378752551069033,
      "tokens_seen": 2407071744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027525252525252526,
      "loss": 2.5254,
      "theoretical_loss": 3.3787451575851115,
      "tokens_seen": 2407137280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752323232323232,
      "loss": 2.5511,
      "theoretical_loss": 3.37873776435884,
      "tokens_seen": 2407202816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027521212121212117,
      "loss": 2.5441,
      "theoretical_loss": 3.3787303713902026,
      "tokens_seen": 2407268352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027519191919191924,
      "loss": 2.5071,
      "theoretical_loss": 3.3787229786791837,
      "tokens_seen": 2407333888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751717171717172,
      "loss": 2.6239,
      "theoretical_loss": 3.378715586225767,
      "tokens_seen": 2407399424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027515151515151515,
      "loss": 2.5842,
      "theoretical_loss": 3.3787081940299366,
      "tokens_seen": 2407464960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751313131313131,
      "loss": 2.8004,
      "theoretical_loss": 3.378700802091677,
      "tokens_seen": 2407530496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027511111111111106,
      "loss": 2.638,
      "theoretical_loss": 3.378693410410971,
      "tokens_seen": 2407596032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027509090909090913,
      "loss": 2.6841,
      "theoretical_loss": 3.378686018987804,
      "tokens_seen": 2407661568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750707070707071,
      "loss": 2.6142,
      "theoretical_loss": 3.378678627822159,
      "tokens_seen": 2407727104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027505050505050504,
      "loss": 2.6844,
      "theoretical_loss": 3.3786712369140206,
      "tokens_seen": 2407792640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000275030303030303,
      "loss": 2.6938,
      "theoretical_loss": 3.3786638462633722,
      "tokens_seen": 2407858176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000275010101010101,
      "loss": 2.3917,
      "theoretical_loss": 3.378656455870199,
      "tokens_seen": 2407923712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274989898989899,
      "loss": 2.4709,
      "theoretical_loss": 3.3786490657344834,
      "tokens_seen": 2407989248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274969696969697,
      "loss": 2.4547,
      "theoretical_loss": 3.378641675856211,
      "tokens_seen": 2408054784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027494949494949494,
      "loss": 2.4256,
      "theoretical_loss": 3.378634286235364,
      "tokens_seen": 2408120320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002749292929292929,
      "loss": 2.5933,
      "theoretical_loss": 3.378626896871929,
      "tokens_seen": 2408185856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027490909090909096,
      "loss": 2.398,
      "theoretical_loss": 3.3786195077658876,
      "tokens_seen": 2408251392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002748888888888889,
      "loss": 2.586,
      "theoretical_loss": 3.378612118917225,
      "tokens_seen": 2408316928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027486868686868687,
      "loss": 2.5633,
      "theoretical_loss": 3.3786047303259252,
      "tokens_seen": 2408382464
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027484848484848483,
      "loss": 2.6146,
      "theoretical_loss": 3.3785973419919717,
      "tokens_seen": 2408448000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027482828282828284,
      "loss": 2.4238,
      "theoretical_loss": 3.378589953915349,
      "tokens_seen": 2408513536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027480808080808085,
      "loss": 2.5799,
      "theoretical_loss": 3.3785825660960414,
      "tokens_seen": 2408579072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747878787878788,
      "loss": 2.6225,
      "theoretical_loss": 3.378575178534032,
      "tokens_seen": 2408644608
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1322701,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3897078037261963,
      "objective/train/theoretical_loss": 3.378567791229306,
      "objective/train/tokens_used": 767569376,
      "theoretical_loss": 3.378567791229306,
      "tokens_seen": 2408710144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027476767676767676,
      "loss": 2.4039,
      "theoretical_loss": 3.378567791229306,
      "tokens_seen": 2408710144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747474747474747,
      "loss": 2.4518,
      "theoretical_loss": 3.3785604041818464,
      "tokens_seen": 2408775680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027472727272727273,
      "loss": 2.6513,
      "theoretical_loss": 3.3785530173916376,
      "tokens_seen": 2408841216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027470707070707074,
      "loss": 2.5527,
      "theoretical_loss": 3.3785456308586643,
      "tokens_seen": 2408906752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746868686868687,
      "loss": 2.5835,
      "theoretical_loss": 3.3785382445829093,
      "tokens_seen": 2408972288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027466666666666666,
      "loss": 2.5775,
      "theoretical_loss": 3.378530858564358,
      "tokens_seen": 2409037824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027464646464646467,
      "loss": 2.6796,
      "theoretical_loss": 3.378523472802993,
      "tokens_seen": 2409103360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746262626262626,
      "loss": 2.7412,
      "theoretical_loss": 3.3785160872987996,
      "tokens_seen": 2409168896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027460606060606064,
      "loss": 2.4294,
      "theoretical_loss": 3.3785087020517612,
      "tokens_seen": 2409234432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745858585858586,
      "loss": 2.4665,
      "theoretical_loss": 3.378501317061862,
      "tokens_seen": 2409299968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027456565656565655,
      "loss": 2.6361,
      "theoretical_loss": 3.378493932329086,
      "tokens_seen": 2409365504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027454545454545456,
      "loss": 2.5521,
      "theoretical_loss": 3.3784865478534174,
      "tokens_seen": 2409431040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745252525252525,
      "loss": 2.6366,
      "theoretical_loss": 3.3784791636348404,
      "tokens_seen": 2409496576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027450505050505053,
      "loss": 2.7747,
      "theoretical_loss": 3.3784717796733386,
      "tokens_seen": 2409562112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744848484848485,
      "loss": 2.6529,
      "theoretical_loss": 3.378464395968896,
      "tokens_seen": 2409627648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744646464646465,
      "loss": 2.6161,
      "theoretical_loss": 3.3784570125214977,
      "tokens_seen": 2409693184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027444444444444445,
      "loss": 2.5866,
      "theoretical_loss": 3.3784496293311266,
      "tokens_seen": 2409758720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744242424242424,
      "loss": 2.6144,
      "theoretical_loss": 3.378442246397767,
      "tokens_seen": 2409824256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744040404040404,
      "loss": 2.5056,
      "theoretical_loss": 3.3784348637214032,
      "tokens_seen": 2409889792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743838383838384,
      "loss": 2.6486,
      "theoretical_loss": 3.3784274813020194,
      "tokens_seen": 2409955328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743636363636364,
      "loss": 2.8171,
      "theoretical_loss": 3.3784200991395994,
      "tokens_seen": 2410020864
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027434343434343435,
      "loss": 2.5236,
      "theoretical_loss": 3.378412717234127,
      "tokens_seen": 2410086400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743232323232323,
      "loss": 2.6253,
      "theoretical_loss": 3.3784053355855868,
      "tokens_seen": 2410151936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743030303030303,
      "loss": 2.4956,
      "theoretical_loss": 3.378397954193963,
      "tokens_seen": 2410217472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742828282828283,
      "loss": 2.5192,
      "theoretical_loss": 3.378390573059239,
      "tokens_seen": 2410283008
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1324080,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.414530038833618,
      "objective/train/theoretical_loss": 3.3783831921813996,
      "objective/train/tokens_used": 769207776,
      "theoretical_loss": 3.3783831921813996,
      "tokens_seen": 2410348544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742626262626263,
      "loss": 2.3967,
      "theoretical_loss": 3.3783831921813996,
      "tokens_seen": 2410348544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027424242424242424,
      "loss": 2.5257,
      "theoretical_loss": 3.378375811560428,
      "tokens_seen": 2410414080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742222222222222,
      "loss": 2.6932,
      "theoretical_loss": 3.378368431196309,
      "tokens_seen": 2410479616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742020202020202,
      "loss": 2.4726,
      "theoretical_loss": 3.3783610510890267,
      "tokens_seen": 2410545152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741818181818182,
      "loss": 2.5827,
      "theoretical_loss": 3.3783536712385644,
      "tokens_seen": 2410610688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741616161616162,
      "loss": 2.7406,
      "theoretical_loss": 3.3783462916449074,
      "tokens_seen": 2410676224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027414141414141413,
      "loss": 2.5695,
      "theoretical_loss": 3.3783389123080387,
      "tokens_seen": 2410741760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741212121212121,
      "loss": 2.68,
      "theoretical_loss": 3.378331533227943,
      "tokens_seen": 2410807296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027410101010101015,
      "loss": 2.6111,
      "theoretical_loss": 3.378324154404604,
      "tokens_seen": 2410872832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740808080808081,
      "loss": 2.5358,
      "theoretical_loss": 3.3783167758380057,
      "tokens_seen": 2410938368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027406060606060607,
      "loss": 2.6948,
      "theoretical_loss": 3.378309397528133,
      "tokens_seen": 2411003904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274040404040404,
      "loss": 2.4084,
      "theoretical_loss": 3.378302019474969,
      "tokens_seen": 2411069440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274020202020202,
      "loss": 2.5457,
      "theoretical_loss": 3.3782946416784982,
      "tokens_seen": 2411134976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027400000000000005,
      "loss": 2.4703,
      "theoretical_loss": 3.378287264138705,
      "tokens_seen": 2411200512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273979797979798,
      "loss": 2.5908,
      "theoretical_loss": 3.378279886855573,
      "tokens_seen": 2411266048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027395959595959596,
      "loss": 2.7809,
      "theoretical_loss": 3.3782725098290864,
      "tokens_seen": 2411331584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002739393939393939,
      "loss": 2.6265,
      "theoretical_loss": 3.3782651330592297,
      "tokens_seen": 2411397120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027391919191919193,
      "loss": 2.6918,
      "theoretical_loss": 3.3782577565459864,
      "tokens_seen": 2411462656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027389898989898994,
      "loss": 2.6622,
      "theoretical_loss": 3.3782503802893413,
      "tokens_seen": 2411528192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738787878787879,
      "loss": 2.5508,
      "theoretical_loss": 3.3782430042892777,
      "tokens_seen": 2411593728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027385858585858585,
      "loss": 2.4676,
      "theoretical_loss": 3.3782356285457804,
      "tokens_seen": 2411659264
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738383838383838,
      "loss": 2.4347,
      "theoretical_loss": 3.3782282530588335,
      "tokens_seen": 2411724800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738181818181818,
      "loss": 2.5036,
      "theoretical_loss": 3.37822087782842,
      "tokens_seen": 2411790336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027379797979797983,
      "loss": 2.7887,
      "theoretical_loss": 3.3782135028545253,
      "tokens_seen": 2411855872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737777777777778,
      "loss": 2.6072,
      "theoretical_loss": 3.378206128137133,
      "tokens_seen": 2411921408
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1324705,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.521892786026001,
      "objective/train/theoretical_loss": 3.3781987536762275,
      "objective/train/tokens_used": 770846176,
      "theoretical_loss": 3.3781987536762275,
      "tokens_seen": 2411986944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027375757575757575,
      "loss": 2.505,
      "theoretical_loss": 3.3781987536762275,
      "tokens_seen": 2411986944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737373737373737,
      "loss": 2.5708,
      "theoretical_loss": 3.378191379471792,
      "tokens_seen": 2412052480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737171717171717,
      "loss": 2.592,
      "theoretical_loss": 3.3781840055238117,
      "tokens_seen": 2412118016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736969696969697,
      "loss": 2.5439,
      "theoretical_loss": 3.3781766318322703,
      "tokens_seen": 2412183552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736767676767677,
      "loss": 2.3507,
      "theoretical_loss": 3.3781692583971514,
      "tokens_seen": 2412249088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027365656565656564,
      "loss": 2.5583,
      "theoretical_loss": 3.3781618852184403,
      "tokens_seen": 2412314624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027363636363636365,
      "loss": 2.5696,
      "theoretical_loss": 3.3781545122961205,
      "tokens_seen": 2412380160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736161616161616,
      "loss": 2.6648,
      "theoretical_loss": 3.378147139630175,
      "tokens_seen": 2412445696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735959595959596,
      "loss": 2.5808,
      "theoretical_loss": 3.37813976722059,
      "tokens_seen": 2412511232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735757575757576,
      "loss": 2.451,
      "theoretical_loss": 3.3781323950673485,
      "tokens_seen": 2412576768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027355555555555553,
      "loss": 2.6893,
      "theoretical_loss": 3.3781250231704343,
      "tokens_seen": 2412642304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027353535353535354,
      "loss": 2.4823,
      "theoretical_loss": 3.378117651529832,
      "tokens_seen": 2412707840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735151515151515,
      "loss": 2.4537,
      "theoretical_loss": 3.378110280145526,
      "tokens_seen": 2412773376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734949494949495,
      "loss": 2.5139,
      "theoretical_loss": 3.3781029090174997,
      "tokens_seen": 2412838912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027347474747474747,
      "loss": 2.7567,
      "theoretical_loss": 3.378095538145738,
      "tokens_seen": 2412904448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734545454545455,
      "loss": 2.6232,
      "theoretical_loss": 3.3780881675302243,
      "tokens_seen": 2412969984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027343434343434343,
      "loss": 2.6075,
      "theoretical_loss": 3.3780807971709432,
      "tokens_seen": 2413035520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734141414141414,
      "loss": 2.6903,
      "theoretical_loss": 3.3780734270678785,
      "tokens_seen": 2413101056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733939393939394,
      "loss": 2.7068,
      "theoretical_loss": 3.378066057221015,
      "tokens_seen": 2413166592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027337373737373736,
      "loss": 2.2726,
      "theoretical_loss": 3.378058687630336,
      "tokens_seen": 2413232128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027335353535353537,
      "loss": 2.4803,
      "theoretical_loss": 3.3780513182958263,
      "tokens_seen": 2413297664
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027333333333333333,
      "loss": 2.3323,
      "theoretical_loss": 3.3780439492174694,
      "tokens_seen": 2413363200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733131313131313,
      "loss": 2.7783,
      "theoretical_loss": 3.3780365803952503,
      "tokens_seen": 2413428736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732929292929293,
      "loss": 2.5186,
      "theoretical_loss": 3.378029211829152,
      "tokens_seen": 2413494272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732727272727273,
      "loss": 2.397,
      "theoretical_loss": 3.3780218435191602,
      "tokens_seen": 2413559808
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1325945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0629208087921143,
      "objective/train/theoretical_loss": 3.3780144754652577,
      "objective/train/tokens_used": 772484576,
      "theoretical_loss": 3.3780144754652577,
      "tokens_seen": 2413625344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027325252525252526,
      "loss": 2.4717,
      "theoretical_loss": 3.3780144754652577,
      "tokens_seen": 2413625344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732323232323232,
      "loss": 2.7437,
      "theoretical_loss": 3.378007107667429,
      "tokens_seen": 2413690880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027321212121212123,
      "loss": 2.544,
      "theoretical_loss": 3.3779997401256585,
      "tokens_seen": 2413756416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731919191919192,
      "loss": 2.8167,
      "theoretical_loss": 3.37799237283993,
      "tokens_seen": 2413821952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731717171717172,
      "loss": 2.6599,
      "theoretical_loss": 3.3779850058102276,
      "tokens_seen": 2413887488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027315151515151516,
      "loss": 2.6438,
      "theoretical_loss": 3.377977639036536,
      "tokens_seen": 2413953024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731313131313131,
      "loss": 2.7503,
      "theoretical_loss": 3.377970272518839,
      "tokens_seen": 2414018560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731111111111111,
      "loss": 2.6223,
      "theoretical_loss": 3.377962906257121,
      "tokens_seen": 2414084096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027309090909090913,
      "loss": 2.6295,
      "theoretical_loss": 3.3779555402513655,
      "tokens_seen": 2414149632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002730707070707071,
      "loss": 2.5036,
      "theoretical_loss": 3.3779481745015576,
      "tokens_seen": 2414215168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027305050505050505,
      "loss": 2.7458,
      "theoretical_loss": 3.3779408090076806,
      "tokens_seen": 2414280704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273030303030303,
      "loss": 2.317,
      "theoretical_loss": 3.3779334437697193,
      "tokens_seen": 2414346240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273010101010101,
      "loss": 2.716,
      "theoretical_loss": 3.3779260787876573,
      "tokens_seen": 2414411776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027298989898989903,
      "loss": 2.5639,
      "theoretical_loss": 3.3779187140614795,
      "tokens_seen": 2414477312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272969696969697,
      "loss": 2.8078,
      "theoretical_loss": 3.377911349591169,
      "tokens_seen": 2414542848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027294949494949494,
      "loss": 2.3878,
      "theoretical_loss": 3.3779039853767108,
      "tokens_seen": 2414608384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729292929292929,
      "loss": 2.434,
      "theoretical_loss": 3.377896621418089,
      "tokens_seen": 2414673920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027290909090909096,
      "loss": 2.5527,
      "theoretical_loss": 3.377889257715288,
      "tokens_seen": 2414739456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728888888888889,
      "loss": 2.5444,
      "theoretical_loss": 3.377881894268291,
      "tokens_seen": 2414804992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728686868686869,
      "loss": 2.5772,
      "theoretical_loss": 3.3778745310770826,
      "tokens_seen": 2414870528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027284848484848483,
      "loss": 2.498,
      "theoretical_loss": 3.3778671681416474,
      "tokens_seen": 2414936064
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728282828282828,
      "loss": 2.4699,
      "theoretical_loss": 3.3778598054619695,
      "tokens_seen": 2415001600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027280808080808086,
      "loss": 2.436,
      "theoretical_loss": 3.377852443038033,
      "tokens_seen": 2415067136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727878787878788,
      "loss": 2.4811,
      "theoretical_loss": 3.3778450808698213,
      "tokens_seen": 2415132672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027276767676767677,
      "loss": 2.6489,
      "theoretical_loss": 3.37783771895732,
      "tokens_seen": 2415198208
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1326379,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.516770601272583,
      "objective/train/theoretical_loss": 3.377830357300512,
      "objective/train/tokens_used": 774122976,
      "theoretical_loss": 3.377830357300512,
      "tokens_seen": 2415263744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727474747474747,
      "loss": 2.5568,
      "theoretical_loss": 3.377830357300512,
      "tokens_seen": 2415263744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027272727272727274,
      "loss": 2.5463,
      "theoretical_loss": 3.377822995899382,
      "tokens_seen": 2415329280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027270707070707075,
      "loss": 2.7598,
      "theoretical_loss": 3.3778156347539143,
      "tokens_seen": 2415394816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002726868686868687,
      "loss": 2.5923,
      "theoretical_loss": 3.377808273864093,
      "tokens_seen": 2415460352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027266666666666666,
      "loss": 2.673,
      "theoretical_loss": 3.377800913229902,
      "tokens_seen": 2415525888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002726464646464646,
      "loss": 2.3286,
      "theoretical_loss": 3.377793552851326,
      "tokens_seen": 2415591424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027262626262626263,
      "loss": 2.6616,
      "theoretical_loss": 3.3777861927283492,
      "tokens_seen": 2415656960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027260606060606064,
      "loss": 2.4835,
      "theoretical_loss": 3.3777788328609555,
      "tokens_seen": 2415722496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725858585858586,
      "loss": 2.6633,
      "theoretical_loss": 3.377771473249129,
      "tokens_seen": 2415788032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027256565656565656,
      "loss": 2.5684,
      "theoretical_loss": 3.3777641138928534,
      "tokens_seen": 2415853568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027254545454545457,
      "loss": 2.5802,
      "theoretical_loss": 3.377756754792114,
      "tokens_seen": 2415919104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725252525252525,
      "loss": 2.7372,
      "theoretical_loss": 3.3777493959468945,
      "tokens_seen": 2415984640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027250505050505053,
      "loss": 2.6714,
      "theoretical_loss": 3.377742037357179,
      "tokens_seen": 2416050176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724848484848485,
      "loss": 2.7078,
      "theoretical_loss": 3.377734679022952,
      "tokens_seen": 2416115712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027246464646464645,
      "loss": 2.7154,
      "theoretical_loss": 3.3777273209441976,
      "tokens_seen": 2416181248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027244444444444446,
      "loss": 2.4861,
      "theoretical_loss": 3.3777199631208994,
      "tokens_seen": 2416246784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724242424242424,
      "loss": 2.5344,
      "theoretical_loss": 3.377712605553042,
      "tokens_seen": 2416312320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724040404040404,
      "loss": 2.5755,
      "theoretical_loss": 3.3777052482406105,
      "tokens_seen": 2416377856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723838383838384,
      "loss": 2.4513,
      "theoretical_loss": 3.3776978911835878,
      "tokens_seen": 2416443392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723636363636364,
      "loss": 2.4515,
      "theoretical_loss": 3.3776905343819585,
      "tokens_seen": 2416508928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027234343434343435,
      "loss": 2.2856,
      "theoretical_loss": 3.3776831778357073,
      "tokens_seen": 2416574464
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723232323232323,
      "loss": 2.4504,
      "theoretical_loss": 3.377675821544818,
      "tokens_seen": 2416640000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723030303030303,
      "loss": 2.529,
      "theoretical_loss": 3.3776684655092746,
      "tokens_seen": 2416705536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722828282828283,
      "loss": 2.4591,
      "theoretical_loss": 3.377661109729061,
      "tokens_seen": 2416771072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722626262626263,
      "loss": 2.3674,
      "theoretical_loss": 3.3776537542041627,
      "tokens_seen": 2416836608
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1327590,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.714198112487793,
      "objective/train/theoretical_loss": 3.377646398934563,
      "objective/train/tokens_used": 775761376,
      "theoretical_loss": 3.377646398934563,
      "tokens_seen": 2416902144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027224242424242424,
      "loss": 2.4671,
      "theoretical_loss": 3.377646398934563,
      "tokens_seen": 2416902144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722222222222222,
      "loss": 2.6425,
      "theoretical_loss": 3.3776390439202464,
      "tokens_seen": 2416967680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722020202020202,
      "loss": 2.6495,
      "theoretical_loss": 3.377631689161197,
      "tokens_seen": 2417033216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721818181818182,
      "loss": 2.6031,
      "theoretical_loss": 3.377624334657399,
      "tokens_seen": 2417098752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721616161616162,
      "loss": 2.4543,
      "theoretical_loss": 3.3776169804088365,
      "tokens_seen": 2417164288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027214141414141414,
      "loss": 2.6633,
      "theoretical_loss": 3.3776096264154942,
      "tokens_seen": 2417229824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721212121212121,
      "loss": 2.6781,
      "theoretical_loss": 3.377602272677356,
      "tokens_seen": 2417295360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721010101010101,
      "loss": 2.4223,
      "theoretical_loss": 3.3775949191944057,
      "tokens_seen": 2417360896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720808080808081,
      "loss": 2.5522,
      "theoretical_loss": 3.3775875659666283,
      "tokens_seen": 2417426432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720606060606061,
      "loss": 2.4789,
      "theoretical_loss": 3.3775802129940073,
      "tokens_seen": 2417491968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027204040404040403,
      "loss": 2.4614,
      "theoretical_loss": 3.377572860276528,
      "tokens_seen": 2417557504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272020202020202,
      "loss": 2.7,
      "theoretical_loss": 3.3775655078141735,
      "tokens_seen": 2417623040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027200000000000005,
      "loss": 2.5628,
      "theoretical_loss": 3.3775581556069283,
      "tokens_seen": 2417688576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271979797979798,
      "loss": 2.4734,
      "theoretical_loss": 3.3775508036547772,
      "tokens_seen": 2417754112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027195959595959597,
      "loss": 2.5551,
      "theoretical_loss": 3.377543451957704,
      "tokens_seen": 2417819648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719393939393939,
      "loss": 2.7689,
      "theoretical_loss": 3.3775361005156928,
      "tokens_seen": 2417885184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719191919191919,
      "loss": 2.4865,
      "theoretical_loss": 3.377528749328728,
      "tokens_seen": 2417950720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027189898989898994,
      "loss": 2.5239,
      "theoretical_loss": 3.377521398396794,
      "tokens_seen": 2418016256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718787878787879,
      "loss": 2.6182,
      "theoretical_loss": 3.377514047719875,
      "tokens_seen": 2418081792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027185858585858586,
      "loss": 2.406,
      "theoretical_loss": 3.377506697297955,
      "tokens_seen": 2418147328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718383838383838,
      "loss": 2.6016,
      "theoretical_loss": 3.377499347131018,
      "tokens_seen": 2418212864
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027181818181818177,
      "loss": 2.4279,
      "theoretical_loss": 3.377491997219049,
      "tokens_seen": 2418278400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027179797979797984,
      "loss": 2.2922,
      "theoretical_loss": 3.377484647562032,
      "tokens_seen": 2418343936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717777777777778,
      "loss": 2.5436,
      "theoretical_loss": 3.377477298159951,
      "tokens_seen": 2418409472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027175757575757575,
      "loss": 2.4884,
      "theoretical_loss": 3.37746994901279,
      "tokens_seen": 2418475008
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1328863,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2463276386260986,
      "objective/train/theoretical_loss": 3.377462600120534,
      "objective/train/tokens_used": 777399776,
      "theoretical_loss": 3.377462600120534,
      "tokens_seen": 2418540544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717373737373737,
      "loss": 2.4146,
      "theoretical_loss": 3.377462600120534,
      "tokens_seen": 2418540544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717171717171717,
      "loss": 2.2153,
      "theoretical_loss": 3.377455251483167,
      "tokens_seen": 2418606080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027169696969696973,
      "loss": 2.5368,
      "theoretical_loss": 3.377447903100673,
      "tokens_seen": 2418671616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716767676767677,
      "loss": 2.521,
      "theoretical_loss": 3.3774405549730364,
      "tokens_seen": 2418737152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027165656565656564,
      "loss": 2.5067,
      "theoretical_loss": 3.3774332071002413,
      "tokens_seen": 2418802688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716363636363636,
      "loss": 2.5375,
      "theoretical_loss": 3.377425859482272,
      "tokens_seen": 2418868224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716161616161616,
      "loss": 2.7116,
      "theoretical_loss": 3.3774185121191134,
      "tokens_seen": 2418933760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715959595959596,
      "loss": 2.4842,
      "theoretical_loss": 3.3774111650107486,
      "tokens_seen": 2418999296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715757575757576,
      "loss": 2.491,
      "theoretical_loss": 3.3774038181571626,
      "tokens_seen": 2419064832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027155555555555554,
      "loss": 2.6163,
      "theoretical_loss": 3.37739647155834,
      "tokens_seen": 2419130368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027153535353535355,
      "loss": 2.7782,
      "theoretical_loss": 3.3773891252142643,
      "tokens_seen": 2419195904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027151515151515156,
      "loss": 2.5928,
      "theoretical_loss": 3.37738177912492,
      "tokens_seen": 2419261440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714949494949495,
      "loss": 2.685,
      "theoretical_loss": 3.3773744332902913,
      "tokens_seen": 2419326976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027147474747474747,
      "loss": 2.6787,
      "theoretical_loss": 3.377367087710363,
      "tokens_seen": 2419392512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027145454545454543,
      "loss": 2.4557,
      "theoretical_loss": 3.3773597423851185,
      "tokens_seen": 2419458048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027143434343434344,
      "loss": 2.5975,
      "theoretical_loss": 3.3773523973145427,
      "tokens_seen": 2419523584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027141414141414145,
      "loss": 2.4962,
      "theoretical_loss": 3.3773450524986197,
      "tokens_seen": 2419589120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713939393939394,
      "loss": 2.5028,
      "theoretical_loss": 3.377337707937334,
      "tokens_seen": 2419654656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027137373737373737,
      "loss": 2.6093,
      "theoretical_loss": 3.3773303636306693,
      "tokens_seen": 2419720192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713535353535354,
      "loss": 2.7072,
      "theoretical_loss": 3.377323019578611,
      "tokens_seen": 2419785728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027133333333333333,
      "loss": 2.6711,
      "theoretical_loss": 3.377315675781142,
      "tokens_seen": 2419851264
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027131313131313134,
      "loss": 2.4018,
      "theoretical_loss": 3.3773083322382473,
      "tokens_seen": 2419916800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712929292929293,
      "loss": 2.5922,
      "theoretical_loss": 3.3773009889499113,
      "tokens_seen": 2419982336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027127272727272726,
      "loss": 2.664,
      "theoretical_loss": 3.377293645916118,
      "tokens_seen": 2420047872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027125252525252527,
      "loss": 2.4764,
      "theoretical_loss": 3.3772863031368514,
      "tokens_seen": 2420113408
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1329609,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3919801712036133,
      "objective/train/theoretical_loss": 3.377278960612096,
      "objective/train/tokens_used": 779038176,
      "theoretical_loss": 3.377278960612096,
      "tokens_seen": 2420178944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712323232323232,
      "loss": 2.421,
      "theoretical_loss": 3.377278960612096,
      "tokens_seen": 2420178944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027121212121212124,
      "loss": 2.6578,
      "theoretical_loss": 3.3772716183418368,
      "tokens_seen": 2420244480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711919191919192,
      "loss": 2.7039,
      "theoretical_loss": 3.3772642763260574,
      "tokens_seen": 2420310016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711717171717172,
      "loss": 2.6739,
      "theoretical_loss": 3.377256934564742,
      "tokens_seen": 2420375552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027115151515151516,
      "loss": 2.538,
      "theoretical_loss": 3.377249593057875,
      "tokens_seen": 2420441088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711313131313131,
      "loss": 2.4294,
      "theoretical_loss": 3.3772422518054412,
      "tokens_seen": 2420506624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027111111111111113,
      "loss": 2.7442,
      "theoretical_loss": 3.377234910807424,
      "tokens_seen": 2420572160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710909090909091,
      "loss": 2.5176,
      "theoretical_loss": 3.3772275700638086,
      "tokens_seen": 2420637696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710707070707071,
      "loss": 2.6513,
      "theoretical_loss": 3.3772202295745783,
      "tokens_seen": 2420703232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027105050505050505,
      "loss": 2.3484,
      "theoretical_loss": 3.3772128893397184,
      "tokens_seen": 2420768768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271030303030303,
      "loss": 2.4639,
      "theoretical_loss": 3.377205549359213,
      "tokens_seen": 2420834304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271010101010101,
      "loss": 2.6157,
      "theoretical_loss": 3.377198209633046,
      "tokens_seen": 2420899840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027098989898989903,
      "loss": 2.6086,
      "theoretical_loss": 3.3771908701612015,
      "tokens_seen": 2420965376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000270969696969697,
      "loss": 2.4896,
      "theoretical_loss": 3.3771835309436646,
      "tokens_seen": 2421030912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027094949494949495,
      "loss": 2.5821,
      "theoretical_loss": 3.3771761919804186,
      "tokens_seen": 2421096448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709292929292929,
      "loss": 2.6494,
      "theoretical_loss": 3.3771688532714492,
      "tokens_seen": 2421161984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709090909090909,
      "loss": 2.4321,
      "theoretical_loss": 3.377161514816739,
      "tokens_seen": 2421227520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708888888888889,
      "loss": 2.736,
      "theoretical_loss": 3.3771541766162736,
      "tokens_seen": 2421293056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708686868686869,
      "loss": 2.2884,
      "theoretical_loss": 3.377146838670037,
      "tokens_seen": 2421358592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027084848484848484,
      "loss": 2.7259,
      "theoretical_loss": 3.3771395009780134,
      "tokens_seen": 2421424128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708282828282828,
      "loss": 2.5679,
      "theoretical_loss": 3.377132163540187,
      "tokens_seen": 2421489664
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027080808080808086,
      "loss": 2.6837,
      "theoretical_loss": 3.3771248263565425,
      "tokens_seen": 2421555200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707878787878788,
      "loss": 2.702,
      "theoretical_loss": 3.377117489427064,
      "tokens_seen": 2421620736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707676767676768,
      "loss": 2.6824,
      "theoretical_loss": 3.3771101527517358,
      "tokens_seen": 2421686272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027074747474747473,
      "loss": 2.6854,
      "theoretical_loss": 3.377102816330542,
      "tokens_seen": 2421751808
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1330720,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.00944447517395,
      "objective/train/theoretical_loss": 3.377095480163467,
      "objective/train/tokens_used": 780676576,
      "theoretical_loss": 3.377095480163467,
      "tokens_seen": 2421817344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707272727272727,
      "loss": 2.5717,
      "theoretical_loss": 3.377095480163467,
      "tokens_seen": 2421817344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027070707070707075,
      "loss": 2.5181,
      "theoretical_loss": 3.3770881442504956,
      "tokens_seen": 2421882880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706868686868687,
      "loss": 2.5089,
      "theoretical_loss": 3.3770808085916113,
      "tokens_seen": 2421948416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027066666666666667,
      "loss": 2.1892,
      "theoretical_loss": 3.3770734731867993,
      "tokens_seen": 2422013952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706464646464646,
      "loss": 2.6041,
      "theoretical_loss": 3.377066138036043,
      "tokens_seen": 2422079488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027062626262626264,
      "loss": 2.5648,
      "theoretical_loss": 3.377058803139328,
      "tokens_seen": 2422145024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027060606060606065,
      "loss": 2.5677,
      "theoretical_loss": 3.3770514684966373,
      "tokens_seen": 2422210560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705858585858586,
      "loss": 2.5145,
      "theoretical_loss": 3.377044134107956,
      "tokens_seen": 2422276096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027056565656565656,
      "loss": 2.6456,
      "theoretical_loss": 3.3770367999732684,
      "tokens_seen": 2422341632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705454545454545,
      "loss": 2.267,
      "theoretical_loss": 3.377029466092558,
      "tokens_seen": 2422407168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027052525252525253,
      "loss": 2.7496,
      "theoretical_loss": 3.3770221324658105,
      "tokens_seen": 2422472704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027050505050505054,
      "loss": 2.5598,
      "theoretical_loss": 3.3770147990930095,
      "tokens_seen": 2422538240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704848484848485,
      "loss": 2.5307,
      "theoretical_loss": 3.377007465974139,
      "tokens_seen": 2422603776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027046464646464645,
      "loss": 2.6812,
      "theoretical_loss": 3.377000133109184,
      "tokens_seen": 2422669312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704444444444444,
      "loss": 2.6716,
      "theoretical_loss": 3.3769928004981282,
      "tokens_seen": 2422734848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704242424242424,
      "loss": 2.517,
      "theoretical_loss": 3.3769854681409566,
      "tokens_seen": 2422800384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027040404040404043,
      "loss": 2.4556,
      "theoretical_loss": 3.3769781360376534,
      "tokens_seen": 2422865920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703838383838384,
      "loss": 2.7594,
      "theoretical_loss": 3.3769708041882023,
      "tokens_seen": 2422931456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027036363636363635,
      "loss": 2.501,
      "theoretical_loss": 3.376963472592588,
      "tokens_seen": 2422996992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027034343434343436,
      "loss": 2.639,
      "theoretical_loss": 3.376956141250796,
      "tokens_seen": 2423062528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703232323232323,
      "loss": 2.5825,
      "theoretical_loss": 3.3769488101628085,
      "tokens_seen": 2423128064
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703030303030303,
      "loss": 2.6151,
      "theoretical_loss": 3.3769414793286114,
      "tokens_seen": 2423193600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002702828282828283,
      "loss": 2.5882,
      "theoretical_loss": 3.3769341487481888,
      "tokens_seen": 2423259136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027026262626262624,
      "loss": 2.6113,
      "theoretical_loss": 3.3769268184215244,
      "tokens_seen": 2423324672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027024242424242425,
      "loss": 2.4535,
      "theoretical_loss": 3.3769194883486033,
      "tokens_seen": 2423390208
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 1331401,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.588832378387451,
      "objective/train/theoretical_loss": 3.37691215852941,
      "objective/train/tokens_used": 782314976,
      "theoretical_loss": 3.37691215852941,
      "tokens_seen": 2423455744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702222222222222,
      "loss": 2.6272,
      "theoretical_loss": 3.37691215852941,
      "tokens_seen": 2423455744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702020202020202,
      "loss": 2.7104,
      "theoretical_loss": 3.376904828963928,
      "tokens_seen": 2423521280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701818181818182,
      "loss": 2.5889,
      "theoretical_loss": 3.376897499652142,
      "tokens_seen": 2423586816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701616161616162,
      "loss": 2.3941,
      "theoretical_loss": 3.376890170594036,
      "tokens_seen": 2423652352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027014141414141414,
      "loss": 2.5046,
      "theoretical_loss": 3.376882841789596,
      "tokens_seen": 2423717888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701212121212121,
      "loss": 2.5633,
      "theoretical_loss": 3.3768755132388044,
      "tokens_seen": 2423783424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701010101010101,
      "loss": 2.4163,
      "theoretical_loss": 3.3768681849416464,
      "tokens_seen": 2423848960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027008080808080807,
      "loss": 2.5344,
      "theoretical_loss": 3.3768608568981064,
      "tokens_seen": 2423914496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700606060606061,
      "loss": 2.4975,
      "theoretical_loss": 3.3768535291081685,
      "tokens_seen": 2423980032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027004040404040404,
      "loss": 2.7312,
      "theoretical_loss": 3.3768462015718175,
      "tokens_seen": 2424045568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000270020202020202,
      "loss": 2.3678,
      "theoretical_loss": 3.3768388742890374,
      "tokens_seen": 2424111104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027,
      "loss": 2.6491,
      "theoretical_loss": 3.3768315472598127,
      "tokens_seen": 2424176640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269979797979798,
      "loss": 2.5082,
      "theoretical_loss": 3.376824220484128,
      "tokens_seen": 2424242176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026995959595959597,
      "loss": 2.3295,
      "theoretical_loss": 3.376816893961967,
      "tokens_seen": 2424307712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026993939393939393,
      "loss": 2.6033,
      "theoretical_loss": 3.3768095676933143,
      "tokens_seen": 2424373248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699191919191919,
      "loss": 2.6762,
      "theoretical_loss": 3.376802241678155,
      "tokens_seen": 2424438784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698989898989899,
      "loss": 2.5041,
      "theoretical_loss": 3.3767949159164727,
      "tokens_seen": 2424504320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698787878787879,
      "loss": 2.4712,
      "theoretical_loss": 3.3767875904082523,
      "tokens_seen": 2424569856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026985858585858586,
      "loss": 2.3945,
      "theoretical_loss": 3.3767802651534775,
      "tokens_seen": 2424635392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698383838383838,
      "loss": 2.7693,
      "theoretical_loss": 3.376772940152133,
      "tokens_seen": 2424700928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026981818181818183,
      "loss": 2.6028,
      "theoretical_loss": 3.376765615404204,
      "tokens_seen": 2424766464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026979797979797984,
      "loss": 2.8442,
      "theoretical_loss": 3.376758290909673,
      "tokens_seen": 2424832000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697777777777778,
      "loss": 2.4963,
      "theoretical_loss": 3.3767509666685265,
      "tokens_seen": 2424897536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026975757575757576,
      "loss": 2.7412,
      "theoretical_loss": 3.3767436426807476,
      "tokens_seen": 2424963072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697373737373737,
      "loss": 2.6452,
      "theoretical_loss": 3.376736318946321,
      "tokens_seen": 2425028608
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1332594,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.556990146636963,
      "objective/train/theoretical_loss": 3.376728995465231,
      "objective/train/tokens_used": 783953376,
      "theoretical_loss": 3.376728995465231,
      "tokens_seen": 2425094144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697171717171717,
      "loss": 2.4698,
      "theoretical_loss": 3.376728995465231,
      "tokens_seen": 2425094144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026969696969696974,
      "loss": 2.4541,
      "theoretical_loss": 3.376721672237462,
      "tokens_seen": 2425159680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696767676767677,
      "loss": 2.4361,
      "theoretical_loss": 3.3767143492629987,
      "tokens_seen": 2425225216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026965656565656565,
      "loss": 2.3783,
      "theoretical_loss": 3.3767070265418253,
      "tokens_seen": 2425290752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696363636363636,
      "loss": 2.6316,
      "theoretical_loss": 3.3766997040739266,
      "tokens_seen": 2425356288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026961616161616167,
      "loss": 2.6758,
      "theoretical_loss": 3.3766923818592858,
      "tokens_seen": 2425421824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026959595959595963,
      "loss": 2.5955,
      "theoretical_loss": 3.3766850598978886,
      "tokens_seen": 2425487360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695757575757576,
      "loss": 2.5651,
      "theoretical_loss": 3.3766777381897186,
      "tokens_seen": 2425552896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026955555555555554,
      "loss": 2.5277,
      "theoretical_loss": 3.3766704167347603,
      "tokens_seen": 2425618432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695353535353535,
      "loss": 2.5461,
      "theoretical_loss": 3.3766630955329986,
      "tokens_seen": 2425683968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026951515151515156,
      "loss": 2.4433,
      "theoretical_loss": 3.3766557745844175,
      "tokens_seen": 2425749504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694949494949495,
      "loss": 2.6348,
      "theoretical_loss": 3.3766484538890014,
      "tokens_seen": 2425815040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694747474747475,
      "loss": 2.4347,
      "theoretical_loss": 3.3766411334467348,
      "tokens_seen": 2425880576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026945454545454543,
      "loss": 2.7299,
      "theoretical_loss": 3.376633813257602,
      "tokens_seen": 2425946112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026943434343434345,
      "loss": 2.5579,
      "theoretical_loss": 3.376626493321588,
      "tokens_seen": 2426011648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026941414141414146,
      "loss": 2.7931,
      "theoretical_loss": 3.376619173638676,
      "tokens_seen": 2426077184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002693939393939394,
      "loss": 2.6122,
      "theoretical_loss": 3.3766118542088517,
      "tokens_seen": 2426142720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026937373737373737,
      "loss": 2.6723,
      "theoretical_loss": 3.3766045350320986,
      "tokens_seen": 2426208256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026935353535353533,
      "loss": 2.3997,
      "theoretical_loss": 3.3765972161084012,
      "tokens_seen": 2426273792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026933333333333334,
      "loss": 2.5429,
      "theoretical_loss": 3.3765898974377446,
      "tokens_seen": 2426339328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026931313131313135,
      "loss": 2.6126,
      "theoretical_loss": 3.3765825790201127,
      "tokens_seen": 2426404864
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692929292929293,
      "loss": 2.7448,
      "theoretical_loss": 3.37657526085549,
      "tokens_seen": 2426470400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026927272727272726,
      "loss": 2.8911,
      "theoretical_loss": 3.376567942943861,
      "tokens_seen": 2426535936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692525252525253,
      "loss": 2.558,
      "theoretical_loss": 3.37656062528521,
      "tokens_seen": 2426601472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026923232323232323,
      "loss": 2.5752,
      "theoretical_loss": 3.3765533078795213,
      "tokens_seen": 2426667008
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1333076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.361471652984619,
      "objective/train/theoretical_loss": 3.3765459907267794,
      "objective/train/tokens_used": 785591776,
      "theoretical_loss": 3.3765459907267794,
      "tokens_seen": 2426732544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026921212121212124,
      "loss": 2.3541,
      "theoretical_loss": 3.3765459907267794,
      "tokens_seen": 2426732544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691919191919192,
      "loss": 2.7144,
      "theoretical_loss": 3.3765386738269694,
      "tokens_seen": 2426798080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026917171717171716,
      "loss": 2.3975,
      "theoretical_loss": 3.3765313571800744,
      "tokens_seen": 2426863616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026915151515151517,
      "loss": 2.8372,
      "theoretical_loss": 3.37652404078608,
      "tokens_seen": 2426929152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691313131313131,
      "loss": 2.6128,
      "theoretical_loss": 3.37651672464497,
      "tokens_seen": 2426994688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026911111111111114,
      "loss": 2.5108,
      "theoretical_loss": 3.3765094087567293,
      "tokens_seen": 2427060224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690909090909091,
      "loss": 2.6143,
      "theoretical_loss": 3.3765020931213416,
      "tokens_seen": 2427125760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690707070707071,
      "loss": 2.8087,
      "theoretical_loss": 3.3764947777387926,
      "tokens_seen": 2427191296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026905050505050506,
      "loss": 2.478,
      "theoretical_loss": 3.3764874626090653,
      "tokens_seen": 2427256832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269030303030303,
      "loss": 2.4421,
      "theoretical_loss": 3.3764801477321447,
      "tokens_seen": 2427322368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026901010101010103,
      "loss": 2.4682,
      "theoretical_loss": 3.3764728331080156,
      "tokens_seen": 2427387904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268989898989899,
      "loss": 2.6778,
      "theoretical_loss": 3.376465518736662,
      "tokens_seen": 2427453440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268969696969697,
      "loss": 2.5871,
      "theoretical_loss": 3.3764582046180687,
      "tokens_seen": 2427518976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026894949494949495,
      "loss": 2.7602,
      "theoretical_loss": 3.3764508907522197,
      "tokens_seen": 2427584512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689292929292929,
      "loss": 2.5292,
      "theoretical_loss": 3.3764435771390997,
      "tokens_seen": 2427650048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689090909090909,
      "loss": 2.4538,
      "theoretical_loss": 3.376436263778693,
      "tokens_seen": 2427715584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026888888888888893,
      "loss": 2.5796,
      "theoretical_loss": 3.3764289506709844,
      "tokens_seen": 2427781120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688686868686869,
      "loss": 2.4144,
      "theoretical_loss": 3.376421637815958,
      "tokens_seen": 2427846656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026884848484848485,
      "loss": 2.4935,
      "theoretical_loss": 3.3764143252135983,
      "tokens_seen": 2427912192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688282828282828,
      "loss": 2.7623,
      "theoretical_loss": 3.37640701286389,
      "tokens_seen": 2427977728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688080808080808,
      "loss": 2.3049,
      "theoretical_loss": 3.376399700766817,
      "tokens_seen": 2428043264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687878787878788,
      "loss": 2.4255,
      "theoretical_loss": 3.3763923889223646,
      "tokens_seen": 2428108800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687676767676768,
      "loss": 2.4888,
      "theoretical_loss": 3.376385077330516,
      "tokens_seen": 2428174336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026874747474747474,
      "loss": 2.6093,
      "theoretical_loss": 3.3763777659912573,
      "tokens_seen": 2428239872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687272727272727,
      "loss": 2.5453,
      "theoretical_loss": 3.376370454904572,
      "tokens_seen": 2428305408
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1334211,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6763081550598145,
      "objective/train/theoretical_loss": 3.376363144070444,
      "objective/train/tokens_used": 787230176,
      "theoretical_loss": 3.376363144070444,
      "tokens_seen": 2428370944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026870707070707076,
      "loss": 2.6711,
      "theoretical_loss": 3.376363144070444,
      "tokens_seen": 2428370944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686868686868687,
      "loss": 2.5968,
      "theoretical_loss": 3.376355833488859,
      "tokens_seen": 2428436480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686666666666667,
      "loss": 2.4966,
      "theoretical_loss": 3.3763485231598005,
      "tokens_seen": 2428502016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026864646464646463,
      "loss": 2.6842,
      "theoretical_loss": 3.3763412130832537,
      "tokens_seen": 2428567552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686262626262626,
      "loss": 2.4217,
      "theoretical_loss": 3.3763339032592024,
      "tokens_seen": 2428633088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026860606060606065,
      "loss": 2.5748,
      "theoretical_loss": 3.376326593687631,
      "tokens_seen": 2428698624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685858585858586,
      "loss": 2.5352,
      "theoretical_loss": 3.376319284368525,
      "tokens_seen": 2428764160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026856565656565657,
      "loss": 2.4552,
      "theoretical_loss": 3.376311975301868,
      "tokens_seen": 2428829696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685454545454545,
      "loss": 2.4599,
      "theoretical_loss": 3.3763046664876444,
      "tokens_seen": 2428895232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685252525252525,
      "loss": 2.4581,
      "theoretical_loss": 3.3762973579258393,
      "tokens_seen": 2428960768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026850505050505055,
      "loss": 2.6801,
      "theoretical_loss": 3.3762900496164363,
      "tokens_seen": 2429026304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684848484848485,
      "loss": 2.6776,
      "theoretical_loss": 3.3762827415594208,
      "tokens_seen": 2429091840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026846464646464646,
      "loss": 2.7273,
      "theoretical_loss": 3.3762754337547767,
      "tokens_seen": 2429157376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684444444444444,
      "loss": 2.6648,
      "theoretical_loss": 3.3762681262024885,
      "tokens_seen": 2429222912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026842424242424243,
      "loss": 2.5583,
      "theoretical_loss": 3.3762608189025407,
      "tokens_seen": 2429288448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026840404040404044,
      "loss": 2.3741,
      "theoretical_loss": 3.3762535118549186,
      "tokens_seen": 2429353984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683838383838384,
      "loss": 2.7823,
      "theoretical_loss": 3.3762462050596054,
      "tokens_seen": 2429419520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026836363636363635,
      "loss": 2.7154,
      "theoretical_loss": 3.3762388985165863,
      "tokens_seen": 2429485056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683434343434343,
      "loss": 2.7091,
      "theoretical_loss": 3.3762315922258455,
      "tokens_seen": 2429550592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683232323232323,
      "loss": 2.726,
      "theoretical_loss": 3.3762242861873677,
      "tokens_seen": 2429616128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026830303030303033,
      "loss": 2.5146,
      "theoretical_loss": 3.376216980401137,
      "tokens_seen": 2429681664
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682828282828283,
      "loss": 2.7387,
      "theoretical_loss": 3.3762096748671384,
      "tokens_seen": 2429747200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026826262626262624,
      "loss": 2.441,
      "theoretical_loss": 3.3762023695853562,
      "tokens_seen": 2429812736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026824242424242426,
      "loss": 2.805,
      "theoretical_loss": 3.376195064555775,
      "tokens_seen": 2429878272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682222222222222,
      "loss": 2.6108,
      "theoretical_loss": 3.376187759778379,
      "tokens_seen": 2429943808
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1334672,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.394357204437256,
      "objective/train/theoretical_loss": 3.376180455253153,
      "objective/train/tokens_used": 788868576,
      "theoretical_loss": 3.376180455253153,
      "tokens_seen": 2430009344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682020202020202,
      "loss": 2.4336,
      "theoretical_loss": 3.376180455253153,
      "tokens_seen": 2430009344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681818181818182,
      "loss": 2.7224,
      "theoretical_loss": 3.3761731509800814,
      "tokens_seen": 2430074880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026816161616161614,
      "loss": 2.5301,
      "theoretical_loss": 3.3761658469591485,
      "tokens_seen": 2430140416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026814141414141415,
      "loss": 2.4406,
      "theoretical_loss": 3.3761585431903387,
      "tokens_seen": 2430205952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681212121212121,
      "loss": 2.3457,
      "theoretical_loss": 3.376151239673637,
      "tokens_seen": 2430271488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681010101010101,
      "loss": 2.439,
      "theoretical_loss": 3.3761439364090275,
      "tokens_seen": 2430337024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002680808080808081,
      "loss": 2.4804,
      "theoretical_loss": 3.376136633396495,
      "tokens_seen": 2430402560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002680606060606061,
      "loss": 2.5079,
      "theoretical_loss": 3.3761293306360236,
      "tokens_seen": 2430468096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026804040404040404,
      "loss": 2.6971,
      "theoretical_loss": 3.3761220281275985,
      "tokens_seen": 2430533632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026802020202020205,
      "loss": 2.7266,
      "theoretical_loss": 3.3761147258712034,
      "tokens_seen": 2430599168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268,
      "loss": 2.6872,
      "theoretical_loss": 3.376107423866823,
      "tokens_seen": 2430664704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026797979797979797,
      "loss": 2.6346,
      "theoretical_loss": 3.376100122114442,
      "tokens_seen": 2430730240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267959595959596,
      "loss": 2.5747,
      "theoretical_loss": 3.376092820614045,
      "tokens_seen": 2430795776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026793939393939393,
      "loss": 2.552,
      "theoretical_loss": 3.376085519365617,
      "tokens_seen": 2430861312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026791919191919194,
      "loss": 2.7006,
      "theoretical_loss": 3.376078218369141,
      "tokens_seen": 2430926848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678989898989899,
      "loss": 2.5641,
      "theoretical_loss": 3.376070917624603,
      "tokens_seen": 2430992384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678787878787879,
      "loss": 2.7543,
      "theoretical_loss": 3.3760636171319867,
      "tokens_seen": 2431057920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026785858585858587,
      "loss": 2.7107,
      "theoretical_loss": 3.3760563168912765,
      "tokens_seen": 2431123456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678383838383838,
      "loss": 2.6819,
      "theoretical_loss": 3.3760490169024573,
      "tokens_seen": 2431188992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026781818181818184,
      "loss": 2.5399,
      "theoretical_loss": 3.376041717165514,
      "tokens_seen": 2431254528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677979797979798,
      "loss": 2.6909,
      "theoretical_loss": 3.3760344176804304,
      "tokens_seen": 2431320064
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677777777777778,
      "loss": 2.6631,
      "theoretical_loss": 3.3760271184471917,
      "tokens_seen": 2431385600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026775757575757576,
      "loss": 2.6917,
      "theoretical_loss": 3.376019819465782,
      "tokens_seen": 2431451136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677373737373737,
      "loss": 2.4872,
      "theoretical_loss": 3.376012520736185,
      "tokens_seen": 2431516672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026771717171717173,
      "loss": 2.5818,
      "theoretical_loss": 3.376005222258387,
      "tokens_seen": 2431582208
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1335677,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.379019260406494,
      "objective/train/theoretical_loss": 3.3759979240323714,
      "objective/train/tokens_used": 790506976,
      "theoretical_loss": 3.3759979240323714,
      "tokens_seen": 2431647744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026769696969696974,
      "loss": 2.5296,
      "theoretical_loss": 3.3759979240323714,
      "tokens_seen": 2431647744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676767676767677,
      "loss": 2.765,
      "theoretical_loss": 3.3759906260581225,
      "tokens_seen": 2431713280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026765656565656566,
      "loss": 2.5465,
      "theoretical_loss": 3.375983328335626,
      "tokens_seen": 2431778816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676363636363636,
      "loss": 2.3647,
      "theoretical_loss": 3.375976030864865,
      "tokens_seen": 2431844352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676161616161616,
      "loss": 2.6953,
      "theoretical_loss": 3.3759687336458253,
      "tokens_seen": 2431909888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026759595959595963,
      "loss": 2.4947,
      "theoretical_loss": 3.375961436678491,
      "tokens_seen": 2431975424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675757575757576,
      "loss": 2.4316,
      "theoretical_loss": 3.375954139962846,
      "tokens_seen": 2432040960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026755555555555555,
      "loss": 2.6234,
      "theoretical_loss": 3.3759468434988755,
      "tokens_seen": 2432106496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675353535353535,
      "loss": 2.5268,
      "theoretical_loss": 3.375939547286564,
      "tokens_seen": 2432172032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026751515151515157,
      "loss": 2.735,
      "theoretical_loss": 3.375932251325896,
      "tokens_seen": 2432237568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674949494949495,
      "loss": 2.6445,
      "theoretical_loss": 3.3759249556168553,
      "tokens_seen": 2432303104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674747474747475,
      "loss": 2.61,
      "theoretical_loss": 3.375917660159428,
      "tokens_seen": 2432368640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026745454545454544,
      "loss": 2.4683,
      "theoretical_loss": 3.3759103649535973,
      "tokens_seen": 2432434176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674343434343434,
      "loss": 2.5302,
      "theoretical_loss": 3.375903069999348,
      "tokens_seen": 2432499712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026741414141414146,
      "loss": 2.7753,
      "theoretical_loss": 3.375895775296665,
      "tokens_seen": 2432565248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673939393939394,
      "loss": 2.4033,
      "theoretical_loss": 3.375888480845533,
      "tokens_seen": 2432630784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673737373737374,
      "loss": 2.6525,
      "theoretical_loss": 3.375881186645936,
      "tokens_seen": 2432696320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026735353535353533,
      "loss": 2.5969,
      "theoretical_loss": 3.3758738926978586,
      "tokens_seen": 2432761856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026733333333333334,
      "loss": 2.6262,
      "theoretical_loss": 3.375866599001286,
      "tokens_seen": 2432827392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026731313131313136,
      "loss": 2.479,
      "theoretical_loss": 3.375859305556202,
      "tokens_seen": 2432892928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672929292929293,
      "loss": 2.5408,
      "theoretical_loss": 3.3758520123625915,
      "tokens_seen": 2432958464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026727272727272727,
      "loss": 2.6221,
      "theoretical_loss": 3.375844719420439,
      "tokens_seen": 2433024000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672525252525252,
      "loss": 2.5281,
      "theoretical_loss": 3.3758374267297286,
      "tokens_seen": 2433089536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026723232323232324,
      "loss": 2.822,
      "theoretical_loss": 3.375830134290446,
      "tokens_seen": 2433155072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026721212121212125,
      "loss": 2.382,
      "theoretical_loss": 3.3758228421025747,
      "tokens_seen": 2433220608
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1336773,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.446194887161255,
      "objective/train/theoretical_loss": 3.3758155501661,
      "objective/train/tokens_used": 792145376,
      "theoretical_loss": 3.3758155501661,
      "tokens_seen": 2433286144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671919191919192,
      "loss": 2.6136,
      "theoretical_loss": 3.3758155501661,
      "tokens_seen": 2433286144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026717171717171716,
      "loss": 2.3919,
      "theoretical_loss": 3.3758082584810056,
      "tokens_seen": 2433351680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671515151515151,
      "loss": 2.2964,
      "theoretical_loss": 3.3758009670472773,
      "tokens_seen": 2433417216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026713131313131313,
      "loss": 2.4072,
      "theoretical_loss": 3.375793675864898,
      "tokens_seen": 2433482752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026711111111111114,
      "loss": 2.6743,
      "theoretical_loss": 3.375786384933854,
      "tokens_seen": 2433548288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002670909090909091,
      "loss": 2.5573,
      "theoretical_loss": 3.3757790942541286,
      "tokens_seen": 2433613824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026707070707070705,
      "loss": 2.6263,
      "theoretical_loss": 3.375771803825707,
      "tokens_seen": 2433679360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026705050505050507,
      "loss": 2.6231,
      "theoretical_loss": 3.3757645136485737,
      "tokens_seen": 2433744896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267030303030303,
      "loss": 2.4567,
      "theoretical_loss": 3.375757223722713,
      "tokens_seen": 2433810432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026701010101010103,
      "loss": 2.4794,
      "theoretical_loss": 3.37574993404811,
      "tokens_seen": 2433875968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266989898989899,
      "loss": 2.5598,
      "theoretical_loss": 3.3757426446247485,
      "tokens_seen": 2433941504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026696969696969695,
      "loss": 2.4144,
      "theoretical_loss": 3.3757353554526137,
      "tokens_seen": 2434007040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026694949494949496,
      "loss": 2.4774,
      "theoretical_loss": 3.3757280665316896,
      "tokens_seen": 2434072576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002669292929292929,
      "loss": 2.6486,
      "theoretical_loss": 3.375720777861962,
      "tokens_seen": 2434138112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002669090909090909,
      "loss": 2.7292,
      "theoretical_loss": 3.375713489443414,
      "tokens_seen": 2434203648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668888888888889,
      "loss": 2.5281,
      "theoretical_loss": 3.3757062012760306,
      "tokens_seen": 2434269184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668686868686869,
      "loss": 2.4617,
      "theoretical_loss": 3.375698913359797,
      "tokens_seen": 2434334720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026684848484848485,
      "loss": 2.5479,
      "theoretical_loss": 3.3756916256946976,
      "tokens_seen": 2434400256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668282828282828,
      "loss": 2.8708,
      "theoretical_loss": 3.375684338280716,
      "tokens_seen": 2434465792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668080808080808,
      "loss": 2.6174,
      "theoretical_loss": 3.3756770511178384,
      "tokens_seen": 2434531328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667878787878788,
      "loss": 2.584,
      "theoretical_loss": 3.375669764206048,
      "tokens_seen": 2434596864
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667676767676768,
      "loss": 2.6701,
      "theoretical_loss": 3.3756624775453306,
      "tokens_seen": 2434662400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026674747474747474,
      "loss": 2.5115,
      "theoretical_loss": 3.3756551911356696,
      "tokens_seen": 2434727936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667272727272727,
      "loss": 2.608,
      "theoretical_loss": 3.37564790497705,
      "tokens_seen": 2434793472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667070707070707,
      "loss": 2.6837,
      "theoretical_loss": 3.375640619069457,
      "tokens_seen": 2434859008
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1337466,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9103434085845947,
      "objective/train/theoretical_loss": 3.3756333334128743,
      "objective/train/tokens_used": 793783776,
      "theoretical_loss": 3.3756333334128743,
      "tokens_seen": 2434924544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666868686868687,
      "loss": 2.5773,
      "theoretical_loss": 3.3756333334128743,
      "tokens_seen": 2434924544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666666666666667,
      "loss": 2.6097,
      "theoretical_loss": 3.375626048007287,
      "tokens_seen": 2434990080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026664646464646464,
      "loss": 2.5248,
      "theoretical_loss": 3.37561876285268,
      "tokens_seen": 2435055616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666262626262626,
      "loss": 2.5488,
      "theoretical_loss": 3.375611477949037,
      "tokens_seen": 2435121152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666060606060606,
      "loss": 2.4839,
      "theoretical_loss": 3.3756041932963434,
      "tokens_seen": 2435186688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665858585858586,
      "loss": 2.5291,
      "theoretical_loss": 3.3755969088945834,
      "tokens_seen": 2435252224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026656565656565657,
      "loss": 2.7532,
      "theoretical_loss": 3.375589624743742,
      "tokens_seen": 2435317760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026654545454545453,
      "loss": 2.523,
      "theoretical_loss": 3.375582340843803,
      "tokens_seen": 2435383296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665252525252525,
      "loss": 2.6294,
      "theoretical_loss": 3.3755750571947516,
      "tokens_seen": 2435448832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026650505050505055,
      "loss": 2.5958,
      "theoretical_loss": 3.375567773796573,
      "tokens_seen": 2435514368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664848484848485,
      "loss": 2.4305,
      "theoretical_loss": 3.3755604906492502,
      "tokens_seen": 2435579904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026646464646464646,
      "loss": 2.6766,
      "theoretical_loss": 3.3755532077527692,
      "tokens_seen": 2435645440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664444444444444,
      "loss": 2.5552,
      "theoretical_loss": 3.3755459251071143,
      "tokens_seen": 2435710976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664242424242424,
      "loss": 2.6694,
      "theoretical_loss": 3.3755386427122698,
      "tokens_seen": 2435776512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026640404040404044,
      "loss": 2.5408,
      "theoretical_loss": 3.3755313605682202,
      "tokens_seen": 2435842048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663838383838384,
      "loss": 2.4806,
      "theoretical_loss": 3.375524078674951,
      "tokens_seen": 2435907584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026636363636363636,
      "loss": 2.3821,
      "theoretical_loss": 3.375516797032446,
      "tokens_seen": 2435973120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663434343434343,
      "loss": 2.6681,
      "theoretical_loss": 3.3755095156406902,
      "tokens_seen": 2436038656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663232323232324,
      "loss": 2.6113,
      "theoretical_loss": 3.3755022344996677,
      "tokens_seen": 2436104192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026630303030303034,
      "loss": 2.6237,
      "theoretical_loss": 3.3754949536093637,
      "tokens_seen": 2436169728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662828282828283,
      "loss": 2.6986,
      "theoretical_loss": 3.3754876729697627,
      "tokens_seen": 2436235264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026626262626262625,
      "loss": 2.4305,
      "theoretical_loss": 3.375480392580849,
      "tokens_seen": 2436300800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662424242424242,
      "loss": 2.5156,
      "theoretical_loss": 3.3754731124426076,
      "tokens_seen": 2436366336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026622222222222227,
      "loss": 2.5138,
      "theoretical_loss": 3.375465832555023,
      "tokens_seen": 2436431872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026620202020202023,
      "loss": 2.6387,
      "theoretical_loss": 3.3754585529180794,
      "tokens_seen": 2436497408
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1338189,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9122860431671143,
      "objective/train/theoretical_loss": 3.3754512735317626,
      "objective/train/tokens_used": 795422176,
      "theoretical_loss": 3.3754512735317626,
      "tokens_seen": 2436562944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661818181818182,
      "loss": 2.6878,
      "theoretical_loss": 3.3754512735317626,
      "tokens_seen": 2436562944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026616161616161614,
      "loss": 2.5895,
      "theoretical_loss": 3.375443994396056,
      "tokens_seen": 2436628480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026614141414141415,
      "loss": 2.6319,
      "theoretical_loss": 3.3754367155109444,
      "tokens_seen": 2436694016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026612121212121217,
      "loss": 2.5464,
      "theoretical_loss": 3.3754294368764133,
      "tokens_seen": 2436759552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661010101010101,
      "loss": 2.6127,
      "theoretical_loss": 3.3754221584924466,
      "tokens_seen": 2436825088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002660808080808081,
      "loss": 2.4074,
      "theoretical_loss": 3.375414880359029,
      "tokens_seen": 2436890624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026606060606060604,
      "loss": 2.6108,
      "theoretical_loss": 3.3754076024761455,
      "tokens_seen": 2436956160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026604040404040405,
      "loss": 2.7403,
      "theoretical_loss": 3.3754003248437803,
      "tokens_seen": 2437021696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026602020202020206,
      "loss": 2.65,
      "theoretical_loss": 3.375393047461918,
      "tokens_seen": 2437087232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266,
      "loss": 2.4183,
      "theoretical_loss": 3.375385770330544,
      "tokens_seen": 2437152768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026597979797979797,
      "loss": 2.7428,
      "theoretical_loss": 3.3753784934496416,
      "tokens_seen": 2437218304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000265959595959596,
      "loss": 2.6406,
      "theoretical_loss": 3.375371216819197,
      "tokens_seen": 2437283840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026593939393939394,
      "loss": 2.5311,
      "theoretical_loss": 3.3753639404391937,
      "tokens_seen": 2437349376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026591919191919195,
      "loss": 2.5561,
      "theoretical_loss": 3.375356664309617,
      "tokens_seen": 2437414912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658989898989899,
      "loss": 2.5476,
      "theoretical_loss": 3.3753493884304513,
      "tokens_seen": 2437480448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026587878787878786,
      "loss": 2.73,
      "theoretical_loss": 3.3753421128016807,
      "tokens_seen": 2437545984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658585858585859,
      "loss": 2.5517,
      "theoretical_loss": 3.375334837423291,
      "tokens_seen": 2437611520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026583838383838383,
      "loss": 2.542,
      "theoretical_loss": 3.375327562295266,
      "tokens_seen": 2437677056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026581818181818184,
      "loss": 2.5594,
      "theoretical_loss": 3.3753202874175905,
      "tokens_seen": 2437742592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657979797979798,
      "loss": 2.635,
      "theoretical_loss": 3.3753130127902495,
      "tokens_seen": 2437808128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657777777777778,
      "loss": 2.6303,
      "theoretical_loss": 3.3753057384132275,
      "tokens_seen": 2437873664
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026575757575757577,
      "loss": 2.6106,
      "theoretical_loss": 3.3752984642865087,
      "tokens_seen": 2437939200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657373737373737,
      "loss": 2.8313,
      "theoretical_loss": 3.3752911904100786,
      "tokens_seen": 2438004736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026571717171717174,
      "loss": 2.4608,
      "theoretical_loss": 3.375283916783921,
      "tokens_seen": 2438070272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656969696969697,
      "loss": 2.8014,
      "theoretical_loss": 3.375276643408021,
      "tokens_seen": 2438135808
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1339345,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.939288377761841,
      "objective/train/theoretical_loss": 3.3752693702823633,
      "objective/train/tokens_used": 797060576,
      "theoretical_loss": 3.3752693702823633,
      "tokens_seen": 2438201344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656767676767677,
      "loss": 2.6516,
      "theoretical_loss": 3.3752693702823633,
      "tokens_seen": 2438201344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026565656565656566,
      "loss": 2.5418,
      "theoretical_loss": 3.3752620974069325,
      "tokens_seen": 2438266880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656363636363636,
      "loss": 2.3661,
      "theoretical_loss": 3.375254824781713,
      "tokens_seen": 2438332416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026561616161616163,
      "loss": 2.2171,
      "theoretical_loss": 3.37524755240669,
      "tokens_seen": 2438397952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026559595959595964,
      "loss": 2.7802,
      "theoretical_loss": 3.3752402802818473,
      "tokens_seen": 2438463488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655757575757576,
      "loss": 2.5123,
      "theoretical_loss": 3.3752330084071707,
      "tokens_seen": 2438529024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026555555555555555,
      "loss": 2.6926,
      "theoretical_loss": 3.3752257367826446,
      "tokens_seen": 2438594560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655353535353535,
      "loss": 2.5719,
      "theoretical_loss": 3.3752184654082527,
      "tokens_seen": 2438660096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655151515151515,
      "loss": 2.6028,
      "theoretical_loss": 3.3752111942839806,
      "tokens_seen": 2438725632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026549494949494953,
      "loss": 2.551,
      "theoretical_loss": 3.375203923409813,
      "tokens_seen": 2438791168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654747474747475,
      "loss": 2.5455,
      "theoretical_loss": 3.375196652785734,
      "tokens_seen": 2438856704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026545454545454545,
      "loss": 2.5723,
      "theoretical_loss": 3.3751893824117287,
      "tokens_seen": 2438922240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654343434343434,
      "loss": 2.5632,
      "theoretical_loss": 3.375182112287782,
      "tokens_seen": 2438987776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026541414141414147,
      "loss": 2.6393,
      "theoretical_loss": 3.3751748424138777,
      "tokens_seen": 2439053312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653939393939394,
      "loss": 2.5782,
      "theoretical_loss": 3.3751675727900015,
      "tokens_seen": 2439118848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653737373737374,
      "loss": 2.6618,
      "theoretical_loss": 3.3751603034161373,
      "tokens_seen": 2439184384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026535353535353534,
      "loss": 2.4814,
      "theoretical_loss": 3.3751530342922704,
      "tokens_seen": 2439249920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653333333333333,
      "loss": 2.5982,
      "theoretical_loss": 3.375145765418385,
      "tokens_seen": 2439315456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026531313131313136,
      "loss": 2.5569,
      "theoretical_loss": 3.3751384967944658,
      "tokens_seen": 2439380992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652929292929293,
      "loss": 2.5505,
      "theoretical_loss": 3.3751312284204977,
      "tokens_seen": 2439446528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652727272727273,
      "loss": 2.6894,
      "theoretical_loss": 3.3751239602964658,
      "tokens_seen": 2439512064
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026525252525252523,
      "loss": 2.5101,
      "theoretical_loss": 3.375116692422354,
      "tokens_seen": 2439577600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652323232323232,
      "loss": 2.6101,
      "theoretical_loss": 3.3751094247981475,
      "tokens_seen": 2439643136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026521212121212125,
      "loss": 2.588,
      "theoretical_loss": 3.3751021574238305,
      "tokens_seen": 2439708672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002651919191919192,
      "loss": 2.5619,
      "theoretical_loss": 3.3750948902993883,
      "tokens_seen": 2439774208
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 1339900,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4708569049835205,
      "objective/train/theoretical_loss": 3.3750876234248053,
      "objective/train/tokens_used": 798698976,
      "theoretical_loss": 3.3750876234248053,
      "tokens_seen": 2439839744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026517171717171717,
      "loss": 2.5385,
      "theoretical_loss": 3.3750876234248053,
      "tokens_seen": 2439839744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651515151515151,
      "loss": 2.54,
      "theoretical_loss": 3.3750803568000665,
      "tokens_seen": 2439905280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026513131313131314,
      "loss": 2.5333,
      "theoretical_loss": 3.375073090425156,
      "tokens_seen": 2439970816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026511111111111115,
      "loss": 2.611,
      "theoretical_loss": 3.3750658243000586,
      "tokens_seen": 2440036352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650909090909091,
      "loss": 2.3361,
      "theoretical_loss": 3.37505855842476,
      "tokens_seen": 2440101888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026507070707070706,
      "loss": 2.3275,
      "theoretical_loss": 3.3750512927992435,
      "tokens_seen": 2440167424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000265050505050505,
      "loss": 2.3733,
      "theoretical_loss": 3.3750440274234945,
      "tokens_seen": 2440232960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026503030303030303,
      "loss": 2.6831,
      "theoretical_loss": 3.3750367622974977,
      "tokens_seen": 2440298496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026501010101010104,
      "loss": 2.2953,
      "theoretical_loss": 3.375029497421238,
      "tokens_seen": 2440364032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264989898989899,
      "loss": 2.5982,
      "theoretical_loss": 3.3750222327946995,
      "tokens_seen": 2440429568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026496969696969695,
      "loss": 2.5708,
      "theoretical_loss": 3.3750149684178674,
      "tokens_seen": 2440495104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026494949494949496,
      "loss": 2.4435,
      "theoretical_loss": 3.3750077042907263,
      "tokens_seen": 2440560640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002649292929292929,
      "loss": 2.5218,
      "theoretical_loss": 3.375000440413261,
      "tokens_seen": 2440626176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026490909090909093,
      "loss": 2.4119,
      "theoretical_loss": 3.374993176785456,
      "tokens_seen": 2440691712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648888888888889,
      "loss": 2.6621,
      "theoretical_loss": 3.374985913407296,
      "tokens_seen": 2440757248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026486868686868685,
      "loss": 2.6159,
      "theoretical_loss": 3.374978650278766,
      "tokens_seen": 2440822784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026484848484848486,
      "loss": 2.7574,
      "theoretical_loss": 3.374971387399851,
      "tokens_seen": 2440888320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648282828282828,
      "loss": 2.3123,
      "theoretical_loss": 3.3749641247705346,
      "tokens_seen": 2440953856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648080808080808,
      "loss": 2.5051,
      "theoretical_loss": 3.3749568623908024,
      "tokens_seen": 2441019392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647878787878788,
      "loss": 2.6248,
      "theoretical_loss": 3.374949600260639,
      "tokens_seen": 2441084928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647676767676768,
      "loss": 2.6467,
      "theoretical_loss": 3.3749423383800288,
      "tokens_seen": 2441150464
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026474747474747475,
      "loss": 2.652,
      "theoretical_loss": 3.374935076748957,
      "tokens_seen": 2441216000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647272727272727,
      "loss": 2.4366,
      "theoretical_loss": 3.374927815367408,
      "tokens_seen": 2441281536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647070707070707,
      "loss": 2.4901,
      "theoretical_loss": 3.374920554235367,
      "tokens_seen": 2441347072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646868686868687,
      "loss": 2.4452,
      "theoretical_loss": 3.374913293352818,
      "tokens_seen": 2441412608
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1341041,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8801026344299316,
      "objective/train/theoretical_loss": 3.374906032719746,
      "objective/train/tokens_used": 800337376,
      "theoretical_loss": 3.374906032719746,
      "tokens_seen": 2441478144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646666666666667,
      "loss": 2.62,
      "theoretical_loss": 3.374906032719746,
      "tokens_seen": 2441478144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026464646464646464,
      "loss": 2.5879,
      "theoretical_loss": 3.3748987723361363,
      "tokens_seen": 2441543680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026462626262626265,
      "loss": 2.2701,
      "theoretical_loss": 3.3748915122019727,
      "tokens_seen": 2441609216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646060606060606,
      "loss": 2.6169,
      "theoretical_loss": 3.3748842523172407,
      "tokens_seen": 2441674752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645858585858586,
      "loss": 2.7176,
      "theoretical_loss": 3.374876992681924,
      "tokens_seen": 2441740288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645656565656566,
      "loss": 2.6777,
      "theoretical_loss": 3.3748697332960087,
      "tokens_seen": 2441805824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026454545454545453,
      "loss": 2.5114,
      "theoretical_loss": 3.374862474159479,
      "tokens_seen": 2441871360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026452525252525255,
      "loss": 2.4141,
      "theoretical_loss": 3.3748552152723192,
      "tokens_seen": 2441936896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645050505050505,
      "loss": 2.4707,
      "theoretical_loss": 3.3748479566345146,
      "tokens_seen": 2442002432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002644848484848485,
      "loss": 2.6373,
      "theoretical_loss": 3.3748406982460497,
      "tokens_seen": 2442067968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026446464646464647,
      "loss": 2.733,
      "theoretical_loss": 3.3748334401069093,
      "tokens_seen": 2442133504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026444444444444443,
      "loss": 2.6576,
      "theoretical_loss": 3.374826182217078,
      "tokens_seen": 2442199040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026442424242424244,
      "loss": 2.7022,
      "theoretical_loss": 3.374818924576541,
      "tokens_seen": 2442264576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026440404040404045,
      "loss": 2.5744,
      "theoretical_loss": 3.3748116671852824,
      "tokens_seen": 2442330112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643838383838384,
      "loss": 2.4534,
      "theoretical_loss": 3.374804410043287,
      "tokens_seen": 2442395648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026436363636363636,
      "loss": 2.7643,
      "theoretical_loss": 3.3747971531505403,
      "tokens_seen": 2442461184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643434343434343,
      "loss": 2.7687,
      "theoretical_loss": 3.374789896507026,
      "tokens_seen": 2442526720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026432323232323233,
      "loss": 2.7009,
      "theoretical_loss": 3.3747826401127297,
      "tokens_seen": 2442592256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026430303030303034,
      "loss": 2.4016,
      "theoretical_loss": 3.374775383967636,
      "tokens_seen": 2442657792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642828282828283,
      "loss": 2.7504,
      "theoretical_loss": 3.3747681280717297,
      "tokens_seen": 2442723328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026426262626262626,
      "loss": 2.7235,
      "theoretical_loss": 3.3747608724249947,
      "tokens_seen": 2442788864
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642424242424242,
      "loss": 2.5554,
      "theoretical_loss": 3.374753617027417,
      "tokens_seen": 2442854400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642222222222223,
      "loss": 2.7524,
      "theoretical_loss": 3.374746361878981,
      "tokens_seen": 2442919936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026420202020202023,
      "loss": 2.5263,
      "theoretical_loss": 3.374739106979671,
      "tokens_seen": 2442985472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641818181818182,
      "loss": 2.4002,
      "theoretical_loss": 3.3747318523294716,
      "tokens_seen": 2443051008
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1341769,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2153728008270264,
      "objective/train/theoretical_loss": 3.3747245979283687,
      "objective/train/tokens_used": 801975776,
      "theoretical_loss": 3.3747245979283687,
      "tokens_seen": 2443116544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026416161616161615,
      "loss": 2.4886,
      "theoretical_loss": 3.3747245979283687,
      "tokens_seen": 2443116544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641414141414141,
      "loss": 2.5573,
      "theoretical_loss": 3.3747173437763456,
      "tokens_seen": 2443182080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026412121212121217,
      "loss": 2.5934,
      "theoretical_loss": 3.374710089873388,
      "tokens_seen": 2443247616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026410101010101013,
      "loss": 2.6106,
      "theoretical_loss": 3.3747028362194813,
      "tokens_seen": 2443313152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640808080808081,
      "loss": 2.5297,
      "theoretical_loss": 3.3746955828146086,
      "tokens_seen": 2443378688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026406060606060604,
      "loss": 2.5335,
      "theoretical_loss": 3.3746883296587558,
      "tokens_seen": 2443444224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026404040404040405,
      "loss": 2.6934,
      "theoretical_loss": 3.3746810767519073,
      "tokens_seen": 2443509760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026402020202020206,
      "loss": 2.6769,
      "theoretical_loss": 3.374673824094048,
      "tokens_seen": 2443575296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264,
      "loss": 2.6634,
      "theoretical_loss": 3.374666571685163,
      "tokens_seen": 2443640832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263979797979798,
      "loss": 2.5802,
      "theoretical_loss": 3.3746593195252363,
      "tokens_seen": 2443706368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026395959595959593,
      "loss": 2.7276,
      "theoretical_loss": 3.3746520676142535,
      "tokens_seen": 2443771904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026393939393939395,
      "loss": 2.4221,
      "theoretical_loss": 3.3746448159521987,
      "tokens_seen": 2443837440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026391919191919196,
      "loss": 2.6148,
      "theoretical_loss": 3.374637564539057,
      "tokens_seen": 2443902976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638989898989899,
      "loss": 2.766,
      "theoretical_loss": 3.374630313374813,
      "tokens_seen": 2443968512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026387878787878787,
      "loss": 2.5676,
      "theoretical_loss": 3.374623062459452,
      "tokens_seen": 2444034048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638585858585858,
      "loss": 2.4868,
      "theoretical_loss": 3.374615811792958,
      "tokens_seen": 2444099584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026383838383838384,
      "loss": 2.6143,
      "theoretical_loss": 3.3746085613753163,
      "tokens_seen": 2444165120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026381818181818185,
      "loss": 2.4979,
      "theoretical_loss": 3.374601311206512,
      "tokens_seen": 2444230656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637979797979798,
      "loss": 2.5034,
      "theoretical_loss": 3.374594061286529,
      "tokens_seen": 2444296192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026377777777777776,
      "loss": 2.5198,
      "theoretical_loss": 3.3745868116153526,
      "tokens_seen": 2444361728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637575757575758,
      "loss": 2.8063,
      "theoretical_loss": 3.3745795621929675,
      "tokens_seen": 2444427264
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026373737373737373,
      "loss": 2.8533,
      "theoretical_loss": 3.374572313019359,
      "tokens_seen": 2444492800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026371717171717174,
      "loss": 2.494,
      "theoretical_loss": 3.374565064094511,
      "tokens_seen": 2444558336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636969696969697,
      "loss": 2.6548,
      "theoretical_loss": 3.374557815418409,
      "tokens_seen": 2444623872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026367676767676766,
      "loss": 2.6307,
      "theoretical_loss": 3.3745505669910374,
      "tokens_seen": 2444689408
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1342804,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6955645084381104,
      "objective/train/theoretical_loss": 3.374543318812381,
      "objective/train/tokens_used": 803614176,
      "theoretical_loss": 3.374543318812381,
      "tokens_seen": 2444754944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026365656565656567,
      "loss": 2.6506,
      "theoretical_loss": 3.374543318812381,
      "tokens_seen": 2444754944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636363636363636,
      "loss": 2.3067,
      "theoretical_loss": 3.374536070882425,
      "tokens_seen": 2444820480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026361616161616163,
      "loss": 2.5639,
      "theoretical_loss": 3.3745288232011537,
      "tokens_seen": 2444886016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635959595959596,
      "loss": 2.4762,
      "theoretical_loss": 3.3745215757685525,
      "tokens_seen": 2444951552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635757575757576,
      "loss": 2.6961,
      "theoretical_loss": 3.3745143285846053,
      "tokens_seen": 2445017088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026355555555555556,
      "loss": 2.6124,
      "theoretical_loss": 3.3745070816492975,
      "tokens_seen": 2445082624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635353535353535,
      "loss": 2.6841,
      "theoretical_loss": 3.3744998349626143,
      "tokens_seen": 2445148160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635151515151515,
      "loss": 2.6576,
      "theoretical_loss": 3.37449258852454,
      "tokens_seen": 2445213696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634949494949495,
      "loss": 2.6613,
      "theoretical_loss": 3.374485342335059,
      "tokens_seen": 2445279232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634747474747475,
      "loss": 2.662,
      "theoretical_loss": 3.374478096394157,
      "tokens_seen": 2445344768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026345454545454545,
      "loss": 2.7094,
      "theoretical_loss": 3.374470850701818,
      "tokens_seen": 2445410304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634343434343434,
      "loss": 2.6962,
      "theoretical_loss": 3.3744636052580272,
      "tokens_seen": 2445475840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634141414141414,
      "loss": 2.6733,
      "theoretical_loss": 3.3744563600627697,
      "tokens_seen": 2445541376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026339393939393943,
      "loss": 2.599,
      "theoretical_loss": 3.37444911511603,
      "tokens_seen": 2445606912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002633737373737374,
      "loss": 2.4048,
      "theoretical_loss": 3.3744418704177925,
      "tokens_seen": 2445672448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026335353535353534,
      "loss": 2.6704,
      "theoretical_loss": 3.3744346259680427,
      "tokens_seen": 2445737984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002633333333333333,
      "loss": 2.7065,
      "theoretical_loss": 3.3744273817667656,
      "tokens_seen": 2445803520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002633131313131313,
      "loss": 2.696,
      "theoretical_loss": 3.374420137813945,
      "tokens_seen": 2445869056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632929292929293,
      "loss": 2.6544,
      "theoretical_loss": 3.3744128941095664,
      "tokens_seen": 2445934592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632727272727273,
      "loss": 2.5253,
      "theoretical_loss": 3.374405650653615,
      "tokens_seen": 2446000128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026325252525252524,
      "loss": 2.58,
      "theoretical_loss": 3.374398407446075,
      "tokens_seen": 2446065664
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632323232323232,
      "loss": 2.8209,
      "theoretical_loss": 3.374391164486931,
      "tokens_seen": 2446131200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026321212121212126,
      "loss": 2.4928,
      "theoretical_loss": 3.374383921776168,
      "tokens_seen": 2446196736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631919191919192,
      "loss": 2.8239,
      "theoretical_loss": 3.3743766793137717,
      "tokens_seen": 2446262272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631717171717172,
      "loss": 2.7722,
      "theoretical_loss": 3.374369437099726,
      "tokens_seen": 2446327808
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1343166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8937578201293945,
      "objective/train/theoretical_loss": 3.374362195134016,
      "objective/train/tokens_used": 805252576,
      "theoretical_loss": 3.374362195134016,
      "tokens_seen": 2446393344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026315151515151513,
      "loss": 2.7007,
      "theoretical_loss": 3.374362195134016,
      "tokens_seen": 2446393344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631313131313131,
      "loss": 2.5962,
      "theoretical_loss": 3.374354953416627,
      "tokens_seen": 2446458880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026311111111111115,
      "loss": 2.642,
      "theoretical_loss": 3.374347711947543,
      "tokens_seen": 2446524416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630909090909091,
      "loss": 2.4228,
      "theoretical_loss": 3.374340470726749,
      "tokens_seen": 2446589952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026307070707070707,
      "loss": 2.741,
      "theoretical_loss": 3.37433322975423,
      "tokens_seen": 2446655488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263050505050505,
      "loss": 2.7276,
      "theoretical_loss": 3.374325989029971,
      "tokens_seen": 2446721024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026303030303030303,
      "loss": 2.8696,
      "theoretical_loss": 3.3743187485539567,
      "tokens_seen": 2446786560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026301010101010104,
      "loss": 2.4678,
      "theoretical_loss": 3.3743115083261723,
      "tokens_seen": 2446852096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262989898989899,
      "loss": 2.6835,
      "theoretical_loss": 3.374304268346602,
      "tokens_seen": 2446917632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026296969696969696,
      "loss": 2.7896,
      "theoretical_loss": 3.374297028615231,
      "tokens_seen": 2446983168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629494949494949,
      "loss": 2.7025,
      "theoretical_loss": 3.374289789132044,
      "tokens_seen": 2447048704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262929292929293,
      "loss": 2.6045,
      "theoretical_loss": 3.3742825498970257,
      "tokens_seen": 2447114240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026290909090909094,
      "loss": 2.5814,
      "theoretical_loss": 3.3742753109101615,
      "tokens_seen": 2447179776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628888888888889,
      "loss": 2.668,
      "theoretical_loss": 3.3742680721714358,
      "tokens_seen": 2447245312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026286868686868685,
      "loss": 2.7779,
      "theoretical_loss": 3.3742608336808337,
      "tokens_seen": 2447310848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026284848484848486,
      "loss": 2.6306,
      "theoretical_loss": 3.37425359543834,
      "tokens_seen": 2447376384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628282828282829,
      "loss": 2.6649,
      "theoretical_loss": 3.3742463574439396,
      "tokens_seen": 2447441920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026280808080808083,
      "loss": 2.5009,
      "theoretical_loss": 3.374239119697617,
      "tokens_seen": 2447507456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627878787878788,
      "loss": 2.7443,
      "theoretical_loss": 3.374231882199357,
      "tokens_seen": 2447572992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026276767676767674,
      "loss": 2.4669,
      "theoretical_loss": 3.374224644949145,
      "tokens_seen": 2447638528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026274747474747475,
      "loss": 2.532,
      "theoretical_loss": 3.374217407946966,
      "tokens_seen": 2447704064
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026272727272727277,
      "loss": 2.5868,
      "theoretical_loss": 3.3742101711928036,
      "tokens_seen": 2447769600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627070707070707,
      "loss": 2.5711,
      "theoretical_loss": 3.374202934686644,
      "tokens_seen": 2447835136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626868686868687,
      "loss": 2.6259,
      "theoretical_loss": 3.374195698428472,
      "tokens_seen": 2447900672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626666666666667,
      "loss": 2.6822,
      "theoretical_loss": 3.3741884624182714,
      "tokens_seen": 2447966208
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1344165,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.270679235458374,
      "objective/train/theoretical_loss": 3.374181226656028,
      "objective/train/tokens_used": 806890976,
      "theoretical_loss": 3.374181226656028,
      "tokens_seen": 2448031744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026264646464646465,
      "loss": 2.9149,
      "theoretical_loss": 3.374181226656028,
      "tokens_seen": 2448031744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026262626262626266,
      "loss": 2.7126,
      "theoretical_loss": 3.3741739911417263,
      "tokens_seen": 2448097280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626060606060606,
      "loss": 2.4629,
      "theoretical_loss": 3.374166755875351,
      "tokens_seen": 2448162816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026258585858585857,
      "loss": 2.7661,
      "theoretical_loss": 3.3741595208568875,
      "tokens_seen": 2448228352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625656565656566,
      "loss": 2.512,
      "theoretical_loss": 3.3741522860863205,
      "tokens_seen": 2448293888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026254545454545454,
      "loss": 2.7218,
      "theoretical_loss": 3.3741450515636346,
      "tokens_seen": 2448359424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026252525252525255,
      "loss": 2.7902,
      "theoretical_loss": 3.3741378172888146,
      "tokens_seen": 2448424960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625050505050505,
      "loss": 2.4496,
      "theoretical_loss": 3.374130583261846,
      "tokens_seen": 2448490496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624848484848485,
      "loss": 2.4347,
      "theoretical_loss": 3.374123349482713,
      "tokens_seen": 2448556032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624646464646465,
      "loss": 2.7818,
      "theoretical_loss": 3.374116115951401,
      "tokens_seen": 2448621568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026244444444444443,
      "loss": 2.805,
      "theoretical_loss": 3.374108882667894,
      "tokens_seen": 2448687104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026242424242424244,
      "loss": 2.7023,
      "theoretical_loss": 3.374101649632178,
      "tokens_seen": 2448752640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624040404040404,
      "loss": 2.7081,
      "theoretical_loss": 3.3740944168442373,
      "tokens_seen": 2448818176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623838383838384,
      "loss": 2.526,
      "theoretical_loss": 3.3740871843040567,
      "tokens_seen": 2448883712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026236363636363637,
      "loss": 2.5459,
      "theoretical_loss": 3.3740799520116216,
      "tokens_seen": 2448949248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623434343434343,
      "loss": 2.6773,
      "theoretical_loss": 3.3740727199669163,
      "tokens_seen": 2449014784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026232323232323234,
      "loss": 2.6045,
      "theoretical_loss": 3.3740654881699257,
      "tokens_seen": 2449080320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026230303030303035,
      "loss": 2.8309,
      "theoretical_loss": 3.3740582566206356,
      "tokens_seen": 2449145856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622828282828283,
      "loss": 2.687,
      "theoretical_loss": 3.3740510253190297,
      "tokens_seen": 2449211392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026226262626262626,
      "loss": 2.6226,
      "theoretical_loss": 3.374043794265093,
      "tokens_seen": 2449276928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622424242424242,
      "loss": 2.5032,
      "theoretical_loss": 3.3740365634588114,
      "tokens_seen": 2449342464
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026222222222222223,
      "loss": 2.5379,
      "theoretical_loss": 3.374029332900169,
      "tokens_seen": 2449408000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026220202020202024,
      "loss": 2.4041,
      "theoretical_loss": 3.3740221025891506,
      "tokens_seen": 2449473536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621818181818182,
      "loss": 2.6704,
      "theoretical_loss": 3.3740148725257417,
      "tokens_seen": 2449539072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026216161616161615,
      "loss": 2.5828,
      "theoretical_loss": 3.374007642709927,
      "tokens_seen": 2449604608
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1344624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0327699184417725,
      "objective/train/theoretical_loss": 3.3740004131416907,
      "objective/train/tokens_used": 808529376,
      "theoretical_loss": 3.3740004131416907,
      "tokens_seen": 2449670144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621414141414141,
      "loss": 2.7585,
      "theoretical_loss": 3.3740004131416907,
      "tokens_seen": 2449670144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621212121212122,
      "loss": 2.5421,
      "theoretical_loss": 3.3739931838210184,
      "tokens_seen": 2449735680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026210101010101013,
      "loss": 2.6363,
      "theoretical_loss": 3.373985954747895,
      "tokens_seen": 2449801216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620808080808081,
      "loss": 2.7274,
      "theoretical_loss": 3.3739787259223055,
      "tokens_seen": 2449866752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026206060606060605,
      "loss": 2.5467,
      "theoretical_loss": 3.3739714973442343,
      "tokens_seen": 2449932288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262040404040404,
      "loss": 2.404,
      "theoretical_loss": 3.3739642690136664,
      "tokens_seen": 2449997824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026202020202020207,
      "loss": 2.7878,
      "theoretical_loss": 3.373957040930587,
      "tokens_seen": 2450063360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262,
      "loss": 2.5061,
      "theoretical_loss": 3.3739498130949808,
      "tokens_seen": 2450128896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261979797979798,
      "loss": 2.6148,
      "theoretical_loss": 3.373942585506833,
      "tokens_seen": 2450194432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026195959595959594,
      "loss": 2.6403,
      "theoretical_loss": 3.3739353581661278,
      "tokens_seen": 2450259968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002619393939393939,
      "loss": 2.9545,
      "theoretical_loss": 3.3739281310728506,
      "tokens_seen": 2450325504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026191919191919196,
      "loss": 2.5943,
      "theoretical_loss": 3.3739209042269867,
      "tokens_seen": 2450391040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618989898989899,
      "loss": 2.8379,
      "theoretical_loss": 3.3739136776285203,
      "tokens_seen": 2450456576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618787878787879,
      "loss": 2.627,
      "theoretical_loss": 3.373906451277437,
      "tokens_seen": 2450522112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026185858585858583,
      "loss": 2.5092,
      "theoretical_loss": 3.373899225173721,
      "tokens_seen": 2450587648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026183838383838384,
      "loss": 2.4616,
      "theoretical_loss": 3.373891999317358,
      "tokens_seen": 2450653184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026181818181818185,
      "loss": 2.5581,
      "theoretical_loss": 3.3738847737083324,
      "tokens_seen": 2450718720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617979797979798,
      "loss": 2.6438,
      "theoretical_loss": 3.373877548346629,
      "tokens_seen": 2450784256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026177777777777777,
      "loss": 2.5025,
      "theoretical_loss": 3.3738703232322327,
      "tokens_seen": 2450849792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617575757575757,
      "loss": 2.6267,
      "theoretical_loss": 3.373863098365129,
      "tokens_seen": 2450915328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026173737373737374,
      "loss": 2.5287,
      "theoretical_loss": 3.373855873745302,
      "tokens_seen": 2450980864
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026171717171717175,
      "loss": 2.6813,
      "theoretical_loss": 3.3738486493727375,
      "tokens_seen": 2451046400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616969696969697,
      "loss": 2.6658,
      "theoretical_loss": 3.37384142524742,
      "tokens_seen": 2451111936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026167676767676766,
      "loss": 2.7852,
      "theoretical_loss": 3.3738342013693345,
      "tokens_seen": 2451177472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026165656565656567,
      "loss": 2.7797,
      "theoretical_loss": 3.373826977738466,
      "tokens_seen": 2451243008
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1344624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.707127571105957,
      "objective/train/theoretical_loss": 3.373819754354799,
      "objective/train/tokens_used": 810167776,
      "theoretical_loss": 3.373819754354799,
      "tokens_seen": 2451308544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026163636363636363,
      "loss": 2.7378,
      "theoretical_loss": 3.373819754354799,
      "tokens_seen": 2451308544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026161616161616164,
      "loss": 2.727,
      "theoretical_loss": 3.3738125312183187,
      "tokens_seen": 2451374080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615959595959596,
      "loss": 2.7233,
      "theoretical_loss": 3.3738053083290103,
      "tokens_seen": 2451439616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026157575757575755,
      "loss": 2.7302,
      "theoretical_loss": 3.3737980856868584,
      "tokens_seen": 2451505152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026155555555555556,
      "loss": 2.6963,
      "theoretical_loss": 3.3737908632918483,
      "tokens_seen": 2451570688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615353535353535,
      "loss": 2.624,
      "theoretical_loss": 3.3737836411439646,
      "tokens_seen": 2451636224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026151515151515153,
      "loss": 2.7915,
      "theoretical_loss": 3.373776419243192,
      "tokens_seen": 2451701760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614949494949495,
      "loss": 2.5758,
      "theoretical_loss": 3.3737691975895157,
      "tokens_seen": 2451767296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614747474747475,
      "loss": 2.6416,
      "theoretical_loss": 3.373761976182921,
      "tokens_seen": 2451832832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026145454545454546,
      "loss": 2.7037,
      "theoretical_loss": 3.3737547550233926,
      "tokens_seen": 2451898368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614343434343434,
      "loss": 2.6136,
      "theoretical_loss": 3.3737475341109153,
      "tokens_seen": 2451963904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614141414141414,
      "loss": 2.748,
      "theoretical_loss": 3.373740313445474,
      "tokens_seen": 2452029440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613939393939394,
      "loss": 2.91,
      "theoretical_loss": 3.373733093027054,
      "tokens_seen": 2452094976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613737373737374,
      "loss": 2.6432,
      "theoretical_loss": 3.37372587285564,
      "tokens_seen": 2452160512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026135353535353535,
      "loss": 2.8583,
      "theoretical_loss": 3.373718652931217,
      "tokens_seen": 2452226048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613333333333333,
      "loss": 2.7154,
      "theoretical_loss": 3.37371143325377,
      "tokens_seen": 2452291584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613131313131313,
      "loss": 2.7039,
      "theoretical_loss": 3.3737042138232836,
      "tokens_seen": 2452357120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026129292929292933,
      "loss": 2.8158,
      "theoretical_loss": 3.373696994639743,
      "tokens_seen": 2452422656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612727272727273,
      "loss": 3.048,
      "theoretical_loss": 3.3736897757031334,
      "tokens_seen": 2452488192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026125252525252524,
      "loss": 2.6986,
      "theoretical_loss": 3.3736825570134394,
      "tokens_seen": 2452553728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612323232323232,
      "loss": 2.5484,
      "theoretical_loss": 3.373675338570646,
      "tokens_seen": 2452619264
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612121212121212,
      "loss": 2.9467,
      "theoretical_loss": 3.3736681203747385,
      "tokens_seen": 2452684800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611919191919192,
      "loss": 2.9616,
      "theoretical_loss": 3.3736609024257014,
      "tokens_seen": 2452750336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611717171717172,
      "loss": 2.6523,
      "theoretical_loss": 3.37365368472352,
      "tokens_seen": 2452815872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026115151515151514,
      "loss": 2.6857,
      "theoretical_loss": 3.373646467268179,
      "tokens_seen": 2452881408
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1345199,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.841261625289917,
      "objective/train/theoretical_loss": 3.373639250059664,
      "objective/train/tokens_used": 811806176,
      "theoretical_loss": 3.373639250059664,
      "tokens_seen": 2452946944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026113131313131315,
      "loss": 2.7445,
      "theoretical_loss": 3.373639250059664,
      "tokens_seen": 2452946944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026111111111111116,
      "loss": 2.7205,
      "theoretical_loss": 3.373632033097959,
      "tokens_seen": 2453012480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610909090909091,
      "loss": 2.6992,
      "theoretical_loss": 3.3736248163830496,
      "tokens_seen": 2453078016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026107070707070707,
      "loss": 2.9203,
      "theoretical_loss": 3.3736175999149203,
      "tokens_seen": 2453143552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026105050505050503,
      "loss": 2.6908,
      "theoretical_loss": 3.3736103836935567,
      "tokens_seen": 2453209088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026103030303030304,
      "loss": 2.8833,
      "theoretical_loss": 3.3736031677189433,
      "tokens_seen": 2453274624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026101010101010105,
      "loss": 2.7665,
      "theoretical_loss": 3.3735959519910654,
      "tokens_seen": 2453340160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260989898989899,
      "loss": 2.8848,
      "theoretical_loss": 3.3735887365099075,
      "tokens_seen": 2453405696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026096969696969696,
      "loss": 2.7735,
      "theoretical_loss": 3.373581521275455,
      "tokens_seen": 2453471232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002609494949494949,
      "loss": 2.6306,
      "theoretical_loss": 3.3735743062876926,
      "tokens_seen": 2453536768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260929292929293,
      "loss": 2.8885,
      "theoretical_loss": 3.3735670915466054,
      "tokens_seen": 2453602304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026090909090909094,
      "loss": 2.7228,
      "theoretical_loss": 3.3735598770521786,
      "tokens_seen": 2453667840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608888888888889,
      "loss": 2.6217,
      "theoretical_loss": 3.3735526628043973,
      "tokens_seen": 2453733376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026086868686868686,
      "loss": 2.766,
      "theoretical_loss": 3.3735454488032457,
      "tokens_seen": 2453798912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608484848484848,
      "loss": 2.8502,
      "theoretical_loss": 3.3735382350487093,
      "tokens_seen": 2453864448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608282828282829,
      "loss": 2.6791,
      "theoretical_loss": 3.3735310215407726,
      "tokens_seen": 2453929984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026080808080808084,
      "loss": 3.0647,
      "theoretical_loss": 3.3735238082794217,
      "tokens_seen": 2453995520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607878787878788,
      "loss": 2.9451,
      "theoretical_loss": 3.37351659526464,
      "tokens_seen": 2454061056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026076767676767675,
      "loss": 2.8323,
      "theoretical_loss": 3.3735093824964144,
      "tokens_seen": 2454126592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026074747474747476,
      "loss": 2.836,
      "theoretical_loss": 3.3735021699747283,
      "tokens_seen": 2454192128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026072727272727277,
      "loss": 2.8259,
      "theoretical_loss": 3.373494957699567,
      "tokens_seen": 2454257664
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026070707070707073,
      "loss": 2.6386,
      "theoretical_loss": 3.3734877456709165,
      "tokens_seen": 2454323200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002606868686868687,
      "loss": 2.8226,
      "theoretical_loss": 3.3734805338887606,
      "tokens_seen": 2454388736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026066666666666664,
      "loss": 2.8312,
      "theoretical_loss": 3.3734733223530844,
      "tokens_seen": 2454454272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026064646464646465,
      "loss": 2.6148,
      "theoretical_loss": 3.3734661110638737,
      "tokens_seen": 2454519808
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1346301,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8723297119140625,
      "objective/train/theoretical_loss": 3.373458900021113,
      "objective/train/tokens_used": 813444576,
      "theoretical_loss": 3.373458900021113,
      "tokens_seen": 2454585344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026062626262626266,
      "loss": 2.6411,
      "theoretical_loss": 3.373458900021113,
      "tokens_seen": 2454585344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002606060606060606,
      "loss": 2.7596,
      "theoretical_loss": 3.3734516892247868,
      "tokens_seen": 2454650880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002605858585858586,
      "loss": 2.7109,
      "theoretical_loss": 3.373444478674881,
      "tokens_seen": 2454716416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026056565656565653,
      "loss": 2.668,
      "theoretical_loss": 3.3734372683713802,
      "tokens_seen": 2454781952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026054545454545455,
      "loss": 2.763,
      "theoretical_loss": 3.37343005831427,
      "tokens_seen": 2454847488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026052525252525256,
      "loss": 2.84,
      "theoretical_loss": 3.373422848503534,
      "tokens_seen": 2454913024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002605050505050505,
      "loss": 2.8758,
      "theoretical_loss": 3.373415638939158,
      "tokens_seen": 2454978560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026048484848484847,
      "loss": 2.6563,
      "theoretical_loss": 3.3734084296211275,
      "tokens_seen": 2455044096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604646464646465,
      "loss": 2.7618,
      "theoretical_loss": 3.3734012205494266,
      "tokens_seen": 2455109632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026044444444444444,
      "loss": 2.6284,
      "theoretical_loss": 3.373394011724041,
      "tokens_seen": 2455175168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026042424242424245,
      "loss": 2.6634,
      "theoretical_loss": 3.3733868031449554,
      "tokens_seen": 2455240704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604040404040404,
      "loss": 2.6675,
      "theoretical_loss": 3.373379594812155,
      "tokens_seen": 2455306240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026038383838383836,
      "loss": 2.7761,
      "theoretical_loss": 3.3733723867256247,
      "tokens_seen": 2455371776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603636363636364,
      "loss": 2.6757,
      "theoretical_loss": 3.373365178885349,
      "tokens_seen": 2455437312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026034343434343433,
      "loss": 2.9078,
      "theoretical_loss": 3.373357971291314,
      "tokens_seen": 2455502848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026032323232323234,
      "loss": 2.8445,
      "theoretical_loss": 3.373350763943504,
      "tokens_seen": 2455568384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603030303030303,
      "loss": 2.8749,
      "theoretical_loss": 3.373343556841904,
      "tokens_seen": 2455633920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602828282828283,
      "loss": 2.9043,
      "theoretical_loss": 3.373336349986499,
      "tokens_seen": 2455699456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026026262626262627,
      "loss": 2.8745,
      "theoretical_loss": 3.3733291433772745,
      "tokens_seen": 2455764992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602424242424242,
      "loss": 2.7883,
      "theoretical_loss": 3.373321937014215,
      "tokens_seen": 2455830528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026022222222222223,
      "loss": 2.9081,
      "theoretical_loss": 3.3733147308973055,
      "tokens_seen": 2455896064
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602020202020202,
      "loss": 2.5358,
      "theoretical_loss": 3.3733075250265316,
      "tokens_seen": 2455961600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002601818181818182,
      "loss": 2.7218,
      "theoretical_loss": 3.373300319401878,
      "tokens_seen": 2456027136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026016161616161616,
      "loss": 2.5663,
      "theoretical_loss": 3.3732931140233293,
      "tokens_seen": 2456092672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002601414141414141,
      "loss": 2.7825,
      "theoretical_loss": 3.373285908890871,
      "tokens_seen": 2456158208
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 1346849,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8223564624786377,
      "objective/train/theoretical_loss": 3.3732787040044885,
      "objective/train/tokens_used": 815082976,
      "theoretical_loss": 3.3732787040044885,
      "tokens_seen": 2456223744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026012121212121213,
      "loss": 2.7546,
      "theoretical_loss": 3.3732787040044885,
      "tokens_seen": 2456223744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026010101010101014,
      "loss": 2.6832,
      "theoretical_loss": 3.373271499364166,
      "tokens_seen": 2456289280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002600808080808081,
      "loss": 2.9912,
      "theoretical_loss": 3.373264294969889,
      "tokens_seen": 2456354816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026006060606060605,
      "loss": 2.7522,
      "theoretical_loss": 3.3732570908216424,
      "tokens_seen": 2456420352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000260040404040404,
      "loss": 2.8512,
      "theoretical_loss": 3.3732498869194116,
      "tokens_seen": 2456485888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000260020202020202,
      "loss": 2.6009,
      "theoretical_loss": 3.373242683263181,
      "tokens_seen": 2456551424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026000000000000003,
      "loss": 2.6549,
      "theoretical_loss": 3.373235479852936,
      "tokens_seen": 2456616960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259979797979798,
      "loss": 2.7297,
      "theoretical_loss": 3.3732282766886614,
      "tokens_seen": 2456682496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025995959595959595,
      "loss": 2.655,
      "theoretical_loss": 3.3732210737703427,
      "tokens_seen": 2456748032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002599393939393939,
      "loss": 2.6077,
      "theoretical_loss": 3.3732138710979647,
      "tokens_seen": 2456813568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025991919191919197,
      "loss": 2.7402,
      "theoretical_loss": 3.3732066686715125,
      "tokens_seen": 2456879104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598989898989899,
      "loss": 2.8478,
      "theoretical_loss": 3.373199466490971,
      "tokens_seen": 2456944640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598787878787879,
      "loss": 2.8656,
      "theoretical_loss": 3.373192264556325,
      "tokens_seen": 2457010176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025985858585858584,
      "loss": 2.7574,
      "theoretical_loss": 3.3731850628675604,
      "tokens_seen": 2457075712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598383838383838,
      "loss": 2.4306,
      "theoretical_loss": 3.373177861424661,
      "tokens_seen": 2457141248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025981818181818186,
      "loss": 2.7078,
      "theoretical_loss": 3.373170660227613,
      "tokens_seen": 2457206784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597979797979798,
      "loss": 2.9213,
      "theoretical_loss": 3.373163459276401,
      "tokens_seen": 2457272320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597777777777778,
      "loss": 2.7051,
      "theoretical_loss": 3.3731562585710098,
      "tokens_seen": 2457337856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025975757575757573,
      "loss": 2.7108,
      "theoretical_loss": 3.3731490581114247,
      "tokens_seen": 2457403392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025973737373737374,
      "loss": 2.8107,
      "theoretical_loss": 3.373141857897631,
      "tokens_seen": 2457468928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025971717171717175,
      "loss": 2.8593,
      "theoretical_loss": 3.3731346579296133,
      "tokens_seen": 2457534464
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002596969696969697,
      "loss": 2.4869,
      "theoretical_loss": 3.373127458207357,
      "tokens_seen": 2457600000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025967676767676767,
      "loss": 2.6961,
      "theoretical_loss": 3.3731202587308475,
      "tokens_seen": 2457665536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002596565656565656,
      "loss": 2.8116,
      "theoretical_loss": 3.373113059500069,
      "tokens_seen": 2457731072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025963636363636363,
      "loss": 2.7153,
      "theoretical_loss": 3.3731058605150066,
      "tokens_seen": 2457796608
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1347964,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6259384155273438,
      "objective/train/theoretical_loss": 3.373098661775646,
      "objective/train/tokens_used": 816721376,
      "theoretical_loss": 3.373098661775646,
      "tokens_seen": 2457862144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025961616161616165,
      "loss": 2.6826,
      "theoretical_loss": 3.373098661775646,
      "tokens_seen": 2457862144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002595959595959596,
      "loss": 2.3772,
      "theoretical_loss": 3.3730914632819715,
      "tokens_seen": 2457927680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025957575757575756,
      "loss": 2.6872,
      "theoretical_loss": 3.3730842650339694,
      "tokens_seen": 2457993216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025955555555555557,
      "loss": 2.7928,
      "theoretical_loss": 3.3730770670316237,
      "tokens_seen": 2458058752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002595353535353535,
      "loss": 2.7637,
      "theoretical_loss": 3.3730698692749197,
      "tokens_seen": 2458124288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025951515151515154,
      "loss": 2.8141,
      "theoretical_loss": 3.3730626717638423,
      "tokens_seen": 2458189824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594949494949495,
      "loss": 2.5939,
      "theoretical_loss": 3.3730554744983774,
      "tokens_seen": 2458255360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025947474747474745,
      "loss": 2.7473,
      "theoretical_loss": 3.373048277478509,
      "tokens_seen": 2458320896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025945454545454546,
      "loss": 2.7952,
      "theoretical_loss": 3.3730410807042226,
      "tokens_seen": 2458386432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594343434343435,
      "loss": 2.7902,
      "theoretical_loss": 3.373033884175504,
      "tokens_seen": 2458451968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025941414141414143,
      "loss": 2.6969,
      "theoretical_loss": 3.3730266878923367,
      "tokens_seen": 2458517504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593939393939394,
      "loss": 2.8438,
      "theoretical_loss": 3.3730194918547074,
      "tokens_seen": 2458583040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593737373737374,
      "loss": 2.776,
      "theoretical_loss": 3.3730122960626,
      "tokens_seen": 2458648576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025935353535353536,
      "loss": 2.7972,
      "theoretical_loss": 3.373005100516,
      "tokens_seen": 2458714112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025933333333333337,
      "loss": 2.5417,
      "theoretical_loss": 3.372997905214893,
      "tokens_seen": 2458779648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593131313131313,
      "loss": 2.757,
      "theoretical_loss": 3.372990710159263,
      "tokens_seen": 2458845184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592929292929293,
      "loss": 2.6848,
      "theoretical_loss": 3.3729835153490964,
      "tokens_seen": 2458910720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592727272727273,
      "loss": 2.4838,
      "theoretical_loss": 3.372976320784377,
      "tokens_seen": 2458976256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025925252525252525,
      "loss": 2.542,
      "theoretical_loss": 3.3729691264650903,
      "tokens_seen": 2459041792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025923232323232326,
      "loss": 2.7042,
      "theoretical_loss": 3.3729619323912217,
      "tokens_seen": 2459107328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592121212121212,
      "loss": 2.6411,
      "theoretical_loss": 3.3729547385627563,
      "tokens_seen": 2459172864
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025919191919191923,
      "loss": 2.6033,
      "theoretical_loss": 3.372947544979679,
      "tokens_seen": 2459238400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591717171717172,
      "loss": 2.7508,
      "theoretical_loss": 3.372940351641975,
      "tokens_seen": 2459303936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025915151515151514,
      "loss": 2.7737,
      "theoretical_loss": 3.372933158549629,
      "tokens_seen": 2459369472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025913131313131315,
      "loss": 2.7519,
      "theoretical_loss": 3.3729259657026263,
      "tokens_seen": 2459435008
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1348509,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.897902011871338,
      "objective/train/theoretical_loss": 3.3729187731009524,
      "objective/train/tokens_used": 818359776,
      "theoretical_loss": 3.3729187731009524,
      "tokens_seen": 2459500544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591111111111111,
      "loss": 2.856,
      "theoretical_loss": 3.3729187731009524,
      "tokens_seen": 2459500544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590909090909091,
      "loss": 2.8222,
      "theoretical_loss": 3.372911580744592,
      "tokens_seen": 2459566080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590707070707071,
      "loss": 2.819,
      "theoretical_loss": 3.37290438863353,
      "tokens_seen": 2459631616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025905050505050503,
      "loss": 2.7431,
      "theoretical_loss": 3.3728971967677523,
      "tokens_seen": 2459697152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025903030303030304,
      "loss": 2.622,
      "theoretical_loss": 3.372890005147243,
      "tokens_seen": 2459762688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025901010101010106,
      "loss": 2.5188,
      "theoretical_loss": 3.3728828137719877,
      "tokens_seen": 2459828224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258989898989899,
      "loss": 2.5795,
      "theoretical_loss": 3.372875622641972,
      "tokens_seen": 2459893760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025896969696969697,
      "loss": 2.7337,
      "theoretical_loss": 3.3728684317571798,
      "tokens_seen": 2459959296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589494949494949,
      "loss": 2.8569,
      "theoretical_loss": 3.3728612411175973,
      "tokens_seen": 2460024832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025892929292929294,
      "loss": 2.7593,
      "theoretical_loss": 3.372854050723209,
      "tokens_seen": 2460090368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025890909090909095,
      "loss": 2.5794,
      "theoretical_loss": 3.372846860574,
      "tokens_seen": 2460155904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588888888888889,
      "loss": 2.6916,
      "theoretical_loss": 3.372839670669956,
      "tokens_seen": 2460221440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025886868686868686,
      "loss": 2.5212,
      "theoretical_loss": 3.3728324810110615,
      "tokens_seen": 2460286976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588484848484848,
      "loss": 2.6511,
      "theoretical_loss": 3.372825291597302,
      "tokens_seen": 2460352512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025882828282828283,
      "loss": 2.7294,
      "theoretical_loss": 3.3728181024286625,
      "tokens_seen": 2460418048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025880808080808084,
      "loss": 2.3928,
      "theoretical_loss": 3.372810913505128,
      "tokens_seen": 2460483584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587878787878788,
      "loss": 2.6117,
      "theoretical_loss": 3.3728037248266833,
      "tokens_seen": 2460549120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025876767676767675,
      "loss": 2.5043,
      "theoretical_loss": 3.372796536393314,
      "tokens_seen": 2460614656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587474747474747,
      "loss": 2.6559,
      "theoretical_loss": 3.3727893482050053,
      "tokens_seen": 2460680192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587272727272728,
      "loss": 2.7152,
      "theoretical_loss": 3.372782160261742,
      "tokens_seen": 2460745728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025870707070707073,
      "loss": 2.7525,
      "theoretical_loss": 3.3727749725635094,
      "tokens_seen": 2460811264
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586868686868687,
      "loss": 2.8658,
      "theoretical_loss": 3.372767785110293,
      "tokens_seen": 2460876800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025866666666666665,
      "loss": 2.7361,
      "theoretical_loss": 3.3727605979020767,
      "tokens_seen": 2460942336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586464646464646,
      "loss": 2.695,
      "theoretical_loss": 3.372753410938847,
      "tokens_seen": 2461007872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025862626262626267,
      "loss": 2.86,
      "theoretical_loss": 3.372746224220588,
      "tokens_seen": 2461073408
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1349876,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.583599328994751,
      "objective/train/theoretical_loss": 3.3727390377472855,
      "objective/train/tokens_used": 819998176,
      "theoretical_loss": 3.3727390377472855,
      "tokens_seen": 2461138944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586060606060606,
      "loss": 2.6283,
      "theoretical_loss": 3.3727390377472855,
      "tokens_seen": 2461138944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002585858585858586,
      "loss": 2.5535,
      "theoretical_loss": 3.372731851518924,
      "tokens_seen": 2461204480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025856565656565654,
      "loss": 2.4328,
      "theoretical_loss": 3.3727246655354897,
      "tokens_seen": 2461270016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025854545454545455,
      "loss": 2.9047,
      "theoretical_loss": 3.3727174797969663,
      "tokens_seen": 2461335552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025852525252525256,
      "loss": 2.6543,
      "theoretical_loss": 3.3727102943033405,
      "tokens_seen": 2461401088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002585050505050505,
      "loss": 2.6586,
      "theoretical_loss": 3.372703109054596,
      "tokens_seen": 2461466624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584848484848485,
      "loss": 2.4107,
      "theoretical_loss": 3.372695924050719,
      "tokens_seen": 2461532160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025846464646464643,
      "loss": 2.6501,
      "theoretical_loss": 3.3726887392916938,
      "tokens_seen": 2461597696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025844444444444444,
      "loss": 2.708,
      "theoretical_loss": 3.372681554777506,
      "tokens_seen": 2461663232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025842424242424246,
      "loss": 2.7723,
      "theoretical_loss": 3.372674370508141,
      "tokens_seen": 2461728768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584040404040404,
      "loss": 2.7073,
      "theoretical_loss": 3.3726671864835835,
      "tokens_seen": 2461794304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025838383838383837,
      "loss": 2.6628,
      "theoretical_loss": 3.3726600027038183,
      "tokens_seen": 2461859840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002583636363636364,
      "loss": 2.61,
      "theoretical_loss": 3.372652819168832,
      "tokens_seen": 2461925376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025834343434343434,
      "loss": 2.5823,
      "theoretical_loss": 3.3726456358786074,
      "tokens_seen": 2461990912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025832323232323235,
      "loss": 2.755,
      "theoretical_loss": 3.3726384528331317,
      "tokens_seen": 2462056448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002583030303030303,
      "loss": 2.5989,
      "theoretical_loss": 3.3726312700323895,
      "tokens_seen": 2462121984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025828282828282826,
      "loss": 2.54,
      "theoretical_loss": 3.3726240874763658,
      "tokens_seen": 2462187520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025826262626262627,
      "loss": 2.929,
      "theoretical_loss": 3.372616905165045,
      "tokens_seen": 2462253056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025824242424242423,
      "loss": 2.5968,
      "theoretical_loss": 3.372609723098414,
      "tokens_seen": 2462318592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025822222222222224,
      "loss": 2.661,
      "theoretical_loss": 3.3726025412764558,
      "tokens_seen": 2462384128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002582020202020202,
      "loss": 2.6983,
      "theoretical_loss": 3.372595359699157,
      "tokens_seen": 2462449664
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581818181818182,
      "loss": 2.6407,
      "theoretical_loss": 3.372588178366503,
      "tokens_seen": 2462515200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025816161616161617,
      "loss": 2.5894,
      "theoretical_loss": 3.372580997278478,
      "tokens_seen": 2462580736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581414141414141,
      "loss": 2.5867,
      "theoretical_loss": 3.3725738164350676,
      "tokens_seen": 2462646272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025812121212121213,
      "loss": 2.8174,
      "theoretical_loss": 3.372566635836257,
      "tokens_seen": 2462711808
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1350584,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0239999294281006,
      "objective/train/theoretical_loss": 3.372559455482031,
      "objective/train/tokens_used": 821636576,
      "theoretical_loss": 3.372559455482031,
      "tokens_seen": 2462777344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581010101010101,
      "loss": 2.8393,
      "theoretical_loss": 3.372559455482031,
      "tokens_seen": 2462777344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002580808080808081,
      "loss": 2.7599,
      "theoretical_loss": 3.3725522753723753,
      "tokens_seen": 2462842880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025806060606060606,
      "loss": 2.9517,
      "theoretical_loss": 3.3725450955072747,
      "tokens_seen": 2462908416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258040404040404,
      "loss": 2.6314,
      "theoretical_loss": 3.3725379158867144,
      "tokens_seen": 2462973952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258020202020202,
      "loss": 2.4321,
      "theoretical_loss": 3.3725307365106794,
      "tokens_seen": 2463039488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025800000000000004,
      "loss": 2.7111,
      "theoretical_loss": 3.3725235573791554,
      "tokens_seen": 2463105024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257979797979798,
      "loss": 2.5913,
      "theoretical_loss": 3.3725163784921275,
      "tokens_seen": 2463170560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025795959595959595,
      "loss": 2.7813,
      "theoretical_loss": 3.37250919984958,
      "tokens_seen": 2463236096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579393939393939,
      "loss": 2.6053,
      "theoretical_loss": 3.372502021451499,
      "tokens_seen": 2463301632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579191919191919,
      "loss": 2.7379,
      "theoretical_loss": 3.3724948432978694,
      "tokens_seen": 2463367168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025789898989898993,
      "loss": 2.5913,
      "theoretical_loss": 3.372487665388676,
      "tokens_seen": 2463432704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002578787878787879,
      "loss": 2.8309,
      "theoretical_loss": 3.3724804877239047,
      "tokens_seen": 2463498240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025785858585858584,
      "loss": 2.5279,
      "theoretical_loss": 3.37247331030354,
      "tokens_seen": 2463563776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002578383838383838,
      "loss": 2.7626,
      "theoretical_loss": 3.372466133127568,
      "tokens_seen": 2463629312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025781818181818187,
      "loss": 2.6182,
      "theoretical_loss": 3.372458956195972,
      "tokens_seen": 2463694848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002577979797979798,
      "loss": 2.5245,
      "theoretical_loss": 3.372451779508739,
      "tokens_seen": 2463760384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002577777777777778,
      "loss": 2.7016,
      "theoretical_loss": 3.372444603065854,
      "tokens_seen": 2463825920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025775757575757574,
      "loss": 2.661,
      "theoretical_loss": 3.3724374268673016,
      "tokens_seen": 2463891456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025773737373737375,
      "loss": 2.4547,
      "theoretical_loss": 3.372430250913067,
      "tokens_seen": 2463956992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025771717171717176,
      "loss": 2.6454,
      "theoretical_loss": 3.3724230752031357,
      "tokens_seen": 2464022528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576969696969697,
      "loss": 2.5513,
      "theoretical_loss": 3.3724158997374922,
      "tokens_seen": 2464088064
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025767676767676767,
      "loss": 2.755,
      "theoretical_loss": 3.3724087245161227,
      "tokens_seen": 2464153600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025765656565656563,
      "loss": 2.6897,
      "theoretical_loss": 3.3724015495390116,
      "tokens_seen": 2464219136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576363636363637,
      "loss": 2.7595,
      "theoretical_loss": 3.3723943748061447,
      "tokens_seen": 2464284672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025761616161616165,
      "loss": 2.3972,
      "theoretical_loss": 3.372387200317507,
      "tokens_seen": 2464350208
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1351934,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.799548864364624,
      "objective/train/theoretical_loss": 3.372380026073083,
      "objective/train/tokens_used": 823274976,
      "theoretical_loss": 3.372380026073083,
      "tokens_seen": 2464415744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575959595959596,
      "loss": 2.624,
      "theoretical_loss": 3.372380026073083,
      "tokens_seen": 2464415744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025757575757575756,
      "loss": 2.8706,
      "theoretical_loss": 3.372372852072859,
      "tokens_seen": 2464481280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575555555555555,
      "loss": 2.7484,
      "theoretical_loss": 3.3723656783168194,
      "tokens_seen": 2464546816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575353535353536,
      "loss": 2.5338,
      "theoretical_loss": 3.3723585048049496,
      "tokens_seen": 2464612352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025751515151515154,
      "loss": 2.51,
      "theoretical_loss": 3.372351331537235,
      "tokens_seen": 2464677888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574949494949495,
      "loss": 2.6639,
      "theoretical_loss": 3.3723441585136604,
      "tokens_seen": 2464743424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025747474747474746,
      "loss": 2.4146,
      "theoretical_loss": 3.3723369857342114,
      "tokens_seen": 2464808960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574545454545454,
      "loss": 2.7182,
      "theoretical_loss": 3.3723298131988733,
      "tokens_seen": 2464874496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574343434343435,
      "loss": 2.8449,
      "theoretical_loss": 3.3723226409076306,
      "tokens_seen": 2464940032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025741414141414144,
      "loss": 2.6352,
      "theoretical_loss": 3.3723154688604695,
      "tokens_seen": 2465005568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002573939393939394,
      "loss": 2.6361,
      "theoretical_loss": 3.372308297057374,
      "tokens_seen": 2465071104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025737373737373735,
      "loss": 2.4359,
      "theoretical_loss": 3.3723011254983306,
      "tokens_seen": 2465136640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025735353535353536,
      "loss": 2.5939,
      "theoretical_loss": 3.3722939541833235,
      "tokens_seen": 2465202176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025733333333333337,
      "loss": 2.6743,
      "theoretical_loss": 3.3722867831123384,
      "tokens_seen": 2465267712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025731313131313133,
      "loss": 2.7261,
      "theoretical_loss": 3.3722796122853604,
      "tokens_seen": 2465333248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572929292929293,
      "loss": 2.5865,
      "theoretical_loss": 3.372272441702375,
      "tokens_seen": 2465398784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025727272727272724,
      "loss": 2.7297,
      "theoretical_loss": 3.3722652713633665,
      "tokens_seen": 2465464320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025725252525252525,
      "loss": 2.5624,
      "theoretical_loss": 3.372258101268321,
      "tokens_seen": 2465529856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025723232323232326,
      "loss": 2.7513,
      "theoretical_loss": 3.372250931417223,
      "tokens_seen": 2465595392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572121212121212,
      "loss": 2.6505,
      "theoretical_loss": 3.3722437618100587,
      "tokens_seen": 2465660928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571919191919192,
      "loss": 2.7273,
      "theoretical_loss": 3.3722365924468125,
      "tokens_seen": 2465726464
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571717171717172,
      "loss": 2.6835,
      "theoretical_loss": 3.3722294233274703,
      "tokens_seen": 2465792000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025715151515151515,
      "loss": 2.893,
      "theoretical_loss": 3.3722222544520166,
      "tokens_seen": 2465857536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025713131313131316,
      "loss": 2.457,
      "theoretical_loss": 3.372215085820437,
      "tokens_seen": 2465923072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571111111111111,
      "loss": 2.6738,
      "theoretical_loss": 3.372207917432717,
      "tokens_seen": 2465988608
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1353245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.568368434906006,
      "objective/train/theoretical_loss": 3.372200749288841,
      "objective/train/tokens_used": 824913376,
      "theoretical_loss": 3.372200749288841,
      "tokens_seen": 2466054144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025709090909090907,
      "loss": 2.7485,
      "theoretical_loss": 3.372200749288841,
      "tokens_seen": 2466054144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570707070707071,
      "loss": 2.6928,
      "theoretical_loss": 3.372193581388795,
      "tokens_seen": 2466119680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025705050505050504,
      "loss": 2.7153,
      "theoretical_loss": 3.372186413732564,
      "tokens_seen": 2466185216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025703030303030305,
      "loss": 2.6976,
      "theoretical_loss": 3.3721792463201328,
      "tokens_seen": 2466250752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257010101010101,
      "loss": 2.6583,
      "theoretical_loss": 3.372172079151487,
      "tokens_seen": 2466316288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256989898989899,
      "loss": 2.3912,
      "theoretical_loss": 3.372164912226612,
      "tokens_seen": 2466381824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256969696969697,
      "loss": 2.4446,
      "theoretical_loss": 3.3721577455454925,
      "tokens_seen": 2466447360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025694949494949493,
      "loss": 2.5305,
      "theoretical_loss": 3.3721505791081148,
      "tokens_seen": 2466512896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025692929292929294,
      "loss": 2.7362,
      "theoretical_loss": 3.372143412914463,
      "tokens_seen": 2466578432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002569090909090909,
      "loss": 2.5031,
      "theoretical_loss": 3.372136246964523,
      "tokens_seen": 2466643968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568888888888889,
      "loss": 2.6468,
      "theoretical_loss": 3.3721290812582794,
      "tokens_seen": 2466709504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025686868686868687,
      "loss": 2.6966,
      "theoretical_loss": 3.372121915795718,
      "tokens_seen": 2466775040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568484848484848,
      "loss": 2.5698,
      "theoretical_loss": 3.3721147505768236,
      "tokens_seen": 2466840576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025682828282828284,
      "loss": 2.6447,
      "theoretical_loss": 3.372107585601582,
      "tokens_seen": 2466906112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025680808080808085,
      "loss": 2.5622,
      "theoretical_loss": 3.3721004208699785,
      "tokens_seen": 2466971648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567878787878788,
      "loss": 2.7665,
      "theoretical_loss": 3.372093256381998,
      "tokens_seen": 2467037184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025676767676767676,
      "loss": 2.4958,
      "theoretical_loss": 3.372086092137625,
      "tokens_seen": 2467102720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567474747474747,
      "loss": 2.7717,
      "theoretical_loss": 3.3720789281368457,
      "tokens_seen": 2467168256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025672727272727273,
      "loss": 2.5001,
      "theoretical_loss": 3.3720717643796454,
      "tokens_seen": 2467233792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025670707070707074,
      "loss": 2.7851,
      "theoretical_loss": 3.3720646008660093,
      "tokens_seen": 2467299328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566868686868687,
      "loss": 2.6388,
      "theoretical_loss": 3.372057437595922,
      "tokens_seen": 2467364864
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025666666666666665,
      "loss": 2.5015,
      "theoretical_loss": 3.3720502745693697,
      "tokens_seen": 2467430400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566464646464646,
      "loss": 2.4632,
      "theoretical_loss": 3.3720431117863363,
      "tokens_seen": 2467495936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566262626262627,
      "loss": 2.6969,
      "theoretical_loss": 3.3720359492468086,
      "tokens_seen": 2467561472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025660606060606063,
      "loss": 2.6717,
      "theoretical_loss": 3.3720287869507706,
      "tokens_seen": 2467627008
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1353826,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4293365478515625,
      "objective/train/theoretical_loss": 3.372021624898209,
      "objective/train/tokens_used": 826551776,
      "theoretical_loss": 3.372021624898209,
      "tokens_seen": 2467692544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565858585858586,
      "loss": 2.7575,
      "theoretical_loss": 3.372021624898209,
      "tokens_seen": 2467692544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025656565656565655,
      "loss": 2.7289,
      "theoretical_loss": 3.3720144630891076,
      "tokens_seen": 2467758080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565454545454545,
      "loss": 2.7331,
      "theoretical_loss": 3.372007301523452,
      "tokens_seen": 2467823616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025652525252525257,
      "loss": 2.6931,
      "theoretical_loss": 3.372000140201228,
      "tokens_seen": 2467889152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565050505050505,
      "loss": 2.6272,
      "theoretical_loss": 3.37199297912242,
      "tokens_seen": 2467954688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002564848484848485,
      "loss": 2.5799,
      "theoretical_loss": 3.3719858182870146,
      "tokens_seen": 2468020224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025646464646464644,
      "loss": 2.6323,
      "theoretical_loss": 3.3719786576949957,
      "tokens_seen": 2468085760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025644444444444445,
      "loss": 2.6646,
      "theoretical_loss": 3.3719714973463493,
      "tokens_seen": 2468151296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025642424242424246,
      "loss": 2.5733,
      "theoretical_loss": 3.3719643372410606,
      "tokens_seen": 2468216832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002564040404040404,
      "loss": 2.7042,
      "theoretical_loss": 3.3719571773791146,
      "tokens_seen": 2468282368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563838383838384,
      "loss": 2.5914,
      "theoretical_loss": 3.371950017760497,
      "tokens_seen": 2468347904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025636363636363633,
      "loss": 2.7883,
      "theoretical_loss": 3.3719428583851925,
      "tokens_seen": 2468413440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025634343434343434,
      "loss": 2.7139,
      "theoretical_loss": 3.3719356992531866,
      "tokens_seen": 2468478976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025632323232323235,
      "loss": 2.6229,
      "theoretical_loss": 3.3719285403644648,
      "tokens_seen": 2468544512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563030303030303,
      "loss": 2.6614,
      "theoretical_loss": 3.3719213817190123,
      "tokens_seen": 2468610048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025628282828282827,
      "loss": 2.5427,
      "theoretical_loss": 3.3719142233168142,
      "tokens_seen": 2468675584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562626262626263,
      "loss": 2.7659,
      "theoretical_loss": 3.371907065157856,
      "tokens_seen": 2468741120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025624242424242424,
      "loss": 2.5301,
      "theoretical_loss": 3.371899907242123,
      "tokens_seen": 2468806656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025622222222222225,
      "loss": 2.6349,
      "theoretical_loss": 3.3718927495696,
      "tokens_seen": 2468872192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562020202020202,
      "loss": 2.6135,
      "theoretical_loss": 3.371885592140273,
      "tokens_seen": 2468937728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025618181818181816,
      "loss": 2.4874,
      "theoretical_loss": 3.3718784349541266,
      "tokens_seen": 2469003264
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025616161616161617,
      "loss": 2.7144,
      "theoretical_loss": 3.371871278011146,
      "tokens_seen": 2469068800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025614141414141413,
      "loss": 2.7758,
      "theoretical_loss": 3.371864121311318,
      "tokens_seen": 2469134336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025612121212121214,
      "loss": 2.6126,
      "theoretical_loss": 3.3718569648546257,
      "tokens_seen": 2469199872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002561010101010101,
      "loss": 2.5413,
      "theoretical_loss": 3.371849808641056,
      "tokens_seen": 2469265408
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1355023,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1884193420410156,
      "objective/train/theoretical_loss": 3.371842652670593,
      "objective/train/tokens_used": 828190176,
      "theoretical_loss": 3.371842652670593,
      "tokens_seen": 2469330944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560808080808081,
      "loss": 2.3851,
      "theoretical_loss": 3.371842652670593,
      "tokens_seen": 2469330944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025606060606060606,
      "loss": 2.5171,
      "theoretical_loss": 3.3718354969432234,
      "tokens_seen": 2469396480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256040404040404,
      "loss": 2.4676,
      "theoretical_loss": 3.371828341458931,
      "tokens_seen": 2469462016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025602020202020203,
      "loss": 2.6147,
      "theoretical_loss": 3.3718211862177023,
      "tokens_seen": 2469527552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256,
      "loss": 2.6157,
      "theoretical_loss": 3.371814031219522,
      "tokens_seen": 2469593088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000255979797979798,
      "loss": 2.5725,
      "theoretical_loss": 3.371806876464375,
      "tokens_seen": 2469658624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025595959595959596,
      "loss": 2.7654,
      "theoretical_loss": 3.371799721952248,
      "tokens_seen": 2469724160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025593939393939397,
      "loss": 2.5189,
      "theoretical_loss": 3.3717925676831246,
      "tokens_seen": 2469789696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002559191919191919,
      "loss": 2.6325,
      "theoretical_loss": 3.3717854136569914,
      "tokens_seen": 2469855232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025589898989898994,
      "loss": 2.7308,
      "theoretical_loss": 3.3717782598738326,
      "tokens_seen": 2469920768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558787878787879,
      "loss": 2.6253,
      "theoretical_loss": 3.3717711063336346,
      "tokens_seen": 2469986304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025585858585858585,
      "loss": 2.5337,
      "theoretical_loss": 3.371763953036382,
      "tokens_seen": 2470051840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025583838383838386,
      "loss": 2.4684,
      "theoretical_loss": 3.37175679998206,
      "tokens_seen": 2470117376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558181818181818,
      "loss": 2.8133,
      "theoretical_loss": 3.3717496471706543,
      "tokens_seen": 2470182912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025579797979797983,
      "loss": 2.5437,
      "theoretical_loss": 3.3717424946021506,
      "tokens_seen": 2470248448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557777777777778,
      "loss": 2.5857,
      "theoretical_loss": 3.3717353422765335,
      "tokens_seen": 2470313984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025575757575757574,
      "loss": 2.5808,
      "theoretical_loss": 3.371728190193788,
      "tokens_seen": 2470379520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025573737373737375,
      "loss": 2.5107,
      "theoretical_loss": 3.3717210383539005,
      "tokens_seen": 2470445056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025571717171717176,
      "loss": 3.0126,
      "theoretical_loss": 3.3717138867568552,
      "tokens_seen": 2470510592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556969696969697,
      "loss": 2.9824,
      "theoretical_loss": 3.3717067354026384,
      "tokens_seen": 2470576128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556767676767677,
      "loss": 2.8434,
      "theoretical_loss": 3.371699584291235,
      "tokens_seen": 2470641664
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025565656565656563,
      "loss": 2.5783,
      "theoretical_loss": 3.37169243342263,
      "tokens_seen": 2470707200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025563636363636365,
      "loss": 2.655,
      "theoretical_loss": 3.371685282796809,
      "tokens_seen": 2470772736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025561616161616166,
      "loss": 2.8134,
      "theoretical_loss": 3.3716781324137575,
      "tokens_seen": 2470838272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555959595959596,
      "loss": 2.4752,
      "theoretical_loss": 3.3716709822734603,
      "tokens_seen": 2470903808
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1355628,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6723058223724365,
      "objective/train/theoretical_loss": 3.371663832375903,
      "objective/train/tokens_used": 829828576,
      "theoretical_loss": 3.371663832375903,
      "tokens_seen": 2470969344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025557575757575757,
      "loss": 2.6169,
      "theoretical_loss": 3.371663832375903,
      "tokens_seen": 2470969344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025555555555555553,
      "loss": 2.7038,
      "theoretical_loss": 3.3716566827210714,
      "tokens_seen": 2471034880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025553535353535354,
      "loss": 2.8198,
      "theoretical_loss": 3.3716495333089505,
      "tokens_seen": 2471100416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025551515151515155,
      "loss": 2.7255,
      "theoretical_loss": 3.371642384139525,
      "tokens_seen": 2471165952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554949494949495,
      "loss": 2.7225,
      "theoretical_loss": 3.371635235212781,
      "tokens_seen": 2471231488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025547474747474746,
      "loss": 2.6788,
      "theoretical_loss": 3.3716280865287036,
      "tokens_seen": 2471297024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554545454545454,
      "loss": 2.5661,
      "theoretical_loss": 3.371620938087278,
      "tokens_seen": 2471362560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554343434343435,
      "loss": 2.3911,
      "theoretical_loss": 3.3716137898884897,
      "tokens_seen": 2471428096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025541414141414144,
      "loss": 2.4517,
      "theoretical_loss": 3.371606641932324,
      "tokens_seen": 2471493632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553939393939394,
      "loss": 2.693,
      "theoretical_loss": 3.371599494218766,
      "tokens_seen": 2471559168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025537373737373736,
      "loss": 2.6192,
      "theoretical_loss": 3.3715923467478013,
      "tokens_seen": 2471624704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553535353535353,
      "loss": 2.586,
      "theoretical_loss": 3.3715851995194153,
      "tokens_seen": 2471690240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553333333333334,
      "loss": 2.5344,
      "theoretical_loss": 3.371578052533593,
      "tokens_seen": 2471755776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025531313131313133,
      "loss": 2.7059,
      "theoretical_loss": 3.3715709057903203,
      "tokens_seen": 2471821312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552929292929293,
      "loss": 2.556,
      "theoretical_loss": 3.3715637592895815,
      "tokens_seen": 2471886848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025527272727272725,
      "loss": 2.6224,
      "theoretical_loss": 3.371556613031363,
      "tokens_seen": 2471952384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025525252525252526,
      "loss": 2.6685,
      "theoretical_loss": 3.3715494670156496,
      "tokens_seen": 2472017920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025523232323232327,
      "loss": 2.5177,
      "theoretical_loss": 3.371542321242427,
      "tokens_seen": 2472083456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025521212121212123,
      "loss": 2.5755,
      "theoretical_loss": 3.3715351757116805,
      "tokens_seen": 2472148992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551919191919192,
      "loss": 2.6052,
      "theoretical_loss": 3.371528030423395,
      "tokens_seen": 2472214528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025517171717171714,
      "loss": 2.5406,
      "theoretical_loss": 3.3715208853775556,
      "tokens_seen": 2472280064
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025515151515151515,
      "loss": 2.6819,
      "theoretical_loss": 3.371513740574149,
      "tokens_seen": 2472345600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025513131313131316,
      "loss": 2.7209,
      "theoretical_loss": 3.3715065960131594,
      "tokens_seen": 2472411136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551111111111111,
      "loss": 2.6593,
      "theoretical_loss": 3.3714994516945724,
      "tokens_seen": 2472476672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002550909090909091,
      "loss": 2.7952,
      "theoretical_loss": 3.3714923076183734,
      "tokens_seen": 2472542208
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 1356720,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1001343727111816,
      "objective/train/theoretical_loss": 3.371485163784548,
      "objective/train/tokens_used": 831466976,
      "theoretical_loss": 3.371485163784548,
      "tokens_seen": 2472607744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002550707070707071,
      "loss": 2.6932,
      "theoretical_loss": 3.371485163784548,
      "tokens_seen": 2472607744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025505050505050504,
      "loss": 2.438,
      "theoretical_loss": 3.371478020193081,
      "tokens_seen": 2472673280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025503030303030306,
      "loss": 2.7169,
      "theoretical_loss": 3.3714708768439583,
      "tokens_seen": 2472738816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000255010101010101,
      "loss": 2.5973,
      "theoretical_loss": 3.371463733737165,
      "tokens_seen": 2472804352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025498989898989897,
      "loss": 2.5285,
      "theoretical_loss": 3.3714565908726866,
      "tokens_seen": 2472869888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254969696969697,
      "loss": 2.835,
      "theoretical_loss": 3.371449448250508,
      "tokens_seen": 2472935424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025494949494949494,
      "loss": 2.6019,
      "theoretical_loss": 3.371442305870615,
      "tokens_seen": 2473000960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025492929292929295,
      "loss": 2.5988,
      "theoretical_loss": 3.3714351637329933,
      "tokens_seen": 2473066496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549090909090909,
      "loss": 2.3961,
      "theoretical_loss": 3.3714280218376276,
      "tokens_seen": 2473132032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548888888888889,
      "loss": 2.8445,
      "theoretical_loss": 3.371420880184503,
      "tokens_seen": 2473197568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548686868686869,
      "loss": 2.6405,
      "theoretical_loss": 3.371413738773606,
      "tokens_seen": 2473263104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025484848484848483,
      "loss": 2.6328,
      "theoretical_loss": 3.371406597604921,
      "tokens_seen": 2473328640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025482828282828284,
      "loss": 2.5317,
      "theoretical_loss": 3.371399456678434,
      "tokens_seen": 2473394176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548080808080808,
      "loss": 2.7039,
      "theoretical_loss": 3.3713923159941297,
      "tokens_seen": 2473459712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547878787878788,
      "loss": 2.6338,
      "theoretical_loss": 3.371385175551994,
      "tokens_seen": 2473525248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025476767676767677,
      "loss": 2.7724,
      "theoretical_loss": 3.371378035352012,
      "tokens_seen": 2473590784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547474747474747,
      "loss": 2.642,
      "theoretical_loss": 3.3713708953941692,
      "tokens_seen": 2473656320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025472727272727273,
      "loss": 2.4035,
      "theoretical_loss": 3.371363755678451,
      "tokens_seen": 2473721856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025470707070707075,
      "loss": 2.6874,
      "theoretical_loss": 3.3713566162048427,
      "tokens_seen": 2473787392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546868686868687,
      "loss": 2.4818,
      "theoretical_loss": 3.3713494769733297,
      "tokens_seen": 2473852928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025466666666666666,
      "loss": 2.5782,
      "theoretical_loss": 3.3713423379838976,
      "tokens_seen": 2473918464
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546464646464646,
      "loss": 2.4878,
      "theoretical_loss": 3.3713351992365315,
      "tokens_seen": 2473984000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546262626262626,
      "loss": 2.7492,
      "theoretical_loss": 3.3713280607312166,
      "tokens_seen": 2474049536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025460606060606064,
      "loss": 2.7206,
      "theoretical_loss": 3.3713209224679384,
      "tokens_seen": 2474115072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545858585858586,
      "loss": 2.7274,
      "theoretical_loss": 3.3713137844466825,
      "tokens_seen": 2474180608
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1357343,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6152400970458984,
      "objective/train/theoretical_loss": 3.3713066466674344,
      "objective/train/tokens_used": 833105376,
      "theoretical_loss": 3.3713066466674344,
      "tokens_seen": 2474246144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025456565656565655,
      "loss": 2.5409,
      "theoretical_loss": 3.3713066466674344,
      "tokens_seen": 2474246144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545454545454545,
      "loss": 2.3603,
      "theoretical_loss": 3.3712995091301794,
      "tokens_seen": 2474311680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545252525252526,
      "loss": 2.6706,
      "theoretical_loss": 3.3712923718349024,
      "tokens_seen": 2474377216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025450505050505053,
      "loss": 2.6257,
      "theoretical_loss": 3.371285234781589,
      "tokens_seen": 2474442752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544848484848485,
      "loss": 2.6054,
      "theoretical_loss": 3.371278097970225,
      "tokens_seen": 2474508288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025446464646464644,
      "loss": 2.608,
      "theoretical_loss": 3.371270961400796,
      "tokens_seen": 2474573824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544444444444444,
      "loss": 2.4753,
      "theoretical_loss": 3.371263825073286,
      "tokens_seen": 2474639360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025442424242424247,
      "loss": 2.6152,
      "theoretical_loss": 3.371256688987682,
      "tokens_seen": 2474704896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544040404040404,
      "loss": 2.5844,
      "theoretical_loss": 3.3712495531439686,
      "tokens_seen": 2474770432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543838383838384,
      "loss": 2.6685,
      "theoretical_loss": 3.371242417542131,
      "tokens_seen": 2474835968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025436363636363634,
      "loss": 2.5292,
      "theoretical_loss": 3.371235282182155,
      "tokens_seen": 2474901504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025434343434343435,
      "loss": 2.0997,
      "theoretical_loss": 3.3712281470640257,
      "tokens_seen": 2474967040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025432323232323236,
      "loss": 2.6647,
      "theoretical_loss": 3.3712210121877293,
      "tokens_seen": 2475032576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543030303030303,
      "loss": 2.5317,
      "theoretical_loss": 3.37121387755325,
      "tokens_seen": 2475098112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025428282828282827,
      "loss": 2.5788,
      "theoretical_loss": 3.371206743160574,
      "tokens_seen": 2475163648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025426262626262623,
      "loss": 2.6145,
      "theoretical_loss": 3.3711996090096865,
      "tokens_seen": 2475229184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542424242424243,
      "loss": 2.5927,
      "theoretical_loss": 3.371192475100573,
      "tokens_seen": 2475294720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025422222222222225,
      "loss": 2.6617,
      "theoretical_loss": 3.3711853414332187,
      "tokens_seen": 2475360256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542020202020202,
      "loss": 2.6614,
      "theoretical_loss": 3.371178208007609,
      "tokens_seen": 2475425792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025418181818181817,
      "loss": 2.6295,
      "theoretical_loss": 3.3711710748237294,
      "tokens_seen": 2475491328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541616161616161,
      "loss": 2.727,
      "theoretical_loss": 3.3711639418815658,
      "tokens_seen": 2475556864
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541414141414142,
      "loss": 2.3632,
      "theoretical_loss": 3.3711568091811026,
      "tokens_seen": 2475622400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025412121212121214,
      "loss": 2.4289,
      "theoretical_loss": 3.371149676722326,
      "tokens_seen": 2475687936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541010101010101,
      "loss": 2.6175,
      "theoretical_loss": 3.371142544505221,
      "tokens_seen": 2475753472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025408080808080806,
      "loss": 2.6898,
      "theoretical_loss": 3.3711354125297732,
      "tokens_seen": 2475819008
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1358722,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.468827486038208,
      "objective/train/theoretical_loss": 3.3711282807959684,
      "objective/train/tokens_used": 834743776,
      "theoretical_loss": 3.3711282807959684,
      "tokens_seen": 2475884544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025406060606060607,
      "loss": 2.6217,
      "theoretical_loss": 3.3711282807959684,
      "tokens_seen": 2475884544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540404040404041,
      "loss": 2.8655,
      "theoretical_loss": 3.371121149303791,
      "tokens_seen": 2475950080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025402020202020204,
      "loss": 2.7529,
      "theoretical_loss": 3.3711140180532273,
      "tokens_seen": 2476015616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254,
      "loss": 2.5568,
      "theoretical_loss": 3.3711068870442626,
      "tokens_seen": 2476081152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025397979797979795,
      "loss": 2.661,
      "theoretical_loss": 3.371099756276882,
      "tokens_seen": 2476146688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025395959595959596,
      "loss": 2.7211,
      "theoretical_loss": 3.3710926257510714,
      "tokens_seen": 2476212224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253939393939394,
      "loss": 2.4742,
      "theoretical_loss": 3.3710854954668155,
      "tokens_seen": 2476277760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025391919191919193,
      "loss": 2.6528,
      "theoretical_loss": 3.3710783654241,
      "tokens_seen": 2476343296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538989898989899,
      "loss": 2.4986,
      "theoretical_loss": 3.371071235622911,
      "tokens_seen": 2476408832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538787878787879,
      "loss": 2.8291,
      "theoretical_loss": 3.371064106063233,
      "tokens_seen": 2476474368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025385858585858585,
      "loss": 2.7524,
      "theoretical_loss": 3.3710569767450522,
      "tokens_seen": 2476539904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025383838383838387,
      "loss": 2.4977,
      "theoretical_loss": 3.3710498476683535,
      "tokens_seen": 2476605440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538181818181818,
      "loss": 2.6181,
      "theoretical_loss": 3.371042718833122,
      "tokens_seen": 2476670976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537979797979798,
      "loss": 2.7013,
      "theoretical_loss": 3.371035590239344,
      "tokens_seen": 2476736512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537777777777778,
      "loss": 2.6835,
      "theoretical_loss": 3.3710284618870046,
      "tokens_seen": 2476802048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025375757575757575,
      "loss": 2.4376,
      "theoretical_loss": 3.3710213337760893,
      "tokens_seen": 2476867584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025373737373737376,
      "loss": 2.6812,
      "theoretical_loss": 3.371014205906583,
      "tokens_seen": 2476933120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537171717171717,
      "loss": 2.5216,
      "theoretical_loss": 3.371007078278472,
      "tokens_seen": 2476998656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536969696969697,
      "loss": 2.9095,
      "theoretical_loss": 3.3709999508917408,
      "tokens_seen": 2477064192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536767676767677,
      "loss": 2.4722,
      "theoretical_loss": 3.3709928237463758,
      "tokens_seen": 2477129728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025365656565656564,
      "loss": 2.3031,
      "theoretical_loss": 3.3709856968423617,
      "tokens_seen": 2477195264
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025363636363636365,
      "loss": 2.7561,
      "theoretical_loss": 3.3709785701796844,
      "tokens_seen": 2477260800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536161616161616,
      "loss": 2.3269,
      "theoretical_loss": 3.370971443758329,
      "tokens_seen": 2477326336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535959595959596,
      "loss": 2.5325,
      "theoretical_loss": 3.3709643175782813,
      "tokens_seen": 2477391872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535757575757576,
      "loss": 2.8376,
      "theoretical_loss": 3.3709571916395262,
      "tokens_seen": 2477457408
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1359325,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4937326908111572,
      "objective/train/theoretical_loss": 3.3709500659420497,
      "objective/train/tokens_used": 836382176,
      "theoretical_loss": 3.3709500659420497,
      "tokens_seen": 2477522944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025355555555555553,
      "loss": 2.4791,
      "theoretical_loss": 3.3709500659420497,
      "tokens_seen": 2477522944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025353535353535354,
      "loss": 2.5437,
      "theoretical_loss": 3.370942940485837,
      "tokens_seen": 2477588480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025351515151515155,
      "loss": 2.6954,
      "theoretical_loss": 3.3709358152708737,
      "tokens_seen": 2477654016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534949494949495,
      "loss": 2.673,
      "theoretical_loss": 3.3709286902971454,
      "tokens_seen": 2477719552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025347474747474747,
      "loss": 2.6018,
      "theoretical_loss": 3.370921565564637,
      "tokens_seen": 2477785088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534545454545454,
      "loss": 2.4899,
      "theoretical_loss": 3.3709144410733343,
      "tokens_seen": 2477850624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025343434343434344,
      "loss": 2.6202,
      "theoretical_loss": 3.3709073168232226,
      "tokens_seen": 2477916160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025341414141414145,
      "loss": 2.4644,
      "theoretical_loss": 3.3709001928142874,
      "tokens_seen": 2477981696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533939393939394,
      "loss": 2.613,
      "theoretical_loss": 3.3708930690465144,
      "tokens_seen": 2478047232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025337373737373736,
      "loss": 2.6285,
      "theoretical_loss": 3.3708859455198885,
      "tokens_seen": 2478112768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533535353535353,
      "loss": 2.6962,
      "theoretical_loss": 3.370878822234396,
      "tokens_seen": 2478178304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533333333333334,
      "loss": 2.5714,
      "theoretical_loss": 3.3708716991900216,
      "tokens_seen": 2478243840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025331313131313134,
      "loss": 2.7378,
      "theoretical_loss": 3.3708645763867513,
      "tokens_seen": 2478309376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532929292929293,
      "loss": 2.7543,
      "theoretical_loss": 3.3708574538245704,
      "tokens_seen": 2478374912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025327272727272725,
      "loss": 2.6731,
      "theoretical_loss": 3.370850331503464,
      "tokens_seen": 2478440448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532525252525252,
      "loss": 2.6831,
      "theoretical_loss": 3.370843209423418,
      "tokens_seen": 2478505984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532323232323233,
      "loss": 2.5356,
      "theoretical_loss": 3.3708360875844177,
      "tokens_seen": 2478571520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025321212121212123,
      "loss": 2.7289,
      "theoretical_loss": 3.3708289659864485,
      "tokens_seen": 2478637056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002531919191919192,
      "loss": 2.7892,
      "theoretical_loss": 3.3708218446294964,
      "tokens_seen": 2478702592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025317171717171715,
      "loss": 2.7162,
      "theoretical_loss": 3.3708147235135457,
      "tokens_seen": 2478768128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025315151515151516,
      "loss": 2.5791,
      "theoretical_loss": 3.370807602638583,
      "tokens_seen": 2478833664
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025313131313131317,
      "loss": 2.394,
      "theoretical_loss": 3.3708004820045936,
      "tokens_seen": 2478899200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002531111111111111,
      "loss": 2.4905,
      "theoretical_loss": 3.3707933616115624,
      "tokens_seen": 2478964736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002530909090909091,
      "loss": 2.5083,
      "theoretical_loss": 3.3707862414594754,
      "tokens_seen": 2479030272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025307070707070704,
      "loss": 2.6632,
      "theoretical_loss": 3.370779121548318,
      "tokens_seen": 2479095808
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1360622,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.839552640914917,
      "objective/train/theoretical_loss": 3.370772001878075,
      "objective/train/tokens_used": 838020576,
      "theoretical_loss": 3.370772001878075,
      "tokens_seen": 2479161344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025305050505050505,
      "loss": 2.5514,
      "theoretical_loss": 3.370772001878075,
      "tokens_seen": 2479161344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025303030303030306,
      "loss": 2.5342,
      "theoretical_loss": 3.370764882448733,
      "tokens_seen": 2479226880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253010101010101,
      "loss": 2.3905,
      "theoretical_loss": 3.370757763260277,
      "tokens_seen": 2479292416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252989898989899,
      "loss": 2.6789,
      "theoretical_loss": 3.3707506443126922,
      "tokens_seen": 2479357952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252969696969697,
      "loss": 2.4022,
      "theoretical_loss": 3.3707435256059646,
      "tokens_seen": 2479423488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025294949494949494,
      "loss": 2.3827,
      "theoretical_loss": 3.370736407140079,
      "tokens_seen": 2479489024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025292929292929295,
      "loss": 2.5959,
      "theoretical_loss": 3.3707292889150215,
      "tokens_seen": 2479554560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529090909090909,
      "loss": 2.5813,
      "theoretical_loss": 3.3707221709307773,
      "tokens_seen": 2479620096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025288888888888887,
      "loss": 2.7431,
      "theoretical_loss": 3.370715053187332,
      "tokens_seen": 2479685632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528686868686869,
      "loss": 2.6039,
      "theoretical_loss": 3.370707935684671,
      "tokens_seen": 2479751168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025284848484848484,
      "loss": 2.596,
      "theoretical_loss": 3.37070081842278,
      "tokens_seen": 2479816704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025282828282828285,
      "loss": 2.7236,
      "theoretical_loss": 3.3706937014016445,
      "tokens_seen": 2479882240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528080808080808,
      "loss": 2.3529,
      "theoretical_loss": 3.3706865846212493,
      "tokens_seen": 2479947776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527878787878788,
      "loss": 2.4214,
      "theoretical_loss": 3.3706794680815806,
      "tokens_seen": 2480013312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025276767676767677,
      "loss": 2.4925,
      "theoretical_loss": 3.3706723517826243,
      "tokens_seen": 2480078848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025274747474747473,
      "loss": 2.8759,
      "theoretical_loss": 3.3706652357243643,
      "tokens_seen": 2480144384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025272727272727274,
      "loss": 2.7379,
      "theoretical_loss": 3.3706581199067878,
      "tokens_seen": 2480209920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527070707070707,
      "loss": 2.568,
      "theoretical_loss": 3.3706510043298796,
      "tokens_seen": 2480275456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002526868686868687,
      "loss": 2.6996,
      "theoretical_loss": 3.370643888993625,
      "tokens_seen": 2480340992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025266666666666666,
      "loss": 2.6869,
      "theoretical_loss": 3.37063677389801,
      "tokens_seen": 2480406528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002526464646464646,
      "loss": 2.5509,
      "theoretical_loss": 3.3706296590430194,
      "tokens_seen": 2480472064
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025262626262626263,
      "loss": 2.3146,
      "theoretical_loss": 3.3706225444286395,
      "tokens_seen": 2480537600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025260606060606064,
      "loss": 2.5581,
      "theoretical_loss": 3.370615430054855,
      "tokens_seen": 2480603136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525858585858586,
      "loss": 2.3669,
      "theoretical_loss": 3.370608315921652,
      "tokens_seen": 2480668672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025256565656565656,
      "loss": 2.5651,
      "theoretical_loss": 3.3706012020290164,
      "tokens_seen": 2480734208
    },
    {
      "debugging/Self-BLEU-5": 0.3554135683424447,
      "debugging/distinct-1-grams": 0.8323901098901099,
      "debugging/distinct-2-grams": 0.9677981864682632,
      "debugging/entropy-1-grams": 5.1378514335831085,
      "debugging/entropy-2-grams": 5.562533421675665,
      "debugging/length": 516.75,
      "debugging/num_segments": 4,
      "epoch": 0.5,
      "objective/train/docs_used": 1361770,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.56823992729187,
      "objective/train/theoretical_loss": 3.370594088376933,
      "objective/train/tokens_used": 839658976,
      "theoretical_loss": 3.370594088376933,
      "tokens_seen": 2480799744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025254545454545457,
      "loss": 2.726,
      "theoretical_loss": 3.370594088376933,
      "tokens_seen": 2480799744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525252525252525,
      "loss": 2.3474,
      "theoretical_loss": 3.370586974965387,
      "tokens_seen": 2480865280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025250505050505054,
      "loss": 2.647,
      "theoretical_loss": 3.370579861794365,
      "tokens_seen": 2480930816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524848484848485,
      "loss": 2.3668,
      "theoretical_loss": 3.370572748863851,
      "tokens_seen": 2480996352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025246464646464645,
      "loss": 2.6206,
      "theoretical_loss": 3.3705656361738323,
      "tokens_seen": 2481061888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025244444444444446,
      "loss": 2.5581,
      "theoretical_loss": 3.370558523724293,
      "tokens_seen": 2481127424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025242424242424247,
      "loss": 2.5056,
      "theoretical_loss": 3.370551411515219,
      "tokens_seen": 2481192960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025240404040404043,
      "loss": 2.7892,
      "theoretical_loss": 3.3705442995465966,
      "tokens_seen": 2481258496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523838383838384,
      "loss": 2.4186,
      "theoretical_loss": 3.3705371878184103,
      "tokens_seen": 2481324032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025236363636363634,
      "loss": 2.7285,
      "theoretical_loss": 3.3705300763306463,
      "tokens_seen": 2481389568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025234343434343435,
      "loss": 2.7742,
      "theoretical_loss": 3.3705229650832895,
      "tokens_seen": 2481455104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025232323232323236,
      "loss": 2.548,
      "theoretical_loss": 3.3705158540763263,
      "tokens_seen": 2481520640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523030303030303,
      "loss": 2.5098,
      "theoretical_loss": 3.370508743309741,
      "tokens_seen": 2481586176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522828282828283,
      "loss": 2.6916,
      "theoretical_loss": 3.3705016327835198,
      "tokens_seen": 2481651712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025226262626262624,
      "loss": 2.4806,
      "theoretical_loss": 3.370494522497649,
      "tokens_seen": 2481717248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025224242424242425,
      "loss": 2.6428,
      "theoretical_loss": 3.3704874124521127,
      "tokens_seen": 2481782784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025222222222222226,
      "loss": 2.2933,
      "theoretical_loss": 3.3704803026468975,
      "tokens_seen": 2481848320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522020202020202,
      "loss": 2.8177,
      "theoretical_loss": 3.3704731930819882,
      "tokens_seen": 2481913856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025218181818181817,
      "loss": 2.6725,
      "theoretical_loss": 3.370466083757371,
      "tokens_seen": 2481979392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025216161616161613,
      "loss": 2.7135,
      "theoretical_loss": 3.3704589746730305,
      "tokens_seen": 2482044928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002521414141414142,
      "loss": 2.4809,
      "theoretical_loss": 3.3704518658289535,
      "tokens_seen": 2482110464
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025212121212121215,
      "loss": 2.5915,
      "theoretical_loss": 3.3704447572251244,
      "tokens_seen": 2482176000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002521010101010101,
      "loss": 2.4467,
      "theoretical_loss": 3.3704376488615293,
      "tokens_seen": 2482241536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025208080808080806,
      "loss": 2.6161,
      "theoretical_loss": 3.3704305407381536,
      "tokens_seen": 2482307072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252060606060606,
      "loss": 2.7558,
      "theoretical_loss": 3.370423432854983,
      "tokens_seen": 2482372608
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1362312,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5898001194000244,
      "objective/train/theoretical_loss": 3.370416325212003,
      "objective/train/tokens_used": 841297376,
      "theoretical_loss": 3.370416325212003,
      "tokens_seen": 2482438144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520404040404041,
      "loss": 2.5715,
      "theoretical_loss": 3.370416325212003,
      "tokens_seen": 2482438144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025202020202020204,
      "loss": 2.6876,
      "theoretical_loss": 3.370409217809199,
      "tokens_seen": 2482503680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252,
      "loss": 2.4408,
      "theoretical_loss": 3.3704021106465567,
      "tokens_seen": 2482569216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025197979797979796,
      "loss": 2.6548,
      "theoretical_loss": 3.3703950037240613,
      "tokens_seen": 2482634752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025195959595959597,
      "loss": 2.4615,
      "theoretical_loss": 3.370387897041699,
      "tokens_seen": 2482700288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251939393939394,
      "loss": 2.7206,
      "theoretical_loss": 3.3703807905994547,
      "tokens_seen": 2482765824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025191919191919194,
      "loss": 2.6135,
      "theoretical_loss": 3.370373684397314,
      "tokens_seen": 2482831360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518989898989899,
      "loss": 2.3843,
      "theoretical_loss": 3.370366578435263,
      "tokens_seen": 2482896896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025187878787878785,
      "loss": 2.5069,
      "theoretical_loss": 3.370359472713287,
      "tokens_seen": 2482962432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025185858585858586,
      "loss": 2.6309,
      "theoretical_loss": 3.370352367231371,
      "tokens_seen": 2483027968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025183838383838387,
      "loss": 2.5311,
      "theoretical_loss": 3.3703452619895016,
      "tokens_seen": 2483093504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025181818181818183,
      "loss": 2.6555,
      "theoretical_loss": 3.370338156987663,
      "tokens_seen": 2483159040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517979797979798,
      "loss": 2.6517,
      "theoretical_loss": 3.370331052225842,
      "tokens_seen": 2483224576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517777777777778,
      "loss": 2.77,
      "theoretical_loss": 3.3703239477040237,
      "tokens_seen": 2483290112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025175757575757575,
      "loss": 2.4177,
      "theoretical_loss": 3.3703168434221937,
      "tokens_seen": 2483355648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025173737373737376,
      "loss": 2.5524,
      "theoretical_loss": 3.370309739380337,
      "tokens_seen": 2483421184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517171717171717,
      "loss": 2.6492,
      "theoretical_loss": 3.3703026355784402,
      "tokens_seen": 2483486720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516969696969697,
      "loss": 2.5522,
      "theoretical_loss": 3.3702955320164882,
      "tokens_seen": 2483552256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516767676767677,
      "loss": 2.4527,
      "theoretical_loss": 3.370288428694467,
      "tokens_seen": 2483617792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025165656565656565,
      "loss": 2.5929,
      "theoretical_loss": 3.3702813256123614,
      "tokens_seen": 2483683328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025163636363636366,
      "loss": 2.6087,
      "theoretical_loss": 3.370274222770157,
      "tokens_seen": 2483748864
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516161616161616,
      "loss": 2.519,
      "theoretical_loss": 3.3702671201678407,
      "tokens_seen": 2483814400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515959595959596,
      "loss": 2.545,
      "theoretical_loss": 3.3702600178053967,
      "tokens_seen": 2483879936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515757575757576,
      "loss": 2.6979,
      "theoretical_loss": 3.370252915682811,
      "tokens_seen": 2483945472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025155555555555554,
      "loss": 2.5112,
      "theoretical_loss": 3.3702458138000693,
      "tokens_seen": 2484011008
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1363163,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.647827625274658,
      "objective/train/theoretical_loss": 3.370238712157157,
      "objective/train/tokens_used": 842935776,
      "theoretical_loss": 3.370238712157157,
      "tokens_seen": 2484076544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025153535353535355,
      "loss": 2.5919,
      "theoretical_loss": 3.370238712157157,
      "tokens_seen": 2484076544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515151515151515,
      "loss": 2.5548,
      "theoretical_loss": 3.3702316107540597,
      "tokens_seen": 2484142080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514949494949495,
      "loss": 2.6185,
      "theoretical_loss": 3.3702245095907633,
      "tokens_seen": 2484207616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514747474747475,
      "loss": 2.3203,
      "theoretical_loss": 3.3702174086672527,
      "tokens_seen": 2484273152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025145454545454543,
      "loss": 2.4668,
      "theoretical_loss": 3.3702103079835144,
      "tokens_seen": 2484338688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025143434343434344,
      "loss": 2.4689,
      "theoretical_loss": 3.3702032075395327,
      "tokens_seen": 2484404224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025141414141414145,
      "loss": 2.4303,
      "theoretical_loss": 3.3701961073352944,
      "tokens_seen": 2484469760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513939393939394,
      "loss": 2.6945,
      "theoretical_loss": 3.3701890073707848,
      "tokens_seen": 2484535296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025137373737373737,
      "loss": 2.5638,
      "theoretical_loss": 3.370181907645989,
      "tokens_seen": 2484600832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513535353535353,
      "loss": 2.5574,
      "theoretical_loss": 3.370174808160893,
      "tokens_seen": 2484666368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025133333333333333,
      "loss": 2.4983,
      "theoretical_loss": 3.3701677089154822,
      "tokens_seen": 2484731904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025131313131313135,
      "loss": 2.5216,
      "theoretical_loss": 3.370160609909742,
      "tokens_seen": 2484797440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512929292929293,
      "loss": 2.6948,
      "theoretical_loss": 3.3701535111436587,
      "tokens_seen": 2484862976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025127272727272726,
      "loss": 2.3537,
      "theoretical_loss": 3.3701464126172174,
      "tokens_seen": 2484928512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512525252525252,
      "loss": 2.7894,
      "theoretical_loss": 3.3701393143304035,
      "tokens_seen": 2484994048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512323232323233,
      "loss": 2.7231,
      "theoretical_loss": 3.370132216283203,
      "tokens_seen": 2485059584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025121212121212124,
      "loss": 2.5109,
      "theoretical_loss": 3.370125118475601,
      "tokens_seen": 2485125120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511919191919192,
      "loss": 2.531,
      "theoretical_loss": 3.370118020907584,
      "tokens_seen": 2485190656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025117171717171715,
      "loss": 2.4603,
      "theoretical_loss": 3.370110923579136,
      "tokens_seen": 2485256192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511515151515151,
      "loss": 2.4416,
      "theoretical_loss": 3.3701038264902445,
      "tokens_seen": 2485321728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511313131313132,
      "loss": 2.6198,
      "theoretical_loss": 3.3700967296408937,
      "tokens_seen": 2485387264
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025111111111111113,
      "loss": 2.5903,
      "theoretical_loss": 3.37008963303107,
      "tokens_seen": 2485452800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002510909090909091,
      "loss": 2.5919,
      "theoretical_loss": 3.3700825366607585,
      "tokens_seen": 2485518336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025107070707070704,
      "loss": 2.5818,
      "theoretical_loss": 3.370075440529945,
      "tokens_seen": 2485583872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025105050505050506,
      "loss": 2.5391,
      "theoretical_loss": 3.370068344638615,
      "tokens_seen": 2485649408
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1363769,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.790882110595703,
      "objective/train/theoretical_loss": 3.370061248986754,
      "objective/train/tokens_used": 844574176,
      "theoretical_loss": 3.370061248986754,
      "tokens_seen": 2485714944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025103030303030307,
      "loss": 2.7666,
      "theoretical_loss": 3.370061248986754,
      "tokens_seen": 2485714944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251010101010101,
      "loss": 2.7097,
      "theoretical_loss": 3.3700541535743485,
      "tokens_seen": 2485780480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000250989898989899,
      "loss": 2.366,
      "theoretical_loss": 3.370047058401383,
      "tokens_seen": 2485846016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025096969696969694,
      "loss": 2.7431,
      "theoretical_loss": 3.3700399634678435,
      "tokens_seen": 2485911552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025094949494949495,
      "loss": 2.4257,
      "theoretical_loss": 3.3700328687737153,
      "tokens_seen": 2485977088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025092929292929296,
      "loss": 2.5628,
      "theoretical_loss": 3.370025774318985,
      "tokens_seen": 2486042624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509090909090909,
      "loss": 2.6971,
      "theoretical_loss": 3.3700186801036374,
      "tokens_seen": 2486108160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508888888888889,
      "loss": 2.6251,
      "theoretical_loss": 3.370011586127658,
      "tokens_seen": 2486173696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025086868686868683,
      "loss": 2.5609,
      "theoretical_loss": 3.3700044923910326,
      "tokens_seen": 2486239232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508484848484849,
      "loss": 2.6603,
      "theoretical_loss": 3.369997398893747,
      "tokens_seen": 2486304768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025082828282828285,
      "loss": 2.6501,
      "theoretical_loss": 3.3699903056357865,
      "tokens_seen": 2486370304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508080808080808,
      "loss": 2.6643,
      "theoretical_loss": 3.3699832126171376,
      "tokens_seen": 2486435840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025078787878787877,
      "loss": 2.6763,
      "theoretical_loss": 3.3699761198377844,
      "tokens_seen": 2486501376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507676767676768,
      "loss": 2.5123,
      "theoretical_loss": 3.369969027297714,
      "tokens_seen": 2486566912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507474747474748,
      "loss": 2.4762,
      "theoretical_loss": 3.369961934996911,
      "tokens_seen": 2486632448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025072727272727275,
      "loss": 2.817,
      "theoretical_loss": 3.369954842935362,
      "tokens_seen": 2486697984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507070707070707,
      "loss": 2.5725,
      "theoretical_loss": 3.3699477511130516,
      "tokens_seen": 2486763520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025068686868686866,
      "loss": 2.3673,
      "theoretical_loss": 3.3699406595299655,
      "tokens_seen": 2486829056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025066666666666667,
      "loss": 2.4388,
      "theoretical_loss": 3.36993356818609,
      "tokens_seen": 2486894592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506464646464647,
      "loss": 2.4689,
      "theoretical_loss": 3.3699264770814104,
      "tokens_seen": 2486960128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025062626262626264,
      "loss": 2.6646,
      "theoretical_loss": 3.3699193862159125,
      "tokens_seen": 2487025664
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506060606060606,
      "loss": 2.5553,
      "theoretical_loss": 3.3699122955895815,
      "tokens_seen": 2487091200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505858585858586,
      "loss": 2.8432,
      "theoretical_loss": 3.3699052052024037,
      "tokens_seen": 2487156736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025056565656565656,
      "loss": 2.72,
      "theoretical_loss": 3.369898115054364,
      "tokens_seen": 2487222272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505454545454546,
      "loss": 2.6203,
      "theoretical_loss": 3.3698910251454484,
      "tokens_seen": 2487287808
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1365266,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5479819774627686,
      "objective/train/theoretical_loss": 3.3698839354756425,
      "objective/train/tokens_used": 846212576,
      "theoretical_loss": 3.3698839354756425,
      "tokens_seen": 2487353344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025052525252525253,
      "loss": 2.4434,
      "theoretical_loss": 3.3698839354756425,
      "tokens_seen": 2487353344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505050505050505,
      "loss": 2.6013,
      "theoretical_loss": 3.369876846044932,
      "tokens_seen": 2487418880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504848484848485,
      "loss": 2.4104,
      "theoretical_loss": 3.3698697568533023,
      "tokens_seen": 2487484416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025046464646464646,
      "loss": 2.7333,
      "theoretical_loss": 3.3698626679007395,
      "tokens_seen": 2487549952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025044444444444447,
      "loss": 2.6047,
      "theoretical_loss": 3.3698555791872287,
      "tokens_seen": 2487615488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504242424242424,
      "loss": 2.4058,
      "theoretical_loss": 3.369848490712756,
      "tokens_seen": 2487681024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025040404040404043,
      "loss": 2.5335,
      "theoretical_loss": 3.3698414024773062,
      "tokens_seen": 2487746560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503838383838384,
      "loss": 2.5476,
      "theoretical_loss": 3.3698343144808662,
      "tokens_seen": 2487812096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025036363636363635,
      "loss": 2.4196,
      "theoretical_loss": 3.369827226723421,
      "tokens_seen": 2487877632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025034343434343436,
      "loss": 2.4556,
      "theoretical_loss": 3.3698201392049563,
      "tokens_seen": 2487943168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503232323232323,
      "loss": 2.5638,
      "theoretical_loss": 3.3698130519254574,
      "tokens_seen": 2488008704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503030303030303,
      "loss": 2.5288,
      "theoretical_loss": 3.3698059648849106,
      "tokens_seen": 2488074240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502828282828283,
      "loss": 2.4758,
      "theoretical_loss": 3.369798878083301,
      "tokens_seen": 2488139776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025026262626262624,
      "loss": 2.7406,
      "theoretical_loss": 3.3697917915206146,
      "tokens_seen": 2488205312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025024242424242425,
      "loss": 2.5244,
      "theoretical_loss": 3.3697847051968366,
      "tokens_seen": 2488270848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025022222222222226,
      "loss": 2.3793,
      "theoretical_loss": 3.3697776191119537,
      "tokens_seen": 2488336384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502020202020202,
      "loss": 2.5258,
      "theoretical_loss": 3.36977053326595,
      "tokens_seen": 2488401920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501818181818182,
      "loss": 2.3719,
      "theoretical_loss": 3.3697634476588125,
      "tokens_seen": 2488467456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025016161616161613,
      "loss": 2.5511,
      "theoretical_loss": 3.369756362290526,
      "tokens_seen": 2488532992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025014141414141414,
      "loss": 2.4933,
      "theoretical_loss": 3.369749277161077,
      "tokens_seen": 2488598528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025012121212121216,
      "loss": 2.6113,
      "theoretical_loss": 3.3697421922704502,
      "tokens_seen": 2488664064
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501010101010101,
      "loss": 2.4248,
      "theoretical_loss": 3.3697351076186317,
      "tokens_seen": 2488729600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025008080808080807,
      "loss": 2.4509,
      "theoretical_loss": 3.3697280232056075,
      "tokens_seen": 2488795136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000250060606060606,
      "loss": 2.5798,
      "theoretical_loss": 3.369720939031363,
      "tokens_seen": 2488860672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002500404040404041,
      "loss": 2.6715,
      "theoretical_loss": 3.3697138550958834,
      "tokens_seen": 2488926208
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 1366649,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3071372509002686,
      "objective/train/theoretical_loss": 3.369706771399155,
      "objective/train/tokens_used": 847850976,
      "theoretical_loss": 3.369706771399155,
      "tokens_seen": 2488991744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025002020202020205,
      "loss": 2.6215,
      "theoretical_loss": 3.369706771399155,
      "tokens_seen": 2488991744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025,
      "loss": 2.6834,
      "theoretical_loss": 3.3696996879411634,
      "tokens_seen": 2489057280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024997979797979796,
      "loss": 2.6139,
      "theoretical_loss": 3.3696926047218936,
      "tokens_seen": 2489122816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249959595959596,
      "loss": 2.5778,
      "theoretical_loss": 3.3696855217413324,
      "tokens_seen": 2489188352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024993939393939393,
      "loss": 2.294,
      "theoretical_loss": 3.3696784389994643,
      "tokens_seen": 2489253888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024991919191919194,
      "loss": 2.5554,
      "theoretical_loss": 3.369671356496276,
      "tokens_seen": 2489319424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498989898989899,
      "loss": 2.7245,
      "theoretical_loss": 3.369664274231752,
      "tokens_seen": 2489384960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024987878787878785,
      "loss": 2.5064,
      "theoretical_loss": 3.36965719220588,
      "tokens_seen": 2489450496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024985858585858587,
      "loss": 2.6125,
      "theoretical_loss": 3.369650110418643,
      "tokens_seen": 2489516032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498383838383838,
      "loss": 2.333,
      "theoretical_loss": 3.3696430288700285,
      "tokens_seen": 2489581568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024981818181818183,
      "loss": 2.4657,
      "theoretical_loss": 3.369635947560022,
      "tokens_seen": 2489647104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497979797979798,
      "loss": 2.4729,
      "theoretical_loss": 3.3696288664886085,
      "tokens_seen": 2489712640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024977777777777775,
      "loss": 2.5708,
      "theoretical_loss": 3.369621785655774,
      "tokens_seen": 2489778176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024975757575757576,
      "loss": 2.447,
      "theoretical_loss": 3.3696147050615046,
      "tokens_seen": 2489843712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497373737373737,
      "loss": 2.6113,
      "theoretical_loss": 3.3696076247057856,
      "tokens_seen": 2489909248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497171717171717,
      "loss": 2.6729,
      "theoretical_loss": 3.369600544588603,
      "tokens_seen": 2489974784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496969696969697,
      "loss": 2.5747,
      "theoretical_loss": 3.3695934647099413,
      "tokens_seen": 2490040320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496767676767677,
      "loss": 2.4099,
      "theoretical_loss": 3.3695863850697876,
      "tokens_seen": 2490105856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024965656565656565,
      "loss": 2.6807,
      "theoretical_loss": 3.369579305668127,
      "tokens_seen": 2490171392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496363636363636,
      "loss": 2.4302,
      "theoretical_loss": 3.3695722265049457,
      "tokens_seen": 2490236928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496161616161616,
      "loss": 2.5694,
      "theoretical_loss": 3.3695651475802286,
      "tokens_seen": 2490302464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495959595959596,
      "loss": 2.4473,
      "theoretical_loss": 3.3695580688939617,
      "tokens_seen": 2490368000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495757575757576,
      "loss": 2.4847,
      "theoretical_loss": 3.3695509904461307,
      "tokens_seen": 2490433536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024955555555555554,
      "loss": 2.8896,
      "theoretical_loss": 3.3695439122367215,
      "tokens_seen": 2490499072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024953535353535355,
      "loss": 2.5302,
      "theoretical_loss": 3.3695368342657197,
      "tokens_seen": 2490564608
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1367045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7037811279296875,
      "objective/train/theoretical_loss": 3.3695297565331104,
      "objective/train/tokens_used": 849489376,
      "theoretical_loss": 3.3695297565331104,
      "tokens_seen": 2490630144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495151515151515,
      "loss": 2.4598,
      "theoretical_loss": 3.3695297565331104,
      "tokens_seen": 2490630144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494949494949495,
      "loss": 2.6655,
      "theoretical_loss": 3.36952267903888,
      "tokens_seen": 2490695680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494747474747475,
      "loss": 2.6655,
      "theoretical_loss": 3.3695156017830143,
      "tokens_seen": 2490761216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024945454545454544,
      "loss": 2.6021,
      "theoretical_loss": 3.3695085247654992,
      "tokens_seen": 2490826752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024943434343434345,
      "loss": 2.4376,
      "theoretical_loss": 3.3695014479863192,
      "tokens_seen": 2490892288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494141414141414,
      "loss": 2.6308,
      "theoretical_loss": 3.369494371445461,
      "tokens_seen": 2490957824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493939393939394,
      "loss": 2.6892,
      "theoretical_loss": 3.36948729514291,
      "tokens_seen": 2491023360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024937373737373737,
      "loss": 2.4959,
      "theoretical_loss": 3.369480219078652,
      "tokens_seen": 2491088896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493535353535354,
      "loss": 2.5588,
      "theoretical_loss": 3.3694731432526726,
      "tokens_seen": 2491154432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024933333333333334,
      "loss": 2.5366,
      "theoretical_loss": 3.3694660676649573,
      "tokens_seen": 2491219968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024931313131313135,
      "loss": 2.6207,
      "theoretical_loss": 3.3694589923154923,
      "tokens_seen": 2491285504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492929292929293,
      "loss": 2.6988,
      "theoretical_loss": 3.369451917204263,
      "tokens_seen": 2491351040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024927272727272727,
      "loss": 2.6747,
      "theoretical_loss": 3.3694448423312555,
      "tokens_seen": 2491416576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492525252525253,
      "loss": 2.5209,
      "theoretical_loss": 3.369437767696455,
      "tokens_seen": 2491482112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024923232323232323,
      "loss": 2.5245,
      "theoretical_loss": 3.3694306932998472,
      "tokens_seen": 2491547648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024921212121212124,
      "loss": 2.5096,
      "theoretical_loss": 3.3694236191414184,
      "tokens_seen": 2491613184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491919191919192,
      "loss": 2.5984,
      "theoretical_loss": 3.3694165452211537,
      "tokens_seen": 2491678720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491717171717172,
      "loss": 2.4263,
      "theoretical_loss": 3.369409471539039,
      "tokens_seen": 2491744256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024915151515151517,
      "loss": 2.5756,
      "theoretical_loss": 3.36940239809506,
      "tokens_seen": 2491809792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491313131313131,
      "loss": 2.6384,
      "theoretical_loss": 3.369395324889203,
      "tokens_seen": 2491875328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024911111111111114,
      "loss": 2.6184,
      "theoretical_loss": 3.369388251921453,
      "tokens_seen": 2491940864
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490909090909091,
      "loss": 2.6661,
      "theoretical_loss": 3.369381179191796,
      "tokens_seen": 2492006400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490707070707071,
      "loss": 2.6451,
      "theoretical_loss": 3.369374106700217,
      "tokens_seen": 2492071936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024905050505050506,
      "loss": 2.5322,
      "theoretical_loss": 3.3693670344467033,
      "tokens_seen": 2492137472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249030303030303,
      "loss": 2.5058,
      "theoretical_loss": 3.3693599624312394,
      "tokens_seen": 2492203008
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1368139,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.515371799468994,
      "objective/train/theoretical_loss": 3.3693528906538113,
      "objective/train/tokens_used": 851127776,
      "theoretical_loss": 3.3693528906538113,
      "tokens_seen": 2492268544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024901010101010103,
      "loss": 2.6865,
      "theoretical_loss": 3.3693528906538113,
      "tokens_seen": 2492268544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248989898989899,
      "loss": 2.3844,
      "theoretical_loss": 3.369345819114405,
      "tokens_seen": 2492334080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248969696969697,
      "loss": 2.5789,
      "theoretical_loss": 3.3693387478130057,
      "tokens_seen": 2492399616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024894949494949495,
      "loss": 2.5623,
      "theoretical_loss": 3.3693316767495993,
      "tokens_seen": 2492465152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489292929292929,
      "loss": 2.757,
      "theoretical_loss": 3.3693246059241724,
      "tokens_seen": 2492530688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489090909090909,
      "loss": 2.8262,
      "theoretical_loss": 3.3693175353367093,
      "tokens_seen": 2492596224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488888888888889,
      "loss": 2.673,
      "theoretical_loss": 3.3693104649871968,
      "tokens_seen": 2492661760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488686868686869,
      "loss": 2.5363,
      "theoretical_loss": 3.36930339487562,
      "tokens_seen": 2492727296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024884848484848485,
      "loss": 2.4885,
      "theoretical_loss": 3.369296325001965,
      "tokens_seen": 2492792832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488282828282828,
      "loss": 2.7316,
      "theoretical_loss": 3.369289255366218,
      "tokens_seen": 2492858368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488080808080808,
      "loss": 2.5233,
      "theoretical_loss": 3.369282185968364,
      "tokens_seen": 2492923904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024878787878787877,
      "loss": 2.7775,
      "theoretical_loss": 3.3692751168083883,
      "tokens_seen": 2492989440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487676767676768,
      "loss": 2.6249,
      "theoretical_loss": 3.369268047886278,
      "tokens_seen": 2493054976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024874747474747474,
      "loss": 2.6697,
      "theoretical_loss": 3.369260979202018,
      "tokens_seen": 2493120512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487272727272727,
      "loss": 2.7337,
      "theoretical_loss": 3.3692539107555937,
      "tokens_seen": 2493186048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487070707070707,
      "loss": 2.3319,
      "theoretical_loss": 3.3692468425469917,
      "tokens_seen": 2493251584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024868686868686866,
      "loss": 2.4852,
      "theoretical_loss": 3.3692397745761973,
      "tokens_seen": 2493317120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486666666666667,
      "loss": 2.6736,
      "theoretical_loss": 3.369232706843196,
      "tokens_seen": 2493382656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024864646464646463,
      "loss": 2.4356,
      "theoretical_loss": 3.3692256393479747,
      "tokens_seen": 2493448192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024862626262626264,
      "loss": 2.637,
      "theoretical_loss": 3.3692185720905172,
      "tokens_seen": 2493513728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486060606060606,
      "loss": 2.476,
      "theoretical_loss": 3.369211505070811,
      "tokens_seen": 2493579264
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024858585858585856,
      "loss": 2.3298,
      "theoretical_loss": 3.3692044382888415,
      "tokens_seen": 2493644800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024856565656565657,
      "loss": 2.4841,
      "theoretical_loss": 3.3691973717445935,
      "tokens_seen": 2493710336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485454545454545,
      "loss": 2.5156,
      "theoretical_loss": 3.369190305438054,
      "tokens_seen": 2493775872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024852525252525254,
      "loss": 2.6971,
      "theoretical_loss": 3.369183239369208,
      "tokens_seen": 2493841408
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1368726,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9163858890533447,
      "objective/train/theoretical_loss": 3.369176173538041,
      "objective/train/tokens_used": 852766176,
      "theoretical_loss": 3.369176173538041,
      "tokens_seen": 2493906944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485050505050505,
      "loss": 2.6727,
      "theoretical_loss": 3.369176173538041,
      "tokens_seen": 2493906944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002484848484848485,
      "loss": 2.7429,
      "theoretical_loss": 3.36916910794454,
      "tokens_seen": 2493972480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024846464646464646,
      "loss": 2.5531,
      "theoretical_loss": 3.3691620425886897,
      "tokens_seen": 2494038016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024844444444444447,
      "loss": 2.6035,
      "theoretical_loss": 3.369154977470476,
      "tokens_seen": 2494103552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024842424242424243,
      "loss": 2.3908,
      "theoretical_loss": 3.3691479125898844,
      "tokens_seen": 2494169088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002484040404040404,
      "loss": 2.5897,
      "theoretical_loss": 3.369140847946902,
      "tokens_seen": 2494234624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483838383838384,
      "loss": 2.5264,
      "theoretical_loss": 3.3691337835415127,
      "tokens_seen": 2494300160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024836363636363635,
      "loss": 2.6173,
      "theoretical_loss": 3.369126719373704,
      "tokens_seen": 2494365696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024834343434343436,
      "loss": 2.6316,
      "theoretical_loss": 3.36911965544346,
      "tokens_seen": 2494431232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483232323232323,
      "loss": 2.4897,
      "theoretical_loss": 3.369112591750768,
      "tokens_seen": 2494496768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024830303030303033,
      "loss": 2.5915,
      "theoretical_loss": 3.369105528295613,
      "tokens_seen": 2494562304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482828282828283,
      "loss": 2.668,
      "theoretical_loss": 3.3690984650779807,
      "tokens_seen": 2494627840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024826262626262625,
      "loss": 2.4532,
      "theoretical_loss": 3.369091402097857,
      "tokens_seen": 2494693376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024824242424242426,
      "loss": 2.3615,
      "theoretical_loss": 3.3690843393552274,
      "tokens_seen": 2494758912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482222222222222,
      "loss": 2.5497,
      "theoretical_loss": 3.3690772768500787,
      "tokens_seen": 2494824448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482020202020202,
      "loss": 2.7915,
      "theoretical_loss": 3.3690702145823956,
      "tokens_seen": 2494889984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481818181818182,
      "loss": 2.3947,
      "theoretical_loss": 3.3690631525521644,
      "tokens_seen": 2494955520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481616161616162,
      "loss": 2.5586,
      "theoretical_loss": 3.3690560907593707,
      "tokens_seen": 2495021056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024814141414141415,
      "loss": 2.6466,
      "theoretical_loss": 3.369049029204,
      "tokens_seen": 2495086592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024812121212121216,
      "loss": 2.6203,
      "theoretical_loss": 3.369041967886039,
      "tokens_seen": 2495152128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481010101010101,
      "loss": 2.4112,
      "theoretical_loss": 3.3690349068054726,
      "tokens_seen": 2495217664
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002480808080808081,
      "loss": 2.5501,
      "theoretical_loss": 3.3690278459622864,
      "tokens_seen": 2495283200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002480606060606061,
      "loss": 2.6416,
      "theoretical_loss": 3.3690207853564673,
      "tokens_seen": 2495348736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024804040404040404,
      "loss": 2.7519,
      "theoretical_loss": 3.369013724988,
      "tokens_seen": 2495414272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024802020202020205,
      "loss": 2.4194,
      "theoretical_loss": 3.369006664856871,
      "tokens_seen": 2495479808
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1369317,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.740476608276367,
      "objective/train/theoretical_loss": 3.368999604963066,
      "objective/train/tokens_used": 854404576,
      "theoretical_loss": 3.368999604963066,
      "tokens_seen": 2495545344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248,
      "loss": 2.5074,
      "theoretical_loss": 3.368999604963066,
      "tokens_seen": 2495545344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024797979797979797,
      "loss": 2.5753,
      "theoretical_loss": 3.36899254530657,
      "tokens_seen": 2495610880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000247959595959596,
      "loss": 2.6143,
      "theoretical_loss": 3.36898548588737,
      "tokens_seen": 2495676416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024793939393939394,
      "loss": 2.3874,
      "theoretical_loss": 3.3689784267054503,
      "tokens_seen": 2495741952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024791919191919195,
      "loss": 2.6073,
      "theoretical_loss": 3.3689713677607984,
      "tokens_seen": 2495807488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478989898989899,
      "loss": 2.6966,
      "theoretical_loss": 3.368964309053399,
      "tokens_seen": 2495873024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024787878787878786,
      "loss": 2.6409,
      "theoretical_loss": 3.368957250583238,
      "tokens_seen": 2495938560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024785858585858587,
      "loss": 2.5733,
      "theoretical_loss": 3.368950192350302,
      "tokens_seen": 2496004096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024783838383838383,
      "loss": 2.941,
      "theoretical_loss": 3.3689431343545753,
      "tokens_seen": 2496069632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024781818181818184,
      "loss": 2.7823,
      "theoretical_loss": 3.368936076596045,
      "tokens_seen": 2496135168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477979797979798,
      "loss": 2.7744,
      "theoretical_loss": 3.3689290190746966,
      "tokens_seen": 2496200704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477777777777778,
      "loss": 2.7765,
      "theoretical_loss": 3.3689219617905155,
      "tokens_seen": 2496266240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024775757575757576,
      "loss": 2.5725,
      "theoretical_loss": 3.368914904743488,
      "tokens_seen": 2496331776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477373737373737,
      "loss": 2.6258,
      "theoretical_loss": 3.3689078479335994,
      "tokens_seen": 2496397312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024771717171717173,
      "loss": 2.5606,
      "theoretical_loss": 3.368900791360836,
      "tokens_seen": 2496462848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476969696969697,
      "loss": 2.5343,
      "theoretical_loss": 3.3688937350251833,
      "tokens_seen": 2496528384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476767676767677,
      "loss": 2.7174,
      "theoretical_loss": 3.368886678926627,
      "tokens_seen": 2496593920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024765656565656566,
      "loss": 2.4547,
      "theoretical_loss": 3.3688796230651534,
      "tokens_seen": 2496659456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476363636363636,
      "loss": 2.4884,
      "theoretical_loss": 3.3688725674407483,
      "tokens_seen": 2496724992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476161616161616,
      "loss": 2.6397,
      "theoretical_loss": 3.3688655120533966,
      "tokens_seen": 2496790528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475959595959596,
      "loss": 2.7922,
      "theoretical_loss": 3.3688584569030855,
      "tokens_seen": 2496856064
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475757575757576,
      "loss": 2.62,
      "theoretical_loss": 3.368851401989799,
      "tokens_seen": 2496921600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024755555555555555,
      "loss": 2.5287,
      "theoretical_loss": 3.368844347313525,
      "tokens_seen": 2496987136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475353535353535,
      "loss": 2.6028,
      "theoretical_loss": 3.3688372928742476,
      "tokens_seen": 2497052672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475151515151515,
      "loss": 2.4231,
      "theoretical_loss": 3.3688302386719537,
      "tokens_seen": 2497118208
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1369981,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4275753498077393,
      "objective/train/theoretical_loss": 3.3688231847066286,
      "objective/train/tokens_used": 856042976,
      "theoretical_loss": 3.3688231847066286,
      "tokens_seen": 2497183744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474949494949495,
      "loss": 2.4614,
      "theoretical_loss": 3.3688231847066286,
      "tokens_seen": 2497183744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474747474747475,
      "loss": 2.4062,
      "theoretical_loss": 3.3688161309782583,
      "tokens_seen": 2497249280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024745454545454544,
      "loss": 2.5543,
      "theoretical_loss": 3.3688090774868287,
      "tokens_seen": 2497314816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024743434343434345,
      "loss": 2.6275,
      "theoretical_loss": 3.368802024232325,
      "tokens_seen": 2497380352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474141414141414,
      "loss": 2.4845,
      "theoretical_loss": 3.3687949712147343,
      "tokens_seen": 2497445888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024739393939393937,
      "loss": 2.608,
      "theoretical_loss": 3.368787918434041,
      "tokens_seen": 2497511424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473737373737374,
      "loss": 2.6319,
      "theoretical_loss": 3.3687808658902316,
      "tokens_seen": 2497576960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024735353535353533,
      "loss": 2.5962,
      "theoretical_loss": 3.3687738135832923,
      "tokens_seen": 2497642496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024733333333333335,
      "loss": 2.4208,
      "theoretical_loss": 3.3687667615132084,
      "tokens_seen": 2497708032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473131313131313,
      "loss": 2.6172,
      "theoretical_loss": 3.368759709679966,
      "tokens_seen": 2497773568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472929292929293,
      "loss": 2.5665,
      "theoretical_loss": 3.3687526580835505,
      "tokens_seen": 2497839104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024727272727272727,
      "loss": 2.4846,
      "theoretical_loss": 3.368745606723948,
      "tokens_seen": 2497904640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472525252525253,
      "loss": 2.5455,
      "theoretical_loss": 3.3687385556011447,
      "tokens_seen": 2497970176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024723232323232324,
      "loss": 2.822,
      "theoretical_loss": 3.368731504715126,
      "tokens_seen": 2498035712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472121212121212,
      "loss": 2.6857,
      "theoretical_loss": 3.3687244540658776,
      "tokens_seen": 2498101248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471919191919192,
      "loss": 2.5812,
      "theoretical_loss": 3.3687174036533856,
      "tokens_seen": 2498166784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024717171717171716,
      "loss": 2.5216,
      "theoretical_loss": 3.368710353477636,
      "tokens_seen": 2498232320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471515151515152,
      "loss": 2.6513,
      "theoretical_loss": 3.3687033035386147,
      "tokens_seen": 2498297856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024713131313131313,
      "loss": 2.5227,
      "theoretical_loss": 3.3686962538363066,
      "tokens_seen": 2498363392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024711111111111114,
      "loss": 2.525,
      "theoretical_loss": 3.368689204370699,
      "tokens_seen": 2498428928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470909090909091,
      "loss": 2.5978,
      "theoretical_loss": 3.3686821551417765,
      "tokens_seen": 2498494464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470707070707071,
      "loss": 2.6067,
      "theoretical_loss": 3.3686751061495253,
      "tokens_seen": 2498560000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024705050505050507,
      "loss": 2.6635,
      "theoretical_loss": 3.368668057393932,
      "tokens_seen": 2498625536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000247030303030303,
      "loss": 2.4435,
      "theoretical_loss": 3.368661008874981,
      "tokens_seen": 2498691072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024701010101010104,
      "loss": 2.6276,
      "theoretical_loss": 3.3686539605926598,
      "tokens_seen": 2498756608
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1371264,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.608736515045166,
      "objective/train/theoretical_loss": 3.3686469125469527,
      "objective/train/tokens_used": 857681376,
      "theoretical_loss": 3.3686469125469527,
      "tokens_seen": 2498822144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246989898989899,
      "loss": 2.5918,
      "theoretical_loss": 3.3686469125469527,
      "tokens_seen": 2498822144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246969696969697,
      "loss": 2.5816,
      "theoretical_loss": 3.3686398647378466,
      "tokens_seen": 2498887680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024694949494949496,
      "loss": 2.4516,
      "theoretical_loss": 3.3686328171653273,
      "tokens_seen": 2498953216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024692929292929297,
      "loss": 2.7305,
      "theoretical_loss": 3.36862576982938,
      "tokens_seen": 2499018752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024690909090909093,
      "loss": 2.4957,
      "theoretical_loss": 3.368618722729991,
      "tokens_seen": 2499084288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468888888888889,
      "loss": 2.6394,
      "theoretical_loss": 3.3686116758671463,
      "tokens_seen": 2499149824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468686868686869,
      "loss": 2.6072,
      "theoretical_loss": 3.3686046292408314,
      "tokens_seen": 2499215360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024684848484848485,
      "loss": 2.779,
      "theoretical_loss": 3.368597582851032,
      "tokens_seen": 2499280896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024682828282828286,
      "loss": 2.6426,
      "theoretical_loss": 3.368590536697735,
      "tokens_seen": 2499346432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468080808080808,
      "loss": 2.7482,
      "theoretical_loss": 3.368583490780925,
      "tokens_seen": 2499411968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467878787878788,
      "loss": 2.4946,
      "theoretical_loss": 3.3685764451005884,
      "tokens_seen": 2499477504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467676767676768,
      "loss": 2.7292,
      "theoretical_loss": 3.368569399656711,
      "tokens_seen": 2499543040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024674747474747475,
      "loss": 2.4935,
      "theoretical_loss": 3.368562354449279,
      "tokens_seen": 2499608576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024672727272727276,
      "loss": 2.5991,
      "theoretical_loss": 3.368555309478278,
      "tokens_seen": 2499674112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467070707070707,
      "loss": 2.7613,
      "theoretical_loss": 3.3685482647436937,
      "tokens_seen": 2499739648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024668686868686867,
      "loss": 2.5811,
      "theoretical_loss": 3.368541220245512,
      "tokens_seen": 2499805184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466666666666667,
      "loss": 2.6333,
      "theoretical_loss": 3.368534175983719,
      "tokens_seen": 2499870720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024664646464646464,
      "loss": 2.5164,
      "theoretical_loss": 3.3685271319583006,
      "tokens_seen": 2499936256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024662626262626265,
      "loss": 2.6044,
      "theoretical_loss": 3.3685200881692423,
      "tokens_seen": 2500001792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466060606060606,
      "loss": 2.5888,
      "theoretical_loss": 3.36851304461653,
      "tokens_seen": 2500067328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024658585858585856,
      "loss": 2.4215,
      "theoretical_loss": 3.36850600130015,
      "tokens_seen": 2500132864
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465656565656566,
      "loss": 2.6991,
      "theoretical_loss": 3.3684989582200884,
      "tokens_seen": 2500198400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024654545454545453,
      "loss": 2.8688,
      "theoretical_loss": 3.3684919153763304,
      "tokens_seen": 2500263936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024652525252525254,
      "loss": 2.5872,
      "theoretical_loss": 3.368484872768862,
      "tokens_seen": 2500329472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465050505050505,
      "loss": 2.6992,
      "theoretical_loss": 3.3684778303976692,
      "tokens_seen": 2500395008
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1372421,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.858076333999634,
      "objective/train/theoretical_loss": 3.3684707882627376,
      "objective/train/tokens_used": 859319776,
      "theoretical_loss": 3.3684707882627376,
      "tokens_seen": 2500460544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024648484848484846,
      "loss": 2.5673,
      "theoretical_loss": 3.3684707882627376,
      "tokens_seen": 2500460544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024646464646464647,
      "loss": 2.5925,
      "theoretical_loss": 3.3684637463640534,
      "tokens_seen": 2500526080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464444444444444,
      "loss": 2.6613,
      "theoretical_loss": 3.368456704701603,
      "tokens_seen": 2500591616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024642424242424243,
      "loss": 2.6533,
      "theoretical_loss": 3.3684496632753715,
      "tokens_seen": 2500657152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464040404040404,
      "loss": 2.541,
      "theoretical_loss": 3.3684426220853445,
      "tokens_seen": 2500722688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463838383838384,
      "loss": 2.5539,
      "theoretical_loss": 3.368435581131509,
      "tokens_seen": 2500788224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024636363636363636,
      "loss": 2.4263,
      "theoretical_loss": 3.36842854041385,
      "tokens_seen": 2500853760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463434343434343,
      "loss": 2.3381,
      "theoretical_loss": 3.368421499932354,
      "tokens_seen": 2500919296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024632323232323233,
      "loss": 2.5069,
      "theoretical_loss": 3.368414459687006,
      "tokens_seen": 2500984832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463030303030303,
      "loss": 2.641,
      "theoretical_loss": 3.3684074196777924,
      "tokens_seen": 2501050368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462828282828283,
      "loss": 2.6466,
      "theoretical_loss": 3.3684003799047,
      "tokens_seen": 2501115904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024626262626262625,
      "loss": 2.6269,
      "theoretical_loss": 3.368393340367713,
      "tokens_seen": 2501181440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024624242424242426,
      "loss": 2.4895,
      "theoretical_loss": 3.3683863010668187,
      "tokens_seen": 2501246976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462222222222222,
      "loss": 2.6638,
      "theoretical_loss": 3.368379262002002,
      "tokens_seen": 2501312512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024620202020202023,
      "loss": 2.5963,
      "theoretical_loss": 3.3683722231732496,
      "tokens_seen": 2501378048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461818181818182,
      "loss": 2.6803,
      "theoretical_loss": 3.3683651845805467,
      "tokens_seen": 2501443584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024616161616161614,
      "loss": 2.5318,
      "theoretical_loss": 3.3683581462238794,
      "tokens_seen": 2501509120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024614141414141416,
      "loss": 2.5654,
      "theoretical_loss": 3.368351108103234,
      "tokens_seen": 2501574656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461212121212121,
      "loss": 2.638,
      "theoretical_loss": 3.3683440702185963,
      "tokens_seen": 2501640192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461010101010101,
      "loss": 2.6747,
      "theoretical_loss": 3.3683370325699515,
      "tokens_seen": 2501705728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460808080808081,
      "loss": 2.4621,
      "theoretical_loss": 3.3683299951572865,
      "tokens_seen": 2501771264
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460606060606061,
      "loss": 2.8799,
      "theoretical_loss": 3.3683229579805865,
      "tokens_seen": 2501836800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024604040404040405,
      "loss": 2.5774,
      "theoretical_loss": 3.368315921039838,
      "tokens_seen": 2501902336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024602020202020206,
      "loss": 2.6532,
      "theoretical_loss": 3.3683088843350264,
      "tokens_seen": 2501967872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246,
      "loss": 2.5827,
      "theoretical_loss": 3.3683018478661375,
      "tokens_seen": 2502033408
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1372950,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.237164258956909,
      "objective/train/theoretical_loss": 3.3682948116331577,
      "objective/train/tokens_used": 860958176,
      "theoretical_loss": 3.3682948116331577,
      "tokens_seen": 2502098944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245979797979798,
      "loss": 2.4457,
      "theoretical_loss": 3.3682948116331577,
      "tokens_seen": 2502098944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245959595959596,
      "loss": 2.6712,
      "theoretical_loss": 3.3682877756360727,
      "tokens_seen": 2502164480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024593939393939394,
      "loss": 2.7075,
      "theoretical_loss": 3.3682807398748684,
      "tokens_seen": 2502230016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024591919191919195,
      "loss": 2.7458,
      "theoretical_loss": 3.3682737043495306,
      "tokens_seen": 2502295552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458989898989899,
      "loss": 2.5921,
      "theoretical_loss": 3.3682666690600453,
      "tokens_seen": 2502361088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458787878787879,
      "loss": 2.6946,
      "theoretical_loss": 3.3682596340063986,
      "tokens_seen": 2502426624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458585858585859,
      "loss": 2.5563,
      "theoretical_loss": 3.368252599188576,
      "tokens_seen": 2502492160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024583838383838383,
      "loss": 2.5767,
      "theoretical_loss": 3.368245564606564,
      "tokens_seen": 2502557696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024581818181818184,
      "loss": 2.4948,
      "theoretical_loss": 3.3682385302603484,
      "tokens_seen": 2502623232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457979797979798,
      "loss": 2.7634,
      "theoretical_loss": 3.3682314961499147,
      "tokens_seen": 2502688768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457777777777778,
      "loss": 2.7744,
      "theoretical_loss": 3.368224462275249,
      "tokens_seen": 2502754304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024575757575757577,
      "loss": 2.4339,
      "theoretical_loss": 3.368217428636337,
      "tokens_seen": 2502819840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457373737373737,
      "loss": 2.5071,
      "theoretical_loss": 3.3682103952331652,
      "tokens_seen": 2502885376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024571717171717174,
      "loss": 2.3734,
      "theoretical_loss": 3.3682033620657195,
      "tokens_seen": 2502950912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456969696969697,
      "loss": 2.6853,
      "theoretical_loss": 3.368196329133985,
      "tokens_seen": 2503016448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456767676767677,
      "loss": 2.3629,
      "theoretical_loss": 3.368189296437949,
      "tokens_seen": 2503081984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024565656565656566,
      "loss": 2.5044,
      "theoretical_loss": 3.368182263977596,
      "tokens_seen": 2503147520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456363636363636,
      "loss": 2.4397,
      "theoretical_loss": 3.3681752317529123,
      "tokens_seen": 2503213056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024561616161616163,
      "loss": 2.7891,
      "theoretical_loss": 3.3681681997638844,
      "tokens_seen": 2503278592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455959595959596,
      "loss": 2.5209,
      "theoretical_loss": 3.368161168010498,
      "tokens_seen": 2503344128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455757575757576,
      "loss": 2.6058,
      "theoretical_loss": 3.368154136492739,
      "tokens_seen": 2503409664
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024555555555555556,
      "loss": 2.6038,
      "theoretical_loss": 3.368147105210593,
      "tokens_seen": 2503475200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455353535353535,
      "loss": 2.7226,
      "theoretical_loss": 3.368140074164047,
      "tokens_seen": 2503540736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455151515151515,
      "loss": 2.7272,
      "theoretical_loss": 3.3681330433530854,
      "tokens_seen": 2503606272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454949494949495,
      "loss": 2.658,
      "theoretical_loss": 3.3681260127776946,
      "tokens_seen": 2503671808
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1374275,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5912024974823,
      "objective/train/theoretical_loss": 3.3681189824378617,
      "objective/train/tokens_used": 862596576,
      "theoretical_loss": 3.3681189824378617,
      "tokens_seen": 2503737344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454747474747475,
      "loss": 2.703,
      "theoretical_loss": 3.3681189824378617,
      "tokens_seen": 2503737344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024545454545454545,
      "loss": 2.7464,
      "theoretical_loss": 3.3681119523335714,
      "tokens_seen": 2503802880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454343434343434,
      "loss": 2.4792,
      "theoretical_loss": 3.36810492246481,
      "tokens_seen": 2503868416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454141414141414,
      "loss": 2.6586,
      "theoretical_loss": 3.3680978928315635,
      "tokens_seen": 2503933952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024539393939393937,
      "loss": 2.4613,
      "theoretical_loss": 3.368090863433818,
      "tokens_seen": 2503999488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453737373737374,
      "loss": 2.6626,
      "theoretical_loss": 3.368083834271559,
      "tokens_seen": 2504065024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024535353535353534,
      "loss": 2.5153,
      "theoretical_loss": 3.3680768053447725,
      "tokens_seen": 2504130560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024533333333333335,
      "loss": 2.3254,
      "theoretical_loss": 3.3680697766534453,
      "tokens_seen": 2504196096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453131313131313,
      "loss": 2.541,
      "theoretical_loss": 3.3680627481975622,
      "tokens_seen": 2504261632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024529292929292927,
      "loss": 2.6514,
      "theoretical_loss": 3.36805571997711,
      "tokens_seen": 2504327168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452727272727273,
      "loss": 2.5615,
      "theoretical_loss": 3.3680486919920742,
      "tokens_seen": 2504392704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024525252525252523,
      "loss": 2.7122,
      "theoretical_loss": 3.368041664242441,
      "tokens_seen": 2504458240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024523232323232324,
      "loss": 2.5342,
      "theoretical_loss": 3.3680346367281957,
      "tokens_seen": 2504523776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452121212121212,
      "loss": 2.496,
      "theoretical_loss": 3.3680276094493253,
      "tokens_seen": 2504589312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451919191919192,
      "loss": 2.4943,
      "theoretical_loss": 3.368020582405815,
      "tokens_seen": 2504654848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024517171717171717,
      "loss": 2.3659,
      "theoretical_loss": 3.3680135555976514,
      "tokens_seen": 2504720384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451515151515152,
      "loss": 2.6034,
      "theoretical_loss": 3.36800652902482,
      "tokens_seen": 2504785920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024513131313131314,
      "loss": 2.6833,
      "theoretical_loss": 3.367999502687306,
      "tokens_seen": 2504851456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451111111111111,
      "loss": 2.476,
      "theoretical_loss": 3.367992476585097,
      "tokens_seen": 2504916992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450909090909091,
      "loss": 2.4239,
      "theoretical_loss": 3.367985450718178,
      "tokens_seen": 2504982528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024507070707070706,
      "loss": 2.5158,
      "theoretical_loss": 3.367978425086535,
      "tokens_seen": 2505048064
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024505050505050507,
      "loss": 2.4461,
      "theoretical_loss": 3.367971399690154,
      "tokens_seen": 2505113600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024503030303030303,
      "loss": 2.622,
      "theoretical_loss": 3.3679643745290213,
      "tokens_seen": 2505179136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024501010101010104,
      "loss": 2.5459,
      "theoretical_loss": 3.367957349603123,
      "tokens_seen": 2505244672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000244989898989899,
      "loss": 2.5114,
      "theoretical_loss": 3.367950324912444,
      "tokens_seen": 2505310208
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 1374858,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.38922381401062,
      "objective/train/theoretical_loss": 3.367943300456971,
      "objective/train/tokens_used": 864234976,
      "theoretical_loss": 3.367943300456971,
      "tokens_seen": 2505375744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024496969696969695,
      "loss": 2.6752,
      "theoretical_loss": 3.367943300456971,
      "tokens_seen": 2505375744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024494949494949497,
      "loss": 2.4714,
      "theoretical_loss": 3.3679362762366902,
      "tokens_seen": 2505441280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449292929292929,
      "loss": 2.7833,
      "theoretical_loss": 3.3679292522515873,
      "tokens_seen": 2505506816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024490909090909093,
      "loss": 2.6454,
      "theoretical_loss": 3.3679222285016484,
      "tokens_seen": 2505572352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448888888888889,
      "loss": 2.6881,
      "theoretical_loss": 3.367915204986859,
      "tokens_seen": 2505637888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448686868686869,
      "loss": 2.4393,
      "theoretical_loss": 3.3679081817072056,
      "tokens_seen": 2505703424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024484848484848486,
      "loss": 2.6554,
      "theoretical_loss": 3.367901158662674,
      "tokens_seen": 2505768960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024482828282828287,
      "loss": 2.607,
      "theoretical_loss": 3.36789413585325,
      "tokens_seen": 2505834496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448080808080808,
      "loss": 2.3392,
      "theoretical_loss": 3.36788711327892,
      "tokens_seen": 2505900032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447878787878788,
      "loss": 2.3996,
      "theoretical_loss": 3.36788009093967,
      "tokens_seen": 2505965568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447676767676768,
      "loss": 2.4207,
      "theoretical_loss": 3.367873068835485,
      "tokens_seen": 2506031104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024474747474747475,
      "loss": 2.493,
      "theoretical_loss": 3.367866046966353,
      "tokens_seen": 2506096640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024472727272727276,
      "loss": 2.6139,
      "theoretical_loss": 3.367859025332258,
      "tokens_seen": 2506162176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447070707070707,
      "loss": 2.3547,
      "theoretical_loss": 3.3678520039331863,
      "tokens_seen": 2506227712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446868686868687,
      "loss": 2.5617,
      "theoretical_loss": 3.3678449827691246,
      "tokens_seen": 2506293248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446666666666667,
      "loss": 2.6531,
      "theoretical_loss": 3.3678379618400585,
      "tokens_seen": 2506358784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024464646464646464,
      "loss": 2.5635,
      "theoretical_loss": 3.367830941145974,
      "tokens_seen": 2506424320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024462626262626265,
      "loss": 2.5852,
      "theoretical_loss": 3.3678239206868574,
      "tokens_seen": 2506489856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446060606060606,
      "loss": 2.6026,
      "theoretical_loss": 3.3678169004626946,
      "tokens_seen": 2506555392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024458585858585857,
      "loss": 2.6732,
      "theoretical_loss": 3.3678098804734713,
      "tokens_seen": 2506620928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445656565656566,
      "loss": 2.6017,
      "theoretical_loss": 3.3678028607191735,
      "tokens_seen": 2506686464
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024454545454545454,
      "loss": 2.5659,
      "theoretical_loss": 3.3677958411997873,
      "tokens_seen": 2506752000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024452525252525255,
      "loss": 2.4344,
      "theoretical_loss": 3.3677888219152985,
      "tokens_seen": 2506817536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445050505050505,
      "loss": 2.5002,
      "theoretical_loss": 3.367781802865694,
      "tokens_seen": 2506883072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024448484848484846,
      "loss": 2.5845,
      "theoretical_loss": 3.3677747840509586,
      "tokens_seen": 2506948608
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1376185,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2249252796173096,
      "objective/train/theoretical_loss": 3.367767765471079,
      "objective/train/tokens_used": 865873376,
      "theoretical_loss": 3.367767765471079,
      "tokens_seen": 2507014144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024446464646464647,
      "loss": 2.4663,
      "theoretical_loss": 3.367767765471079,
      "tokens_seen": 2507014144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024444444444444443,
      "loss": 2.4525,
      "theoretical_loss": 3.367760747126041,
      "tokens_seen": 2507079680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024442424242424244,
      "loss": 2.6758,
      "theoretical_loss": 3.367753729015831,
      "tokens_seen": 2507145216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444040404040404,
      "loss": 2.4877,
      "theoretical_loss": 3.367746711140434,
      "tokens_seen": 2507210752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443838383838384,
      "loss": 2.6594,
      "theoretical_loss": 3.3677396934998374,
      "tokens_seen": 2507276288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024436363636363636,
      "loss": 2.5969,
      "theoretical_loss": 3.367732676094026,
      "tokens_seen": 2507341824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443434343434343,
      "loss": 2.6704,
      "theoretical_loss": 3.367725658922986,
      "tokens_seen": 2507407360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024432323232323233,
      "loss": 2.6806,
      "theoretical_loss": 3.3677186419867042,
      "tokens_seen": 2507472896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443030303030303,
      "loss": 2.3309,
      "theoretical_loss": 3.3677116252851658,
      "tokens_seen": 2507538432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442828282828283,
      "loss": 2.4701,
      "theoretical_loss": 3.3677046088183573,
      "tokens_seen": 2507603968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024426262626262626,
      "loss": 2.5221,
      "theoretical_loss": 3.367697592586264,
      "tokens_seen": 2507669504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442424242424242,
      "loss": 2.6411,
      "theoretical_loss": 3.367690576588873,
      "tokens_seen": 2507735040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442222222222222,
      "loss": 2.5891,
      "theoretical_loss": 3.3676835608261695,
      "tokens_seen": 2507800576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442020202020202,
      "loss": 2.617,
      "theoretical_loss": 3.36767654529814,
      "tokens_seen": 2507866112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441818181818182,
      "loss": 2.6677,
      "theoretical_loss": 3.36766953000477,
      "tokens_seen": 2507931648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024416161616161615,
      "loss": 2.6607,
      "theoretical_loss": 3.367662514946046,
      "tokens_seen": 2507997184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024414141414141416,
      "loss": 2.4414,
      "theoretical_loss": 3.3676555001219537,
      "tokens_seen": 2508062720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024412121212121215,
      "loss": 2.5747,
      "theoretical_loss": 3.367648485532479,
      "tokens_seen": 2508128256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441010101010101,
      "loss": 2.7421,
      "theoretical_loss": 3.367641471177609,
      "tokens_seen": 2508193792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024408080808080809,
      "loss": 2.4892,
      "theoretical_loss": 3.367634457057328,
      "tokens_seen": 2508259328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024406060606060607,
      "loss": 2.4781,
      "theoretical_loss": 3.367627443171623,
      "tokens_seen": 2508324864
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024404040404040403,
      "loss": 2.4819,
      "theoretical_loss": 3.36762042952048,
      "tokens_seen": 2508390400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024402020202020204,
      "loss": 2.4511,
      "theoretical_loss": 3.367613416103885,
      "tokens_seen": 2508455936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000244,
      "loss": 2.6446,
      "theoretical_loss": 3.3676064029218242,
      "tokens_seen": 2508521472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000243979797979798,
      "loss": 2.5491,
      "theoretical_loss": 3.3675993899742833,
      "tokens_seen": 2508587008
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1376516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9129412174224854,
      "objective/train/theoretical_loss": 3.3675923772612486,
      "objective/train/tokens_used": 867511776,
      "theoretical_loss": 3.3675923772612486,
      "tokens_seen": 2508652544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024395959595959596,
      "loss": 2.923,
      "theoretical_loss": 3.3675923772612486,
      "tokens_seen": 2508652544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024393939393939392,
      "loss": 2.5342,
      "theoretical_loss": 3.367585364782706,
      "tokens_seen": 2508718080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024391919191919193,
      "loss": 2.5773,
      "theoretical_loss": 3.3675783525386414,
      "tokens_seen": 2508783616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002438989898989899,
      "loss": 2.7633,
      "theoretical_loss": 3.367571340529041,
      "tokens_seen": 2508849152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002438787878787879,
      "loss": 2.6471,
      "theoretical_loss": 3.3675643287538906,
      "tokens_seen": 2508914688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024385858585858586,
      "loss": 2.6584,
      "theoretical_loss": 3.3675573172131763,
      "tokens_seen": 2508980224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024383838383838384,
      "loss": 2.7547,
      "theoretical_loss": 3.3675503059068848,
      "tokens_seen": 2509045760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024381818181818182,
      "loss": 2.3317,
      "theoretical_loss": 3.3675432948350013,
      "tokens_seen": 2509111296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437979797979798,
      "loss": 2.6839,
      "theoretical_loss": 3.367536283997512,
      "tokens_seen": 2509176832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437777777777778,
      "loss": 2.5136,
      "theoretical_loss": 3.3675292733944033,
      "tokens_seen": 2509242368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024375757575757575,
      "loss": 2.5828,
      "theoretical_loss": 3.367522263025661,
      "tokens_seen": 2509307904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024373737373737373,
      "loss": 2.7007,
      "theoretical_loss": 3.3675152528912715,
      "tokens_seen": 2509373440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024371717171717172,
      "loss": 2.6331,
      "theoretical_loss": 3.36750824299122,
      "tokens_seen": 2509438976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002436969696969697,
      "loss": 2.7396,
      "theoretical_loss": 3.367501233325493,
      "tokens_seen": 2509504512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024367676767676768,
      "loss": 2.6129,
      "theoretical_loss": 3.3674942238940773,
      "tokens_seen": 2509570048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024365656565656567,
      "loss": 2.6242,
      "theoretical_loss": 3.3674872146969577,
      "tokens_seen": 2509635584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024363636363636362,
      "loss": 2.593,
      "theoretical_loss": 3.367480205734121,
      "tokens_seen": 2509701120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024361616161616164,
      "loss": 2.6288,
      "theoretical_loss": 3.367473197005553,
      "tokens_seen": 2509766656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002435959595959596,
      "loss": 2.4618,
      "theoretical_loss": 3.36746618851124,
      "tokens_seen": 2509832192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024357575757575758,
      "loss": 2.5529,
      "theoretical_loss": 3.3674591802511675,
      "tokens_seen": 2509897728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024355555555555556,
      "loss": 2.6104,
      "theoretical_loss": 3.367452172225322,
      "tokens_seen": 2509963264
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024353535353535352,
      "loss": 2.4727,
      "theoretical_loss": 3.36744516443369,
      "tokens_seen": 2510028800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024351515151515153,
      "loss": 2.5098,
      "theoretical_loss": 3.367438156876257,
      "tokens_seen": 2510094336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024349494949494949,
      "loss": 2.6105,
      "theoretical_loss": 3.3674311495530085,
      "tokens_seen": 2510159872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002434747474747475,
      "loss": 2.6655,
      "theoretical_loss": 3.3674241424639315,
      "tokens_seen": 2510225408
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1376516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8044562339782715,
      "objective/train/theoretical_loss": 3.3674171356090117,
      "objective/train/tokens_used": 869150176,
      "theoretical_loss": 3.3674171356090117,
      "tokens_seen": 2510290944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024345454545454545,
      "loss": 2.6998,
      "theoretical_loss": 3.3674171356090117,
      "tokens_seen": 2510290944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024343434343434346,
      "loss": 2.6286,
      "theoretical_loss": 3.367410128988235,
      "tokens_seen": 2510356480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024341414141414142,
      "loss": 2.6773,
      "theoretical_loss": 3.367403122601588,
      "tokens_seen": 2510422016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024339393939393938,
      "loss": 2.9049,
      "theoretical_loss": 3.367396116449056,
      "tokens_seen": 2510487552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002433737373737374,
      "loss": 2.597,
      "theoretical_loss": 3.3673891105306257,
      "tokens_seen": 2510553088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024335353535353535,
      "loss": 2.7125,
      "theoretical_loss": 3.3673821048462833,
      "tokens_seen": 2510618624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024333333333333336,
      "loss": 2.7389,
      "theoretical_loss": 3.367375099396014,
      "tokens_seen": 2510684160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024331313131313131,
      "loss": 2.5175,
      "theoretical_loss": 3.3673680941798048,
      "tokens_seen": 2510749696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002432929292929293,
      "loss": 2.6552,
      "theoretical_loss": 3.367361089197641,
      "tokens_seen": 2510815232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024327272727272728,
      "loss": 2.5474,
      "theoretical_loss": 3.3673540844495093,
      "tokens_seen": 2510880768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024325252525252527,
      "loss": 2.5989,
      "theoretical_loss": 3.3673470799353953,
      "tokens_seen": 2510946304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024323232323232325,
      "loss": 2.7037,
      "theoretical_loss": 3.367340075655285,
      "tokens_seen": 2511011840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002432121212121212,
      "loss": 2.6618,
      "theoretical_loss": 3.367333071609165,
      "tokens_seen": 2511077376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002431919191919192,
      "loss": 2.7728,
      "theoretical_loss": 3.3673260677970216,
      "tokens_seen": 2511142912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024317171717171717,
      "loss": 2.5152,
      "theoretical_loss": 3.36731906421884,
      "tokens_seen": 2511208448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024315151515151516,
      "loss": 2.6549,
      "theoretical_loss": 3.3673120608746068,
      "tokens_seen": 2511273984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024313131313131314,
      "loss": 2.7633,
      "theoretical_loss": 3.3673050577643076,
      "tokens_seen": 2511339520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024311111111111113,
      "loss": 2.4972,
      "theoretical_loss": 3.367298054887929,
      "tokens_seen": 2511405056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024309090909090908,
      "loss": 2.5136,
      "theoretical_loss": 3.3672910522454575,
      "tokens_seen": 2511470592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002430707070707071,
      "loss": 2.8358,
      "theoretical_loss": 3.367284049836878,
      "tokens_seen": 2511536128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024305050505050505,
      "loss": 2.5822,
      "theoretical_loss": 3.367277047662177,
      "tokens_seen": 2511601664
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024303030303030304,
      "loss": 2.8525,
      "theoretical_loss": 3.3672700457213414,
      "tokens_seen": 2511667200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024301010101010102,
      "loss": 2.6291,
      "theoretical_loss": 3.3672630440143565,
      "tokens_seen": 2511732736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024298989898989898,
      "loss": 2.4815,
      "theoretical_loss": 3.367256042541208,
      "tokens_seen": 2511798272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242969696969697,
      "loss": 2.6621,
      "theoretical_loss": 3.3672490413018834,
      "tokens_seen": 2511863808
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1377404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.514570951461792,
      "objective/train/theoretical_loss": 3.3672420402963676,
      "objective/train/tokens_used": 870788576,
      "theoretical_loss": 3.3672420402963676,
      "tokens_seen": 2511929344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024294949494949494,
      "loss": 2.7543,
      "theoretical_loss": 3.3672420402963676,
      "tokens_seen": 2511929344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024292929292929296,
      "loss": 2.642,
      "theoretical_loss": 3.367235039524647,
      "tokens_seen": 2511994880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002429090909090909,
      "loss": 2.6038,
      "theoretical_loss": 3.3672280389867075,
      "tokens_seen": 2512060416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024288888888888887,
      "loss": 2.6941,
      "theoretical_loss": 3.3672210386825356,
      "tokens_seen": 2512125952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024286868686868688,
      "loss": 2.6416,
      "theoretical_loss": 3.3672140386121177,
      "tokens_seen": 2512191488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024284848484848484,
      "loss": 2.6449,
      "theoretical_loss": 3.3672070387754385,
      "tokens_seen": 2512257024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024282828282828285,
      "loss": 2.6641,
      "theoretical_loss": 3.3672000391724857,
      "tokens_seen": 2512322560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002428080808080808,
      "loss": 2.8028,
      "theoretical_loss": 3.367193039803244,
      "tokens_seen": 2512388096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002427878787878788,
      "loss": 2.6898,
      "theoretical_loss": 3.367186040667701,
      "tokens_seen": 2512453632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024276767676767677,
      "loss": 2.5409,
      "theoretical_loss": 3.3671790417658416,
      "tokens_seen": 2512519168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024274747474747476,
      "loss": 2.5854,
      "theoretical_loss": 3.3671720430976526,
      "tokens_seen": 2512584704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024272727272727274,
      "loss": 2.5927,
      "theoretical_loss": 3.3671650446631194,
      "tokens_seen": 2512650240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002427070707070707,
      "loss": 2.6868,
      "theoretical_loss": 3.367158046462229,
      "tokens_seen": 2512715776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024268686868686868,
      "loss": 2.5709,
      "theoretical_loss": 3.367151048494967,
      "tokens_seen": 2512781312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024266666666666667,
      "loss": 2.6336,
      "theoretical_loss": 3.367144050761319,
      "tokens_seen": 2512846848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024264646464646465,
      "loss": 2.7417,
      "theoretical_loss": 3.367137053261272,
      "tokens_seen": 2512912384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024262626262626263,
      "loss": 2.6892,
      "theoretical_loss": 3.367130055994812,
      "tokens_seen": 2512977920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024260606060606062,
      "loss": 2.532,
      "theoretical_loss": 3.367123058961924,
      "tokens_seen": 2513043456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002425858585858586,
      "loss": 2.52,
      "theoretical_loss": 3.3671160621625957,
      "tokens_seen": 2513108992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024256565656565659,
      "loss": 2.6054,
      "theoretical_loss": 3.3671090655968126,
      "tokens_seen": 2513174528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024254545454545454,
      "loss": 2.4477,
      "theoretical_loss": 3.3671020692645603,
      "tokens_seen": 2513240064
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024252525252525253,
      "loss": 2.7192,
      "theoretical_loss": 3.3670950731658253,
      "tokens_seen": 2513305600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002425050505050505,
      "loss": 2.6857,
      "theoretical_loss": 3.3670880773005942,
      "tokens_seen": 2513371136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424848484848485,
      "loss": 2.5658,
      "theoretical_loss": 3.3670810816688523,
      "tokens_seen": 2513436672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024246464646464648,
      "loss": 2.6159,
      "theoretical_loss": 3.367074086270586,
      "tokens_seen": 2513502208
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1378058,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6538960933685303,
      "objective/train/theoretical_loss": 3.367067091105782,
      "objective/train/tokens_used": 872426976,
      "theoretical_loss": 3.367067091105782,
      "tokens_seen": 2513567744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024244444444444443,
      "loss": 2.7245,
      "theoretical_loss": 3.367067091105782,
      "tokens_seen": 2513567744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024242424242424245,
      "loss": 2.6674,
      "theoretical_loss": 3.3670600961744253,
      "tokens_seen": 2513633280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424040404040404,
      "loss": 2.708,
      "theoretical_loss": 3.367053101476503,
      "tokens_seen": 2513698816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024238383838383841,
      "loss": 2.6521,
      "theoretical_loss": 3.367046107012001,
      "tokens_seen": 2513764352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024236363636363637,
      "loss": 2.6415,
      "theoretical_loss": 3.3670391127809047,
      "tokens_seen": 2513829888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024234343434343433,
      "loss": 2.7366,
      "theoretical_loss": 3.3670321187832015,
      "tokens_seen": 2513895424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024232323232323234,
      "loss": 2.6968,
      "theoretical_loss": 3.3670251250188765,
      "tokens_seen": 2513960960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002423030303030303,
      "loss": 2.6258,
      "theoretical_loss": 3.3670181314879164,
      "tokens_seen": 2514026496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002422828282828283,
      "loss": 2.6262,
      "theoretical_loss": 3.3670111381903065,
      "tokens_seen": 2514092032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024226262626262626,
      "loss": 2.663,
      "theoretical_loss": 3.3670041451260344,
      "tokens_seen": 2514157568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024224242424242425,
      "loss": 2.7351,
      "theoretical_loss": 3.3669971522950846,
      "tokens_seen": 2514223104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024222222222222223,
      "loss": 2.6203,
      "theoretical_loss": 3.3669901596974445,
      "tokens_seen": 2514288640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002422020202020202,
      "loss": 2.6981,
      "theoretical_loss": 3.3669831673330997,
      "tokens_seen": 2514354176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002421818181818182,
      "loss": 2.4737,
      "theoretical_loss": 3.366976175202036,
      "tokens_seen": 2514419712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024216161616161616,
      "loss": 2.7378,
      "theoretical_loss": 3.3669691833042403,
      "tokens_seen": 2514485248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024214141414141414,
      "loss": 2.784,
      "theoretical_loss": 3.3669621916396983,
      "tokens_seen": 2514550784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024212121212121212,
      "loss": 2.5699,
      "theoretical_loss": 3.366955200208396,
      "tokens_seen": 2514616320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002421010101010101,
      "loss": 2.8178,
      "theoretical_loss": 3.36694820901032,
      "tokens_seen": 2514681856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002420808080808081,
      "loss": 2.5103,
      "theoretical_loss": 3.366941218045456,
      "tokens_seen": 2514747392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024206060606060608,
      "loss": 2.6782,
      "theoretical_loss": 3.36693422731379,
      "tokens_seen": 2514812928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024204040404040403,
      "loss": 2.6132,
      "theoretical_loss": 3.366927236815309,
      "tokens_seen": 2514878464
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024202020202020202,
      "loss": 2.7149,
      "theoretical_loss": 3.3669202465499986,
      "tokens_seen": 2514944000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242,
      "loss": 2.565,
      "theoretical_loss": 3.3669132565178446,
      "tokens_seen": 2515009536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024197979797979798,
      "loss": 2.6122,
      "theoretical_loss": 3.366906266718834,
      "tokens_seen": 2515075072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024195959595959597,
      "loss": 2.4542,
      "theoretical_loss": 3.366899277152952,
      "tokens_seen": 2515140608
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1379276,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5083673000335693,
      "objective/train/theoretical_loss": 3.366892287820185,
      "objective/train/tokens_used": 874065376,
      "theoretical_loss": 3.366892287820185,
      "tokens_seen": 2515206144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024193939393939393,
      "loss": 2.5853,
      "theoretical_loss": 3.366892287820185,
      "tokens_seen": 2515206144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024191919191919194,
      "loss": 2.6708,
      "theoretical_loss": 3.36688529872052,
      "tokens_seen": 2515271680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002418989898989899,
      "loss": 2.604,
      "theoretical_loss": 3.366878309853942,
      "tokens_seen": 2515337216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002418787878787879,
      "loss": 2.7872,
      "theoretical_loss": 3.3668713212204375,
      "tokens_seen": 2515402752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024185858585858586,
      "loss": 2.8209,
      "theoretical_loss": 3.3668643328199934,
      "tokens_seen": 2515468288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024183838383838382,
      "loss": 2.7967,
      "theoretical_loss": 3.366857344652595,
      "tokens_seen": 2515533824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024181818181818183,
      "loss": 2.5932,
      "theoretical_loss": 3.3668503567182286,
      "tokens_seen": 2515599360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024179797979797979,
      "loss": 2.5671,
      "theoretical_loss": 3.3668433690168804,
      "tokens_seen": 2515664896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002417777777777778,
      "loss": 2.6938,
      "theoretical_loss": 3.366836381548537,
      "tokens_seen": 2515730432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024175757575757575,
      "loss": 2.6772,
      "theoretical_loss": 3.366829394313184,
      "tokens_seen": 2515795968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024173737373737376,
      "loss": 2.4069,
      "theoretical_loss": 3.3668224073108077,
      "tokens_seen": 2515861504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024171717171717172,
      "loss": 2.7365,
      "theoretical_loss": 3.3668154205413945,
      "tokens_seen": 2515927040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416969696969697,
      "loss": 2.8026,
      "theoretical_loss": 3.3668084340049305,
      "tokens_seen": 2515992576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416767676767677,
      "loss": 2.5881,
      "theoretical_loss": 3.366801447701402,
      "tokens_seen": 2516058112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024165656565656565,
      "loss": 2.4471,
      "theoretical_loss": 3.366794461630794,
      "tokens_seen": 2516123648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024163636363636366,
      "loss": 2.461,
      "theoretical_loss": 3.3667874757930942,
      "tokens_seen": 2516189184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024161616161616161,
      "loss": 2.6363,
      "theoretical_loss": 3.366780490188288,
      "tokens_seen": 2516254720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002415959595959596,
      "loss": 2.6407,
      "theoretical_loss": 3.3667735048163623,
      "tokens_seen": 2516320256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024157575757575758,
      "loss": 2.6058,
      "theoretical_loss": 3.366766519677302,
      "tokens_seen": 2516385792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024155555555555557,
      "loss": 2.7431,
      "theoretical_loss": 3.366759534771094,
      "tokens_seen": 2516451328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024153535353535355,
      "loss": 2.5965,
      "theoretical_loss": 3.3667525500977247,
      "tokens_seen": 2516516864
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024151515151515153,
      "loss": 2.6952,
      "theoretical_loss": 3.36674556565718,
      "tokens_seen": 2516582400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002414949494949495,
      "loss": 2.5901,
      "theoretical_loss": 3.366738581449446,
      "tokens_seen": 2516647936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024147474747474748,
      "loss": 2.4266,
      "theoretical_loss": 3.3667315974745096,
      "tokens_seen": 2516713472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024145454545454546,
      "loss": 2.8792,
      "theoretical_loss": 3.3667246137323557,
      "tokens_seen": 2516779008
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1380577,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4944279193878174,
      "objective/train/theoretical_loss": 3.366717630222971,
      "objective/train/tokens_used": 875703776,
      "theoretical_loss": 3.366717630222971,
      "tokens_seen": 2516844544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024143434343434344,
      "loss": 2.5337,
      "theoretical_loss": 3.366717630222971,
      "tokens_seen": 2516844544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024141414141414143,
      "loss": 2.8,
      "theoretical_loss": 3.366710646946342,
      "tokens_seen": 2516910080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024139393939393938,
      "loss": 2.6108,
      "theoretical_loss": 3.366703663902455,
      "tokens_seen": 2516975616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413737373737374,
      "loss": 2.5537,
      "theoretical_loss": 3.366696681091296,
      "tokens_seen": 2517041152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024135353535353535,
      "loss": 2.6381,
      "theoretical_loss": 3.3666896985128503,
      "tokens_seen": 2517106688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024133333333333334,
      "loss": 2.7723,
      "theoretical_loss": 3.3666827161671056,
      "tokens_seen": 2517172224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024131313131313132,
      "loss": 2.7055,
      "theoretical_loss": 3.366675734054047,
      "tokens_seen": 2517237760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024129292929292928,
      "loss": 2.5981,
      "theoretical_loss": 3.366668752173661,
      "tokens_seen": 2517303296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002412727272727273,
      "loss": 2.6816,
      "theoretical_loss": 3.3666617705259343,
      "tokens_seen": 2517368832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024125252525252524,
      "loss": 2.6118,
      "theoretical_loss": 3.366654789110852,
      "tokens_seen": 2517434368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024123232323232326,
      "loss": 2.6029,
      "theoretical_loss": 3.366647807928401,
      "tokens_seen": 2517499904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002412121212121212,
      "loss": 2.5859,
      "theoretical_loss": 3.3666408269785677,
      "tokens_seen": 2517565440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002411919191919192,
      "loss": 2.737,
      "theoretical_loss": 3.3666338462613377,
      "tokens_seen": 2517630976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024117171717171718,
      "loss": 2.638,
      "theoretical_loss": 3.3666268657766976,
      "tokens_seen": 2517696512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024115151515151514,
      "loss": 2.5856,
      "theoretical_loss": 3.3666198855246336,
      "tokens_seen": 2517762048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024113131313131315,
      "loss": 2.6656,
      "theoretical_loss": 3.3666129055051317,
      "tokens_seen": 2517827584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002411111111111111,
      "loss": 2.5916,
      "theoretical_loss": 3.3666059257181784,
      "tokens_seen": 2517893120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002410909090909091,
      "loss": 2.4775,
      "theoretical_loss": 3.366598946163759,
      "tokens_seen": 2517958656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024107070707070707,
      "loss": 2.5556,
      "theoretical_loss": 3.366591966841861,
      "tokens_seen": 2518024192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024105050505050506,
      "loss": 2.4534,
      "theoretical_loss": 3.3665849877524696,
      "tokens_seen": 2518089728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024103030303030304,
      "loss": 2.6444,
      "theoretical_loss": 3.366578008895572,
      "tokens_seen": 2518155264
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024101010101010102,
      "loss": 2.366,
      "theoretical_loss": 3.366571030271153,
      "tokens_seen": 2518220800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024098989898989898,
      "loss": 2.5635,
      "theoretical_loss": 3.3665640518792,
      "tokens_seen": 2518286336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024096969696969697,
      "loss": 2.5745,
      "theoretical_loss": 3.366557073719699,
      "tokens_seen": 2518351872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024094949494949495,
      "loss": 2.6506,
      "theoretical_loss": 3.3665500957926358,
      "tokens_seen": 2518417408
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1381008,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.737955093383789,
      "objective/train/theoretical_loss": 3.366543118097997,
      "objective/train/tokens_used": 877342176,
      "theoretical_loss": 3.366543118097997,
      "tokens_seen": 2518482944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024092929292929293,
      "loss": 2.5444,
      "theoretical_loss": 3.366543118097997,
      "tokens_seen": 2518482944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024090909090909092,
      "loss": 2.7326,
      "theoretical_loss": 3.366536140635768,
      "tokens_seen": 2518548480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408888888888889,
      "loss": 2.537,
      "theoretical_loss": 3.366529163405936,
      "tokens_seen": 2518614016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024086868686868689,
      "loss": 2.7069,
      "theoretical_loss": 3.366522186408487,
      "tokens_seen": 2518679552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024084848484848484,
      "loss": 2.7154,
      "theoretical_loss": 3.3665152096434072,
      "tokens_seen": 2518745088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024082828282828285,
      "loss": 2.546,
      "theoretical_loss": 3.3665082331106824,
      "tokens_seen": 2518810624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408080808080808,
      "loss": 2.546,
      "theoretical_loss": 3.366501256810299,
      "tokens_seen": 2518876160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002407878787878788,
      "loss": 2.4728,
      "theoretical_loss": 3.3664942807422436,
      "tokens_seen": 2518941696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024076767676767678,
      "loss": 2.4573,
      "theoretical_loss": 3.366487304906502,
      "tokens_seen": 2519007232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024074747474747474,
      "loss": 2.6196,
      "theoretical_loss": 3.3664803293030605,
      "tokens_seen": 2519072768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024072727272727275,
      "loss": 2.4906,
      "theoretical_loss": 3.366473353931905,
      "tokens_seen": 2519138304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002407070707070707,
      "loss": 2.6368,
      "theoretical_loss": 3.366466378793023,
      "tokens_seen": 2519203840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024068686868686871,
      "loss": 2.7033,
      "theoretical_loss": 3.3664594038863993,
      "tokens_seen": 2519269376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024066666666666667,
      "loss": 2.4871,
      "theoretical_loss": 3.366452429212021,
      "tokens_seen": 2519334912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024064646464646463,
      "loss": 2.6816,
      "theoretical_loss": 3.3664454547698734,
      "tokens_seen": 2519400448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024062626262626264,
      "loss": 2.4434,
      "theoretical_loss": 3.366438480559944,
      "tokens_seen": 2519465984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002406060606060606,
      "loss": 2.8236,
      "theoretical_loss": 3.3664315065822175,
      "tokens_seen": 2519531520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002405858585858586,
      "loss": 2.7616,
      "theoretical_loss": 3.3664245328366817,
      "tokens_seen": 2519597056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024056565656565656,
      "loss": 2.523,
      "theoretical_loss": 3.3664175593233217,
      "tokens_seen": 2519662592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024054545454545455,
      "loss": 2.498,
      "theoretical_loss": 3.366410586042124,
      "tokens_seen": 2519728128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024052525252525253,
      "loss": 2.5989,
      "theoretical_loss": 3.366403612993075,
      "tokens_seen": 2519793664
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024050505050505052,
      "loss": 2.6097,
      "theoretical_loss": 3.3663966401761614,
      "tokens_seen": 2519859200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002404848484848485,
      "loss": 2.5093,
      "theoretical_loss": 3.3663896675913683,
      "tokens_seen": 2519924736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024046464646464646,
      "loss": 2.7254,
      "theoretical_loss": 3.366382695238683,
      "tokens_seen": 2519990272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024044444444444444,
      "loss": 2.468,
      "theoretical_loss": 3.366375723118091,
      "tokens_seen": 2520055808
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1382202,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.422396421432495,
      "objective/train/theoretical_loss": 3.366368751229579,
      "objective/train/tokens_used": 878980576,
      "theoretical_loss": 3.366368751229579,
      "tokens_seen": 2520121344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024042424242424242,
      "loss": 2.5207,
      "theoretical_loss": 3.366368751229579,
      "tokens_seen": 2520121344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002404040404040404,
      "loss": 2.5935,
      "theoretical_loss": 3.3663617795731327,
      "tokens_seen": 2520186880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002403838383838384,
      "loss": 2.5413,
      "theoretical_loss": 3.3663548081487393,
      "tokens_seen": 2520252416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024036363636363638,
      "loss": 2.5449,
      "theoretical_loss": 3.366347836956384,
      "tokens_seen": 2520317952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024034343434343433,
      "loss": 2.5786,
      "theoretical_loss": 3.3663408659960536,
      "tokens_seen": 2520383488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024032323232323234,
      "loss": 2.5854,
      "theoretical_loss": 3.3663338952677346,
      "tokens_seen": 2520449024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002403030303030303,
      "loss": 2.5456,
      "theoretical_loss": 3.3663269247714123,
      "tokens_seen": 2520514560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024028282828282828,
      "loss": 2.5977,
      "theoretical_loss": 3.366319954507074,
      "tokens_seen": 2520580096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024026262626262627,
      "loss": 2.6083,
      "theoretical_loss": 3.366312984474705,
      "tokens_seen": 2520645632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024024242424242423,
      "loss": 2.5057,
      "theoretical_loss": 3.3663060146742927,
      "tokens_seen": 2520711168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024022222222222224,
      "loss": 2.5763,
      "theoretical_loss": 3.3662990451058223,
      "tokens_seen": 2520776704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002402020202020202,
      "loss": 2.6627,
      "theoretical_loss": 3.3662920757692802,
      "tokens_seen": 2520842240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002401818181818182,
      "loss": 2.611,
      "theoretical_loss": 3.3662851066646535,
      "tokens_seen": 2520907776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024016161616161616,
      "loss": 2.6441,
      "theoretical_loss": 3.3662781377919275,
      "tokens_seen": 2520973312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024014141414141415,
      "loss": 2.5646,
      "theoretical_loss": 3.366271169151089,
      "tokens_seen": 2521038848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024012121212121213,
      "loss": 2.5792,
      "theoretical_loss": 3.3662642007421235,
      "tokens_seen": 2521104384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024010101010101009,
      "loss": 2.5278,
      "theoretical_loss": 3.366257232565018,
      "tokens_seen": 2521169920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002400808080808081,
      "loss": 2.6671,
      "theoretical_loss": 3.366250264619759,
      "tokens_seen": 2521235456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024006060606060605,
      "loss": 2.6492,
      "theoretical_loss": 3.366243296906332,
      "tokens_seen": 2521300992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024004040404040407,
      "loss": 2.681,
      "theoretical_loss": 3.3662363294247237,
      "tokens_seen": 2521366528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024002020202020202,
      "loss": 2.4801,
      "theoretical_loss": 3.36622936217492,
      "tokens_seen": 2521432064
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024,
      "loss": 2.83,
      "theoretical_loss": 3.3662223951569077,
      "tokens_seen": 2521497600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000239979797979798,
      "loss": 2.3534,
      "theoretical_loss": 3.3662154283706727,
      "tokens_seen": 2521563136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00023995959595959597,
      "loss": 2.3948,
      "theoretical_loss": 3.3662084618162016,
      "tokens_seen": 2521628672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00023993939393939396,
      "loss": 2.6524,
      "theoretical_loss": 3.3662014954934802,
      "tokens_seen": 2521694208
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 1382657,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4263954162597656,
      "objective/train/theoretical_loss": 3.366194529402495,
      "objective/train/tokens_used": 880618976,
      "theoretical_loss": 3.366194529402495,
      "tokens_seen": 2521759744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023991919191919191,
      "loss": 2.647,
      "theoretical_loss": 3.366194529402495,
      "tokens_seen": 2521759744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002398989898989899,
      "loss": 2.604,
      "theoretical_loss": 3.366187563543232,
      "tokens_seen": 2521825280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023987878787878788,
      "loss": 2.9063,
      "theoretical_loss": 3.366180597915678,
      "tokens_seen": 2521890816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023985858585858587,
      "loss": 2.4861,
      "theoretical_loss": 3.3661736325198186,
      "tokens_seen": 2521956352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023983838383838385,
      "loss": 2.7311,
      "theoretical_loss": 3.366166667355641,
      "tokens_seen": 2522021888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023981818181818183,
      "loss": 2.6749,
      "theoretical_loss": 3.366159702423131,
      "tokens_seen": 2522087424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002397979797979798,
      "loss": 2.8589,
      "theoretical_loss": 3.366152737722274,
      "tokens_seen": 2522152960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002397777777777778,
      "loss": 2.6195,
      "theoretical_loss": 3.366145773253058,
      "tokens_seen": 2522218496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023975757575757576,
      "loss": 2.7144,
      "theoretical_loss": 3.366138809015468,
      "tokens_seen": 2522284032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023973737373737374,
      "loss": 2.6001,
      "theoretical_loss": 3.3661318450094906,
      "tokens_seen": 2522349568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023971717171717173,
      "loss": 2.5175,
      "theoretical_loss": 3.366124881235112,
      "tokens_seen": 2522415104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023969696969696968,
      "loss": 2.5426,
      "theoretical_loss": 3.3661179176923186,
      "tokens_seen": 2522480640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002396767676767677,
      "loss": 2.3801,
      "theoretical_loss": 3.3661109543810968,
      "tokens_seen": 2522546176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023965656565656565,
      "loss": 2.5913,
      "theoretical_loss": 3.3661039913014332,
      "tokens_seen": 2522611712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023963636363636366,
      "loss": 2.4643,
      "theoretical_loss": 3.3660970284533134,
      "tokens_seen": 2522677248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023961616161616162,
      "loss": 2.512,
      "theoretical_loss": 3.3660900658367234,
      "tokens_seen": 2522742784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023959595959595958,
      "loss": 2.5461,
      "theoretical_loss": 3.3660831034516505,
      "tokens_seen": 2522808320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002395757575757576,
      "loss": 2.6507,
      "theoretical_loss": 3.3660761412980804,
      "tokens_seen": 2522873856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023955555555555554,
      "loss": 2.7701,
      "theoretical_loss": 3.366069179376,
      "tokens_seen": 2522939392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023953535353535356,
      "loss": 2.6033,
      "theoretical_loss": 3.3660622176853945,
      "tokens_seen": 2523004928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002395151515151515,
      "loss": 2.5006,
      "theoretical_loss": 3.366055256226251,
      "tokens_seen": 2523070464
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394949494949495,
      "loss": 2.5732,
      "theoretical_loss": 3.3660482949985555,
      "tokens_seen": 2523136000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023947474747474748,
      "loss": 2.6068,
      "theoretical_loss": 3.3660413340022943,
      "tokens_seen": 2523201536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023945454545454546,
      "loss": 2.6709,
      "theoretical_loss": 3.366034373237454,
      "tokens_seen": 2523267072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023943434343434345,
      "loss": 2.6736,
      "theoretical_loss": 3.36602741270402,
      "tokens_seen": 2523332608
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1383478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6999809741973877,
      "objective/train/theoretical_loss": 3.36602045240198,
      "objective/train/tokens_used": 882257376,
      "theoretical_loss": 3.36602045240198,
      "tokens_seen": 2523398144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394141414141414,
      "loss": 2.7498,
      "theoretical_loss": 3.36602045240198,
      "tokens_seen": 2523398144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002393939393939394,
      "loss": 2.786,
      "theoretical_loss": 3.3660134923313194,
      "tokens_seen": 2523463680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023937373737373737,
      "loss": 2.6101,
      "theoretical_loss": 3.3660065324920247,
      "tokens_seen": 2523529216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023935353535353536,
      "loss": 2.5801,
      "theoretical_loss": 3.365999572884082,
      "tokens_seen": 2523594752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023933333333333334,
      "loss": 2.5639,
      "theoretical_loss": 3.365992613507478,
      "tokens_seen": 2523660288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023931313131313133,
      "loss": 2.5456,
      "theoretical_loss": 3.3659856543621984,
      "tokens_seen": 2523725824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023929292929292928,
      "loss": 2.5165,
      "theoretical_loss": 3.3659786954482303,
      "tokens_seen": 2523791360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002392727272727273,
      "loss": 2.672,
      "theoretical_loss": 3.3659717367655593,
      "tokens_seen": 2523856896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023925252525252525,
      "loss": 2.703,
      "theoretical_loss": 3.3659647783141717,
      "tokens_seen": 2523922432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023923232323232323,
      "loss": 2.7265,
      "theoretical_loss": 3.3659578200940548,
      "tokens_seen": 2523987968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023921212121212122,
      "loss": 2.7405,
      "theoretical_loss": 3.3659508621051937,
      "tokens_seen": 2524053504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002391919191919192,
      "loss": 2.7003,
      "theoretical_loss": 3.365943904347575,
      "tokens_seen": 2524119040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023917171717171719,
      "loss": 2.7965,
      "theoretical_loss": 3.3659369468211855,
      "tokens_seen": 2524184576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023915151515151514,
      "loss": 2.6441,
      "theoretical_loss": 3.3659299895260113,
      "tokens_seen": 2524250112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023913131313131315,
      "loss": 2.5258,
      "theoretical_loss": 3.365923032462039,
      "tokens_seen": 2524315648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002391111111111111,
      "loss": 2.6987,
      "theoretical_loss": 3.365916075629254,
      "tokens_seen": 2524381184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023909090909090912,
      "loss": 2.6635,
      "theoretical_loss": 3.365909119027643,
      "tokens_seen": 2524446720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023907070707070708,
      "loss": 2.6035,
      "theoretical_loss": 3.3659021626571928,
      "tokens_seen": 2524512256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023905050505050504,
      "loss": 2.4848,
      "theoretical_loss": 3.36589520651789,
      "tokens_seen": 2524577792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023903030303030305,
      "loss": 2.4361,
      "theoretical_loss": 3.3658882506097196,
      "tokens_seen": 2524643328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239010101010101,
      "loss": 2.7254,
      "theoretical_loss": 3.3658812949326684,
      "tokens_seen": 2524708864
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023898989898989901,
      "loss": 2.662,
      "theoretical_loss": 3.365874339486724,
      "tokens_seen": 2524774400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023896969696969697,
      "loss": 2.7091,
      "theoretical_loss": 3.365867384271871,
      "tokens_seen": 2524839936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023894949494949496,
      "loss": 2.4473,
      "theoretical_loss": 3.3658604292880963,
      "tokens_seen": 2524905472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023892929292929294,
      "loss": 2.6125,
      "theoretical_loss": 3.3658534745353865,
      "tokens_seen": 2524971008
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1384021,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5116851329803467,
      "objective/train/theoretical_loss": 3.3658465200137275,
      "objective/train/tokens_used": 883895776,
      "theoretical_loss": 3.3658465200137275,
      "tokens_seen": 2525036544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002389090909090909,
      "loss": 2.7281,
      "theoretical_loss": 3.3658465200137275,
      "tokens_seen": 2525036544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002388888888888889,
      "loss": 2.4563,
      "theoretical_loss": 3.3658395657231064,
      "tokens_seen": 2525102080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023886868686868686,
      "loss": 2.8797,
      "theoretical_loss": 3.365832611663509,
      "tokens_seen": 2525167616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023884848484848485,
      "loss": 2.6666,
      "theoretical_loss": 3.3658256578349213,
      "tokens_seen": 2525233152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023882828282828283,
      "loss": 2.603,
      "theoretical_loss": 3.36581870423733,
      "tokens_seen": 2525298688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023880808080808082,
      "loss": 2.7394,
      "theoretical_loss": 3.365811750870722,
      "tokens_seen": 2525364224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002387878787878788,
      "loss": 2.6197,
      "theoretical_loss": 3.3658047977350822,
      "tokens_seen": 2525429760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023876767676767678,
      "loss": 2.6502,
      "theoretical_loss": 3.3657978448303982,
      "tokens_seen": 2525495296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023874747474747474,
      "loss": 2.6006,
      "theoretical_loss": 3.365790892156656,
      "tokens_seen": 2525560832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023872727272727272,
      "loss": 2.5446,
      "theoretical_loss": 3.365783939713842,
      "tokens_seen": 2525626368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002387070707070707,
      "loss": 2.5886,
      "theoretical_loss": 3.365776987501942,
      "tokens_seen": 2525691904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386868686868687,
      "loss": 2.6032,
      "theoretical_loss": 3.365770035520943,
      "tokens_seen": 2525757440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023866666666666668,
      "loss": 2.5466,
      "theoretical_loss": 3.365763083770831,
      "tokens_seen": 2525822976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023864646464646463,
      "loss": 2.7014,
      "theoretical_loss": 3.3657561322515925,
      "tokens_seen": 2525888512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023862626262626264,
      "loss": 2.7514,
      "theoretical_loss": 3.3657491809632134,
      "tokens_seen": 2525954048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386060606060606,
      "loss": 2.5629,
      "theoretical_loss": 3.365742229905681,
      "tokens_seen": 2526019584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002385858585858586,
      "loss": 2.7119,
      "theoretical_loss": 3.3657352790789807,
      "tokens_seen": 2526085120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023856565656565657,
      "loss": 2.6874,
      "theoretical_loss": 3.365728328483099,
      "tokens_seen": 2526150656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023854545454545453,
      "loss": 2.5777,
      "theoretical_loss": 3.365721378118023,
      "tokens_seen": 2526216192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023852525252525254,
      "loss": 2.6121,
      "theoretical_loss": 3.365714427983738,
      "tokens_seen": 2526281728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002385050505050505,
      "loss": 2.5176,
      "theoretical_loss": 3.365707478080231,
      "tokens_seen": 2526347264
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002384848484848485,
      "loss": 2.6428,
      "theoretical_loss": 3.3657005284074883,
      "tokens_seen": 2526412800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023846464646464646,
      "loss": 2.6427,
      "theoretical_loss": 3.3656935789654963,
      "tokens_seen": 2526478336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023844444444444445,
      "loss": 2.5281,
      "theoretical_loss": 3.3656866297542405,
      "tokens_seen": 2526543872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023842424242424243,
      "loss": 2.5407,
      "theoretical_loss": 3.3656796807737086,
      "tokens_seen": 2526609408
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1385247,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5841684341430664,
      "objective/train/theoretical_loss": 3.365672732023886,
      "objective/train/tokens_used": 885534176,
      "theoretical_loss": 3.365672732023886,
      "tokens_seen": 2526674944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023840404040404041,
      "loss": 2.5472,
      "theoretical_loss": 3.365672732023886,
      "tokens_seen": 2526674944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002383838383838384,
      "loss": 2.4669,
      "theoretical_loss": 3.3656657835047596,
      "tokens_seen": 2526740480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023836363636363635,
      "loss": 2.6163,
      "theoretical_loss": 3.3656588352163155,
      "tokens_seen": 2526806016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023834343434343437,
      "loss": 2.5835,
      "theoretical_loss": 3.3656518871585397,
      "tokens_seen": 2526871552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023832323232323232,
      "loss": 2.4678,
      "theoretical_loss": 3.3656449393314194,
      "tokens_seen": 2526937088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002383030303030303,
      "loss": 2.7687,
      "theoretical_loss": 3.36563799173494,
      "tokens_seen": 2527002624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382828282828283,
      "loss": 2.7377,
      "theoretical_loss": 3.365631044369089,
      "tokens_seen": 2527068160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023826262626262627,
      "loss": 2.4071,
      "theoretical_loss": 3.3656240972338516,
      "tokens_seen": 2527133696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023824242424242426,
      "loss": 2.517,
      "theoretical_loss": 3.3656171503292147,
      "tokens_seen": 2527199232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023822222222222224,
      "loss": 2.6606,
      "theoretical_loss": 3.365610203655165,
      "tokens_seen": 2527264768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382020202020202,
      "loss": 2.5508,
      "theoretical_loss": 3.3656032572116885,
      "tokens_seen": 2527330304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023818181818181818,
      "loss": 2.5295,
      "theoretical_loss": 3.365596310998771,
      "tokens_seen": 2527395840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023816161616161617,
      "loss": 2.6017,
      "theoretical_loss": 3.3655893650164,
      "tokens_seen": 2527461376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023814141414141415,
      "loss": 2.3504,
      "theoretical_loss": 3.3655824192645616,
      "tokens_seen": 2527526912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023812121212121214,
      "loss": 2.5375,
      "theoretical_loss": 3.3655754737432413,
      "tokens_seen": 2527592448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002381010101010101,
      "loss": 2.4184,
      "theoretical_loss": 3.3655685284524264,
      "tokens_seen": 2527657984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002380808080808081,
      "loss": 2.6848,
      "theoretical_loss": 3.3655615833921027,
      "tokens_seen": 2527723520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023806060606060606,
      "loss": 2.5189,
      "theoretical_loss": 3.365554638562257,
      "tokens_seen": 2527789056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023804040404040404,
      "loss": 2.7523,
      "theoretical_loss": 3.3655476939628755,
      "tokens_seen": 2527854592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023802020202020203,
      "loss": 2.7487,
      "theoretical_loss": 3.3655407495939444,
      "tokens_seen": 2527920128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023799999999999998,
      "loss": 2.6504,
      "theoretical_loss": 3.3655338054554504,
      "tokens_seen": 2527985664
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000237979797979798,
      "loss": 2.6846,
      "theoretical_loss": 3.36552686154738,
      "tokens_seen": 2528051200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023795959595959595,
      "loss": 2.6029,
      "theoretical_loss": 3.365519917869719,
      "tokens_seen": 2528116736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023793939393939396,
      "loss": 2.4627,
      "theoretical_loss": 3.365512974422454,
      "tokens_seen": 2528182272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023791919191919192,
      "loss": 2.6922,
      "theoretical_loss": 3.365506031205572,
      "tokens_seen": 2528247808
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1385945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7589683532714844,
      "objective/train/theoretical_loss": 3.3654990882190585,
      "objective/train/tokens_used": 887172576,
      "theoretical_loss": 3.3654990882190585,
      "tokens_seen": 2528313344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378989898989899,
      "loss": 2.3046,
      "theoretical_loss": 3.3654990882190585,
      "tokens_seen": 2528313344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378787878787879,
      "loss": 2.5149,
      "theoretical_loss": 3.3654921454629,
      "tokens_seen": 2528378880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023785858585858585,
      "loss": 2.5721,
      "theoretical_loss": 3.3654852029370836,
      "tokens_seen": 2528444416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023783838383838386,
      "loss": 2.8291,
      "theoretical_loss": 3.365478260641595,
      "tokens_seen": 2528509952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378181818181818,
      "loss": 2.3108,
      "theoretical_loss": 3.365471318576421,
      "tokens_seen": 2528575488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002377979797979798,
      "loss": 2.6363,
      "theoretical_loss": 3.3654643767415475,
      "tokens_seen": 2528641024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023777777777777778,
      "loss": 2.5172,
      "theoretical_loss": 3.3654574351369617,
      "tokens_seen": 2528706560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023775757575757577,
      "loss": 2.4522,
      "theoretical_loss": 3.365450493762649,
      "tokens_seen": 2528772096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023773737373737375,
      "loss": 2.5228,
      "theoretical_loss": 3.365443552618596,
      "tokens_seen": 2528837632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023771717171717173,
      "loss": 2.5641,
      "theoretical_loss": 3.3654366117047902,
      "tokens_seen": 2528903168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002376969696969697,
      "loss": 2.526,
      "theoretical_loss": 3.365429671021217,
      "tokens_seen": 2528968704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023767676767676767,
      "loss": 2.3843,
      "theoretical_loss": 3.365422730567863,
      "tokens_seen": 2529034240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023765656565656566,
      "loss": 2.517,
      "theoretical_loss": 3.365415790344714,
      "tokens_seen": 2529099776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023763636363636364,
      "loss": 2.6491,
      "theoretical_loss": 3.3654088503517574,
      "tokens_seen": 2529165312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023761616161616163,
      "loss": 2.5191,
      "theoretical_loss": 3.365401910588979,
      "tokens_seen": 2529230848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023759595959595958,
      "loss": 2.6534,
      "theoretical_loss": 3.3653949710563653,
      "tokens_seen": 2529296384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002375757575757576,
      "loss": 2.557,
      "theoretical_loss": 3.365388031753903,
      "tokens_seen": 2529361920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023755555555555555,
      "loss": 2.5502,
      "theoretical_loss": 3.365381092681578,
      "tokens_seen": 2529427456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023753535353535356,
      "loss": 2.5511,
      "theoretical_loss": 3.3653741538393778,
      "tokens_seen": 2529492992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023751515151515152,
      "loss": 2.6639,
      "theoretical_loss": 3.365367215227287,
      "tokens_seen": 2529558528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374949494949495,
      "loss": 2.7265,
      "theoretical_loss": 3.3653602768452937,
      "tokens_seen": 2529624064
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023747474747474749,
      "loss": 2.4818,
      "theoretical_loss": 3.3653533386933834,
      "tokens_seen": 2529689600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023745454545454544,
      "loss": 2.5552,
      "theoretical_loss": 3.3653464007715423,
      "tokens_seen": 2529755136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023743434343434345,
      "loss": 2.5522,
      "theoretical_loss": 3.3653394630797577,
      "tokens_seen": 2529820672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374141414141414,
      "loss": 2.6629,
      "theoretical_loss": 3.3653325256180153,
      "tokens_seen": 2529886208
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1387296,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6340339183807373,
      "objective/train/theoretical_loss": 3.365325588386302,
      "objective/train/tokens_used": 888810976,
      "theoretical_loss": 3.365325588386302,
      "tokens_seen": 2529951744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023739393939393942,
      "loss": 2.5533,
      "theoretical_loss": 3.365325588386302,
      "tokens_seen": 2529951744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023737373737373738,
      "loss": 2.4233,
      "theoretical_loss": 3.3653186513846034,
      "tokens_seen": 2530017280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023735353535353534,
      "loss": 2.5152,
      "theoretical_loss": 3.365311714612907,
      "tokens_seen": 2530082816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023733333333333335,
      "loss": 2.6877,
      "theoretical_loss": 3.3653047780711987,
      "tokens_seen": 2530148352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002373131313131313,
      "loss": 2.588,
      "theoretical_loss": 3.365297841759465,
      "tokens_seen": 2530213888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023729292929292931,
      "loss": 2.6193,
      "theoretical_loss": 3.365290905677692,
      "tokens_seen": 2530279424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023727272727272727,
      "loss": 2.5231,
      "theoretical_loss": 3.365283969825866,
      "tokens_seen": 2530344960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023725252525252526,
      "loss": 2.6264,
      "theoretical_loss": 3.365277034203974,
      "tokens_seen": 2530410496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023723232323232324,
      "loss": 2.7149,
      "theoretical_loss": 3.365270098812003,
      "tokens_seen": 2530476032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023721212121212122,
      "loss": 2.6112,
      "theoretical_loss": 3.3652631636499377,
      "tokens_seen": 2530541568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002371919191919192,
      "loss": 2.4873,
      "theoretical_loss": 3.3652562287177656,
      "tokens_seen": 2530607104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023717171717171716,
      "loss": 2.5812,
      "theoretical_loss": 3.3652492940154732,
      "tokens_seen": 2530672640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023715151515151515,
      "loss": 2.4426,
      "theoretical_loss": 3.3652423595430467,
      "tokens_seen": 2530738176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023713131313131313,
      "loss": 2.5431,
      "theoretical_loss": 3.3652354253004724,
      "tokens_seen": 2530803712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023711111111111112,
      "loss": 2.5856,
      "theoretical_loss": 3.365228491287737,
      "tokens_seen": 2530869248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002370909090909091,
      "loss": 2.5204,
      "theoretical_loss": 3.3652215575048268,
      "tokens_seen": 2530934784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023707070707070708,
      "loss": 2.4103,
      "theoretical_loss": 3.365214623951728,
      "tokens_seen": 2531000320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023705050505050504,
      "loss": 2.552,
      "theoretical_loss": 3.365207690628427,
      "tokens_seen": 2531065856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023703030303030305,
      "loss": 2.6296,
      "theoretical_loss": 3.365200757534911,
      "tokens_seen": 2531131392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000237010101010101,
      "loss": 2.6525,
      "theoretical_loss": 3.365193824671166,
      "tokens_seen": 2531196928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000236989898989899,
      "loss": 2.5798,
      "theoretical_loss": 3.365186892037178,
      "tokens_seen": 2531262464
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023696969696969698,
      "loss": 2.7886,
      "theoretical_loss": 3.3651799596329344,
      "tokens_seen": 2531328000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023694949494949493,
      "loss": 2.634,
      "theoretical_loss": 3.3651730274584204,
      "tokens_seen": 2531393536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023692929292929294,
      "loss": 2.5433,
      "theoretical_loss": 3.365166095513623,
      "tokens_seen": 2531459072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002369090909090909,
      "loss": 2.6011,
      "theoretical_loss": 3.365159163798529,
      "tokens_seen": 2531524608
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1387695,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.649080514907837,
      "objective/train/theoretical_loss": 3.3651522323131244,
      "objective/train/tokens_used": 890449376,
      "theoretical_loss": 3.3651522323131244,
      "tokens_seen": 2531590144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368888888888889,
      "loss": 2.4917,
      "theoretical_loss": 3.3651522323131244,
      "tokens_seen": 2531590144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023686868686868687,
      "loss": 2.7032,
      "theoretical_loss": 3.3651453010573964,
      "tokens_seen": 2531655680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023684848484848485,
      "loss": 2.5129,
      "theoretical_loss": 3.36513837003133,
      "tokens_seen": 2531721216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023682828282828284,
      "loss": 2.5074,
      "theoretical_loss": 3.365131439234913,
      "tokens_seen": 2531786752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368080808080808,
      "loss": 2.8184,
      "theoretical_loss": 3.3651245086681314,
      "tokens_seen": 2531852288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002367878787878788,
      "loss": 2.6483,
      "theoretical_loss": 3.3651175783309712,
      "tokens_seen": 2531917824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023676767676767676,
      "loss": 2.5097,
      "theoretical_loss": 3.3651106482234194,
      "tokens_seen": 2531983360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023674747474747475,
      "loss": 2.7178,
      "theoretical_loss": 3.3651037183454626,
      "tokens_seen": 2532048896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023672727272727273,
      "loss": 2.9447,
      "theoretical_loss": 3.3650967886970866,
      "tokens_seen": 2532114432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023670707070707071,
      "loss": 2.7031,
      "theoretical_loss": 3.365089859278278,
      "tokens_seen": 2532179968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002366868686868687,
      "loss": 2.4232,
      "theoretical_loss": 3.365082930089024,
      "tokens_seen": 2532245504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023666666666666668,
      "loss": 2.7103,
      "theoretical_loss": 3.36507600112931,
      "tokens_seen": 2532311040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023664646464646464,
      "loss": 2.4652,
      "theoretical_loss": 3.3650690723991232,
      "tokens_seen": 2532376576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023662626262626262,
      "loss": 2.4705,
      "theoretical_loss": 3.36506214389845,
      "tokens_seen": 2532442112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002366060606060606,
      "loss": 2.4316,
      "theoretical_loss": 3.3650552156272764,
      "tokens_seen": 2532507648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002365858585858586,
      "loss": 2.5798,
      "theoretical_loss": 3.3650482875855894,
      "tokens_seen": 2532573184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023656565656565657,
      "loss": 2.7893,
      "theoretical_loss": 3.365041359773375,
      "tokens_seen": 2532638720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023654545454545456,
      "loss": 2.6254,
      "theoretical_loss": 3.36503443219062,
      "tokens_seen": 2532704256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023652525252525254,
      "loss": 2.7588,
      "theoretical_loss": 3.36502750483731,
      "tokens_seen": 2532769792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002365050505050505,
      "loss": 2.5603,
      "theoretical_loss": 3.365020577713433,
      "tokens_seen": 2532835328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002364848484848485,
      "loss": 2.5845,
      "theoretical_loss": 3.365013650818974,
      "tokens_seen": 2532900864
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023646464646464647,
      "loss": 2.533,
      "theoretical_loss": 3.3650067241539205,
      "tokens_seen": 2532966400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023644444444444445,
      "loss": 2.5883,
      "theoretical_loss": 3.3649997977182586,
      "tokens_seen": 2533031936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023642424242424244,
      "loss": 2.7452,
      "theoretical_loss": 3.3649928715119746,
      "tokens_seen": 2533097472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002364040404040404,
      "loss": 2.6755,
      "theoretical_loss": 3.3649859455350555,
      "tokens_seen": 2533163008
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1388555,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5344579219818115,
      "objective/train/theoretical_loss": 3.364979019787487,
      "objective/train/tokens_used": 892087776,
      "theoretical_loss": 3.364979019787487,
      "tokens_seen": 2533228544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002363838383838384,
      "loss": 2.6651,
      "theoretical_loss": 3.364979019787487,
      "tokens_seen": 2533228544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023636363636363636,
      "loss": 2.5549,
      "theoretical_loss": 3.3649720942692563,
      "tokens_seen": 2533294080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023634343434343437,
      "loss": 2.4487,
      "theoretical_loss": 3.364965168980349,
      "tokens_seen": 2533359616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023632323232323233,
      "loss": 2.6822,
      "theoretical_loss": 3.3649582439207526,
      "tokens_seen": 2533425152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023630303030303028,
      "loss": 2.7001,
      "theoretical_loss": 3.3649513190904528,
      "tokens_seen": 2533490688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362828282828283,
      "loss": 2.6022,
      "theoretical_loss": 3.364944394489436,
      "tokens_seen": 2533556224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023626262626262625,
      "loss": 2.5296,
      "theoretical_loss": 3.36493747011769,
      "tokens_seen": 2533621760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023624242424242426,
      "loss": 2.5767,
      "theoretical_loss": 3.3649305459751995,
      "tokens_seen": 2533687296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023622222222222222,
      "loss": 2.7167,
      "theoretical_loss": 3.3649236220619523,
      "tokens_seen": 2533752832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362020202020202,
      "loss": 2.5037,
      "theoretical_loss": 3.364916698377934,
      "tokens_seen": 2533818368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361818181818182,
      "loss": 2.6725,
      "theoretical_loss": 3.364909774923132,
      "tokens_seen": 2533883904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023616161616161617,
      "loss": 2.4854,
      "theoretical_loss": 3.3649028516975314,
      "tokens_seen": 2533949440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023614141414141416,
      "loss": 2.5974,
      "theoretical_loss": 3.36489592870112,
      "tokens_seen": 2534014976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361212121212121,
      "loss": 2.7305,
      "theoretical_loss": 3.364889005933884,
      "tokens_seen": 2534080512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361010101010101,
      "loss": 2.5157,
      "theoretical_loss": 3.364882083395809,
      "tokens_seen": 2534146048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023608080808080808,
      "loss": 2.4699,
      "theoretical_loss": 3.364875161086883,
      "tokens_seen": 2534211584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023606060606060607,
      "loss": 2.561,
      "theoretical_loss": 3.364868239007091,
      "tokens_seen": 2534277120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023604040404040405,
      "loss": 2.7101,
      "theoretical_loss": 3.3648613171564206,
      "tokens_seen": 2534342656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023602020202020203,
      "loss": 2.3929,
      "theoretical_loss": 3.364854395534858,
      "tokens_seen": 2534408192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000236,
      "loss": 2.8132,
      "theoretical_loss": 3.3648474741423895,
      "tokens_seen": 2534473728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000235979797979798,
      "loss": 2.3725,
      "theoretical_loss": 3.3648405529790013,
      "tokens_seen": 2534539264
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023595959595959596,
      "loss": 2.4755,
      "theoretical_loss": 3.36483363204468,
      "tokens_seen": 2534604800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023593939393939394,
      "loss": 2.5545,
      "theoretical_loss": 3.364826711339413,
      "tokens_seen": 2534670336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023591919191919193,
      "loss": 2.3091,
      "theoretical_loss": 3.364819790863186,
      "tokens_seen": 2534735872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023589898989898988,
      "loss": 2.6303,
      "theoretical_loss": 3.364812870615985,
      "tokens_seen": 2534801408
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1389642,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.644278049468994,
      "objective/train/theoretical_loss": 3.364805950597798,
      "objective/train/tokens_used": 893726176,
      "theoretical_loss": 3.364805950597798,
      "tokens_seen": 2534866944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002358787878787879,
      "loss": 2.6679,
      "theoretical_loss": 3.364805950597798,
      "tokens_seen": 2534866944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023585858585858585,
      "loss": 2.5396,
      "theoretical_loss": 3.3647990308086104,
      "tokens_seen": 2534932480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023583838383838386,
      "loss": 2.66,
      "theoretical_loss": 3.3647921112484087,
      "tokens_seen": 2534998016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023581818181818182,
      "loss": 2.5682,
      "theoretical_loss": 3.36478519191718,
      "tokens_seen": 2535063552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357979797979798,
      "loss": 2.6106,
      "theoretical_loss": 3.36477827281491,
      "tokens_seen": 2535129088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357777777777778,
      "loss": 2.5681,
      "theoretical_loss": 3.3647713539415856,
      "tokens_seen": 2535194624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023575757575757574,
      "loss": 2.5381,
      "theoretical_loss": 3.364764435297194,
      "tokens_seen": 2535260160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023573737373737375,
      "loss": 2.5639,
      "theoretical_loss": 3.3647575168817205,
      "tokens_seen": 2535325696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357171717171717,
      "loss": 2.4618,
      "theoretical_loss": 3.364750598695152,
      "tokens_seen": 2535391232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023569696969696972,
      "loss": 2.6874,
      "theoretical_loss": 3.3647436807374755,
      "tokens_seen": 2535456768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023567676767676768,
      "loss": 2.4307,
      "theoretical_loss": 3.3647367630086773,
      "tokens_seen": 2535522304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023565656565656566,
      "loss": 2.8406,
      "theoretical_loss": 3.3647298455087435,
      "tokens_seen": 2535587840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023563636363636365,
      "loss": 2.7256,
      "theoretical_loss": 3.364722928237661,
      "tokens_seen": 2535653376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002356161616161616,
      "loss": 2.6546,
      "theoretical_loss": 3.3647160111954166,
      "tokens_seen": 2535718912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023559595959595962,
      "loss": 2.6839,
      "theoretical_loss": 3.364709094381996,
      "tokens_seen": 2535784448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023557575757575757,
      "loss": 2.7654,
      "theoretical_loss": 3.364702177797386,
      "tokens_seen": 2535849984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023555555555555556,
      "loss": 2.534,
      "theoretical_loss": 3.364695261441574,
      "tokens_seen": 2535915520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023553535353535354,
      "loss": 2.4804,
      "theoretical_loss": 3.364688345314545,
      "tokens_seen": 2535981056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023551515151515152,
      "loss": 2.7425,
      "theoretical_loss": 3.364681429416287,
      "tokens_seen": 2536046592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354949494949495,
      "loss": 2.6467,
      "theoretical_loss": 3.3646745137467855,
      "tokens_seen": 2536112128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354747474747475,
      "loss": 2.546,
      "theoretical_loss": 3.3646675983060272,
      "tokens_seen": 2536177664
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023545454545454545,
      "loss": 2.7171,
      "theoretical_loss": 3.364660683093999,
      "tokens_seen": 2536243200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023543434343434343,
      "loss": 2.55,
      "theoretical_loss": 3.364653768110687,
      "tokens_seen": 2536308736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023541414141414142,
      "loss": 2.4196,
      "theoretical_loss": 3.3646468533560787,
      "tokens_seen": 2536374272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002353939393939394,
      "loss": 2.7091,
      "theoretical_loss": 3.364639938830159,
      "tokens_seen": 2536439808
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1390171,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7468748092651367,
      "objective/train/theoretical_loss": 3.3646330245329157,
      "objective/train/tokens_used": 895364576,
      "theoretical_loss": 3.3646330245329157,
      "tokens_seen": 2536505344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023537373737373738,
      "loss": 2.6023,
      "theoretical_loss": 3.3646330245329157,
      "tokens_seen": 2536505344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023535353535353534,
      "loss": 2.8589,
      "theoretical_loss": 3.364626110464335,
      "tokens_seen": 2536570880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023533333333333335,
      "loss": 2.5933,
      "theoretical_loss": 3.364619196624403,
      "tokens_seen": 2536636416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002353131313131313,
      "loss": 2.6814,
      "theoretical_loss": 3.3646122830131064,
      "tokens_seen": 2536701952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023529292929292932,
      "loss": 2.6057,
      "theoretical_loss": 3.3646053696304326,
      "tokens_seen": 2536767488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023527272727272728,
      "loss": 2.6537,
      "theoretical_loss": 3.364598456476367,
      "tokens_seen": 2536833024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023525252525252523,
      "loss": 2.5661,
      "theoretical_loss": 3.3645915435508966,
      "tokens_seen": 2536898560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023523232323232325,
      "loss": 2.6555,
      "theoretical_loss": 3.364584630854008,
      "tokens_seen": 2536964096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002352121212121212,
      "loss": 2.6126,
      "theoretical_loss": 3.3645777183856875,
      "tokens_seen": 2537029632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002351919191919192,
      "loss": 2.7495,
      "theoretical_loss": 3.364570806145922,
      "tokens_seen": 2537095168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023517171717171717,
      "loss": 2.2758,
      "theoretical_loss": 3.364563894134698,
      "tokens_seen": 2537160704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023515151515151515,
      "loss": 2.4934,
      "theoretical_loss": 3.3645569823520014,
      "tokens_seen": 2537226240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023513131313131314,
      "loss": 2.5878,
      "theoretical_loss": 3.3645500707978195,
      "tokens_seen": 2537291776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023511111111111112,
      "loss": 2.6998,
      "theoretical_loss": 3.3645431594721384,
      "tokens_seen": 2537357312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002350909090909091,
      "loss": 2.7549,
      "theoretical_loss": 3.364536248374945,
      "tokens_seen": 2537422848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023507070707070706,
      "loss": 2.6067,
      "theoretical_loss": 3.3645293375062253,
      "tokens_seen": 2537488384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023505050505050505,
      "loss": 2.4318,
      "theoretical_loss": 3.3645224268659666,
      "tokens_seen": 2537553920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023503030303030303,
      "loss": 2.494,
      "theoretical_loss": 3.364515516454155,
      "tokens_seen": 2537619456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023501010101010101,
      "loss": 2.6901,
      "theoretical_loss": 3.364508606270777,
      "tokens_seen": 2537684992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000234989898989899,
      "loss": 2.904,
      "theoretical_loss": 3.364501696315819,
      "tokens_seen": 2537750528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023496969696969698,
      "loss": 2.6135,
      "theoretical_loss": 3.364494786589268,
      "tokens_seen": 2537816064
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023494949494949494,
      "loss": 2.5393,
      "theoretical_loss": 3.3644878770911104,
      "tokens_seen": 2537881600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023492929292929295,
      "loss": 2.6888,
      "theoretical_loss": 3.3644809678213328,
      "tokens_seen": 2537947136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002349090909090909,
      "loss": 2.6243,
      "theoretical_loss": 3.364474058779922,
      "tokens_seen": 2538012672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002348888888888889,
      "loss": 2.6966,
      "theoretical_loss": 3.3644671499668632,
      "tokens_seen": 2538078208
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 1391171,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3983640670776367,
      "objective/train/theoretical_loss": 3.3644602413821447,
      "objective/train/tokens_used": 897002976,
      "theoretical_loss": 3.3644602413821447,
      "tokens_seen": 2538143744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023486868686868688,
      "loss": 2.3945,
      "theoretical_loss": 3.3644602413821447,
      "tokens_seen": 2538143744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023484848484848486,
      "loss": 2.8113,
      "theoretical_loss": 3.364453333025752,
      "tokens_seen": 2538209280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023482828282828284,
      "loss": 2.576,
      "theoretical_loss": 3.3644464248976718,
      "tokens_seen": 2538274816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002348080808080808,
      "loss": 2.6328,
      "theoretical_loss": 3.364439516997891,
      "tokens_seen": 2538340352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002347878787878788,
      "loss": 2.6638,
      "theoretical_loss": 3.3644326093263963,
      "tokens_seen": 2538405888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023476767676767677,
      "loss": 2.5303,
      "theoretical_loss": 3.3644257018831736,
      "tokens_seen": 2538471424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023474747474747475,
      "loss": 2.5935,
      "theoretical_loss": 3.36441879466821,
      "tokens_seen": 2538536960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023472727272727274,
      "loss": 2.8733,
      "theoretical_loss": 3.3644118876814924,
      "tokens_seen": 2538602496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002347070707070707,
      "loss": 2.5653,
      "theoretical_loss": 3.364404980923006,
      "tokens_seen": 2538668032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002346868686868687,
      "loss": 2.5322,
      "theoretical_loss": 3.3643980743927386,
      "tokens_seen": 2538733568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023466666666666666,
      "loss": 2.6673,
      "theoretical_loss": 3.3643911680906764,
      "tokens_seen": 2538799104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023464646464646467,
      "loss": 2.4581,
      "theoretical_loss": 3.364384262016806,
      "tokens_seen": 2538864640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023462626262626263,
      "loss": 2.4531,
      "theoretical_loss": 3.3643773561711137,
      "tokens_seen": 2538930176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002346060606060606,
      "loss": 2.5356,
      "theoretical_loss": 3.3643704505535865,
      "tokens_seen": 2538995712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345858585858586,
      "loss": 2.3763,
      "theoretical_loss": 3.3643635451642107,
      "tokens_seen": 2539061248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023456565656565655,
      "loss": 2.5074,
      "theoretical_loss": 3.364356640002973,
      "tokens_seen": 2539126784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023454545454545456,
      "loss": 2.5965,
      "theoretical_loss": 3.36434973506986,
      "tokens_seen": 2539192320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023452525252525252,
      "loss": 2.608,
      "theoretical_loss": 3.364342830364858,
      "tokens_seen": 2539257856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345050505050505,
      "loss": 2.6128,
      "theoretical_loss": 3.364335925887954,
      "tokens_seen": 2539323392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002344848484848485,
      "loss": 2.8063,
      "theoretical_loss": 3.364329021639134,
      "tokens_seen": 2539388928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023446464646464647,
      "loss": 2.6855,
      "theoretical_loss": 3.3643221176183853,
      "tokens_seen": 2539454464
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023444444444444446,
      "loss": 2.5181,
      "theoretical_loss": 3.364315213825694,
      "tokens_seen": 2539520000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023442424242424244,
      "loss": 2.7349,
      "theoretical_loss": 3.364308310261047,
      "tokens_seen": 2539585536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002344040404040404,
      "loss": 2.6431,
      "theoretical_loss": 3.3643014069244304,
      "tokens_seen": 2539651072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023438383838383838,
      "loss": 2.5689,
      "theoretical_loss": 3.364294503815831,
      "tokens_seen": 2539716608
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1391800,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2752268314361572,
      "objective/train/theoretical_loss": 3.3642876009352354,
      "objective/train/tokens_used": 898641376,
      "theoretical_loss": 3.3642876009352354,
      "tokens_seen": 2539782144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023436363636363637,
      "loss": 2.566,
      "theoretical_loss": 3.3642876009352354,
      "tokens_seen": 2539782144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023434343434343435,
      "loss": 2.4836,
      "theoretical_loss": 3.3642806982826308,
      "tokens_seen": 2539847680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023432323232323233,
      "loss": 2.7731,
      "theoretical_loss": 3.3642737958580025,
      "tokens_seen": 2539913216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002343030303030303,
      "loss": 2.4425,
      "theoretical_loss": 3.364266893661338,
      "tokens_seen": 2539978752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342828282828283,
      "loss": 2.6432,
      "theoretical_loss": 3.3642599916926237,
      "tokens_seen": 2540044288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023426262626262626,
      "loss": 2.5975,
      "theoretical_loss": 3.3642530899518466,
      "tokens_seen": 2540109824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023424242424242427,
      "loss": 2.5968,
      "theoretical_loss": 3.3642461884389925,
      "tokens_seen": 2540175360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023422222222222223,
      "loss": 2.5379,
      "theoretical_loss": 3.3642392871540485,
      "tokens_seen": 2540240896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023420202020202018,
      "loss": 2.574,
      "theoretical_loss": 3.364232386097001,
      "tokens_seen": 2540306432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341818181818182,
      "loss": 2.6247,
      "theoretical_loss": 3.364225485267837,
      "tokens_seen": 2540371968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023416161616161615,
      "loss": 2.6482,
      "theoretical_loss": 3.364218584666542,
      "tokens_seen": 2540437504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023414141414141416,
      "loss": 2.7569,
      "theoretical_loss": 3.364211684293104,
      "tokens_seen": 2540503040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023412121212121212,
      "loss": 2.6188,
      "theoretical_loss": 3.3642047841475087,
      "tokens_seen": 2540568576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341010101010101,
      "loss": 2.5926,
      "theoretical_loss": 3.364197884229743,
      "tokens_seen": 2540634112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002340808080808081,
      "loss": 2.6181,
      "theoretical_loss": 3.3641909845397935,
      "tokens_seen": 2540699648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023406060606060604,
      "loss": 2.5699,
      "theoretical_loss": 3.364184085077647,
      "tokens_seen": 2540765184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023404040404040405,
      "loss": 2.471,
      "theoretical_loss": 3.3641771858432894,
      "tokens_seen": 2540830720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000234020202020202,
      "loss": 2.5748,
      "theoretical_loss": 3.364170286836708,
      "tokens_seen": 2540896256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023400000000000002,
      "loss": 2.5245,
      "theoretical_loss": 3.3641633880578894,
      "tokens_seen": 2540961792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023397979797979798,
      "loss": 2.7463,
      "theoretical_loss": 3.3641564895068194,
      "tokens_seen": 2541027328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023395959595959596,
      "loss": 2.4495,
      "theoretical_loss": 3.3641495911834856,
      "tokens_seen": 2541092864
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023393939393939395,
      "loss": 2.62,
      "theoretical_loss": 3.3641426930878744,
      "tokens_seen": 2541158400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023391919191919193,
      "loss": 2.5062,
      "theoretical_loss": 3.3641357952199717,
      "tokens_seen": 2541223936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023389898989898992,
      "loss": 2.3945,
      "theoretical_loss": 3.3641288975797647,
      "tokens_seen": 2541289472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023387878787878787,
      "loss": 2.5318,
      "theoretical_loss": 3.36412200016724,
      "tokens_seen": 2541355008
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1392945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5443549156188965,
      "objective/train/theoretical_loss": 3.3641151029823844,
      "objective/train/tokens_used": 900279776,
      "theoretical_loss": 3.3641151029823844,
      "tokens_seen": 2541420544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023385858585858586,
      "loss": 2.7348,
      "theoretical_loss": 3.3641151029823844,
      "tokens_seen": 2541420544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023383838383838384,
      "loss": 2.5947,
      "theoretical_loss": 3.364108206025184,
      "tokens_seen": 2541486080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023381818181818182,
      "loss": 2.533,
      "theoretical_loss": 3.3641013092956253,
      "tokens_seen": 2541551616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337979797979798,
      "loss": 2.5768,
      "theoretical_loss": 3.364094412793696,
      "tokens_seen": 2541617152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337777777777778,
      "loss": 2.6141,
      "theoretical_loss": 3.3640875165193815,
      "tokens_seen": 2541682688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023375757575757575,
      "loss": 2.5833,
      "theoretical_loss": 3.364080620472669,
      "tokens_seen": 2541748224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023373737373737376,
      "loss": 2.7621,
      "theoretical_loss": 3.3640737246535455,
      "tokens_seen": 2541813760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023371717171717172,
      "loss": 2.7913,
      "theoretical_loss": 3.364066829061997,
      "tokens_seen": 2541879296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002336969696969697,
      "loss": 2.5465,
      "theoretical_loss": 3.36405993369801,
      "tokens_seen": 2541944832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023367676767676768,
      "loss": 2.4993,
      "theoretical_loss": 3.3640530385615715,
      "tokens_seen": 2542010368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023365656565656564,
      "loss": 2.4185,
      "theoretical_loss": 3.364046143652668,
      "tokens_seen": 2542075904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023363636363636365,
      "loss": 2.6962,
      "theoretical_loss": 3.3640392489712863,
      "tokens_seen": 2542141440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002336161616161616,
      "loss": 2.6121,
      "theoretical_loss": 3.3640323545174127,
      "tokens_seen": 2542206976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023359595959595962,
      "loss": 2.6012,
      "theoretical_loss": 3.3640254602910344,
      "tokens_seen": 2542272512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023357575757575758,
      "loss": 2.5755,
      "theoretical_loss": 3.3640185662921374,
      "tokens_seen": 2542338048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023355555555555556,
      "loss": 2.4853,
      "theoretical_loss": 3.3640116725207085,
      "tokens_seen": 2542403584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023353535353535355,
      "loss": 2.3537,
      "theoretical_loss": 3.3640047789767342,
      "tokens_seen": 2542469120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002335151515151515,
      "loss": 2.7106,
      "theoretical_loss": 3.363997885660202,
      "tokens_seen": 2542534656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002334949494949495,
      "loss": 2.5192,
      "theoretical_loss": 3.3639909925710976,
      "tokens_seen": 2542600192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023347474747474747,
      "loss": 2.6282,
      "theoretical_loss": 3.3639840997094077,
      "tokens_seen": 2542665728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023345454545454545,
      "loss": 2.5555,
      "theoretical_loss": 3.363977207075119,
      "tokens_seen": 2542731264
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023343434343434344,
      "loss": 2.4653,
      "theoretical_loss": 3.3639703146682187,
      "tokens_seen": 2542796800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023341414141414142,
      "loss": 2.6739,
      "theoretical_loss": 3.363963422488693,
      "tokens_seen": 2542862336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002333939393939394,
      "loss": 2.6557,
      "theoretical_loss": 3.3639565305365284,
      "tokens_seen": 2542927872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002333737373737374,
      "loss": 2.6826,
      "theoretical_loss": 3.363949638811712,
      "tokens_seen": 2542993408
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1394002,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.704057455062866,
      "objective/train/theoretical_loss": 3.36394274731423,
      "objective/train/tokens_used": 901918176,
      "theoretical_loss": 3.36394274731423,
      "tokens_seen": 2543058944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023335353535353535,
      "loss": 2.6489,
      "theoretical_loss": 3.36394274731423,
      "tokens_seen": 2543058944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023333333333333333,
      "loss": 2.4723,
      "theoretical_loss": 3.3639358560440686,
      "tokens_seen": 2543124480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023331313131313131,
      "loss": 2.9125,
      "theoretical_loss": 3.3639289650012154,
      "tokens_seen": 2543190016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002332929292929293,
      "loss": 2.5235,
      "theoretical_loss": 3.363922074185657,
      "tokens_seen": 2543255552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023327272727272728,
      "loss": 2.5792,
      "theoretical_loss": 3.3639151835973795,
      "tokens_seen": 2543321088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023325252525252524,
      "loss": 2.5602,
      "theoretical_loss": 3.3639082932363698,
      "tokens_seen": 2543386624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023323232323232325,
      "loss": 2.4142,
      "theoretical_loss": 3.363901403102614,
      "tokens_seen": 2543452160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002332121212121212,
      "loss": 2.5644,
      "theoretical_loss": 3.3638945131961,
      "tokens_seen": 2543517696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002331919191919192,
      "loss": 2.6675,
      "theoretical_loss": 3.3638876235168134,
      "tokens_seen": 2543583232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023317171717171718,
      "loss": 2.635,
      "theoretical_loss": 3.3638807340647414,
      "tokens_seen": 2543648768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023315151515151516,
      "loss": 2.4709,
      "theoretical_loss": 3.36387384483987,
      "tokens_seen": 2543714304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023313131313131314,
      "loss": 2.7813,
      "theoretical_loss": 3.3638669558421865,
      "tokens_seen": 2543779840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002331111111111111,
      "loss": 2.5628,
      "theoretical_loss": 3.3638600670716774,
      "tokens_seen": 2543845376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002330909090909091,
      "loss": 2.4428,
      "theoretical_loss": 3.363853178528329,
      "tokens_seen": 2543910912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023307070707070707,
      "loss": 2.5687,
      "theoretical_loss": 3.3638462902121287,
      "tokens_seen": 2543976448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023305050505050508,
      "loss": 2.3504,
      "theoretical_loss": 3.3638394021230624,
      "tokens_seen": 2544041984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023303030303030304,
      "loss": 2.5489,
      "theoretical_loss": 3.363832514261117,
      "tokens_seen": 2544107520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000233010101010101,
      "loss": 2.6251,
      "theoretical_loss": 3.3638256266262796,
      "tokens_seen": 2544173056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000232989898989899,
      "loss": 2.6629,
      "theoretical_loss": 3.3638187392185364,
      "tokens_seen": 2544238592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023296969696969696,
      "loss": 2.6492,
      "theoretical_loss": 3.363811852037874,
      "tokens_seen": 2544304128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023294949494949497,
      "loss": 2.6118,
      "theoretical_loss": 3.363804965084279,
      "tokens_seen": 2544369664
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023292929292929293,
      "loss": 2.566,
      "theoretical_loss": 3.3637980783577386,
      "tokens_seen": 2544435200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002329090909090909,
      "loss": 2.7243,
      "theoretical_loss": 3.363791191858239,
      "tokens_seen": 2544500736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002328888888888889,
      "loss": 2.6767,
      "theoretical_loss": 3.3637843055857672,
      "tokens_seen": 2544566272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023286868686868688,
      "loss": 2.538,
      "theoretical_loss": 3.3637774195403094,
      "tokens_seen": 2544631808
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1394146,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.181093454360962,
      "objective/train/theoretical_loss": 3.3637705337218526,
      "objective/train/tokens_used": 903556576,
      "theoretical_loss": 3.3637705337218526,
      "tokens_seen": 2544697344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023284848484848486,
      "loss": 2.4151,
      "theoretical_loss": 3.3637705337218526,
      "tokens_seen": 2544697344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023282828282828282,
      "loss": 2.7687,
      "theoretical_loss": 3.3637636481303836,
      "tokens_seen": 2544762880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002328080808080808,
      "loss": 2.4268,
      "theoretical_loss": 3.3637567627658886,
      "tokens_seen": 2544828416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002327878787878788,
      "loss": 2.7664,
      "theoretical_loss": 3.3637498776283548,
      "tokens_seen": 2544893952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023276767676767677,
      "loss": 2.4098,
      "theoretical_loss": 3.3637429927177687,
      "tokens_seen": 2544959488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023274747474747476,
      "loss": 2.5337,
      "theoretical_loss": 3.363736108034117,
      "tokens_seen": 2545025024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023272727272727274,
      "loss": 2.5796,
      "theoretical_loss": 3.363729223577386,
      "tokens_seen": 2545090560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002327070707070707,
      "loss": 2.6336,
      "theoretical_loss": 3.3637223393475626,
      "tokens_seen": 2545156096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002326868686868687,
      "loss": 2.6943,
      "theoretical_loss": 3.363715455344634,
      "tokens_seen": 2545221632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023266666666666667,
      "loss": 2.609,
      "theoretical_loss": 3.363708571568586,
      "tokens_seen": 2545287168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023264646464646465,
      "loss": 2.392,
      "theoretical_loss": 3.3637016880194057,
      "tokens_seen": 2545352704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023262626262626263,
      "loss": 2.6878,
      "theoretical_loss": 3.3636948046970803,
      "tokens_seen": 2545418240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002326060606060606,
      "loss": 2.5737,
      "theoretical_loss": 3.3636879216015956,
      "tokens_seen": 2545483776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002325858585858586,
      "loss": 2.7241,
      "theoretical_loss": 3.3636810387329383,
      "tokens_seen": 2545549312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023256565656565656,
      "loss": 2.5932,
      "theoretical_loss": 3.363674156091096,
      "tokens_seen": 2545614848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023254545454545457,
      "loss": 2.6117,
      "theoretical_loss": 3.3636672736760547,
      "tokens_seen": 2545680384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023252525252525253,
      "loss": 2.5194,
      "theoretical_loss": 3.363660391487801,
      "tokens_seen": 2545745920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023250505050505048,
      "loss": 2.7884,
      "theoretical_loss": 3.3636535095263222,
      "tokens_seen": 2545811456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002324848484848485,
      "loss": 2.5799,
      "theoretical_loss": 3.3636466277916046,
      "tokens_seen": 2545876992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023246464646464645,
      "loss": 2.6815,
      "theoretical_loss": 3.3636397462836345,
      "tokens_seen": 2545942528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023244444444444446,
      "loss": 2.5922,
      "theoretical_loss": 3.3636328650023994,
      "tokens_seen": 2546008064
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023242424242424242,
      "loss": 2.7509,
      "theoretical_loss": 3.3636259839478853,
      "tokens_seen": 2546073600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002324040404040404,
      "loss": 2.8669,
      "theoretical_loss": 3.363619103120079,
      "tokens_seen": 2546139136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002323838383838384,
      "loss": 2.3998,
      "theoretical_loss": 3.3636122225189675,
      "tokens_seen": 2546204672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023236363636363637,
      "loss": 2.4883,
      "theoretical_loss": 3.3636053421445378,
      "tokens_seen": 2546270208
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1395112,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8437981605529785,
      "objective/train/theoretical_loss": 3.3635984619967756,
      "objective/train/tokens_used": 905194976,
      "theoretical_loss": 3.3635984619967756,
      "tokens_seen": 2546335744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023234343434343436,
      "loss": 2.6607,
      "theoretical_loss": 3.3635984619967756,
      "tokens_seen": 2546335744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002323232323232323,
      "loss": 2.4318,
      "theoretical_loss": 3.3635915820756686,
      "tokens_seen": 2546401280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023230303030303032,
      "loss": 2.5522,
      "theoretical_loss": 3.363584702381203,
      "tokens_seen": 2546466816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023228282828282828,
      "loss": 2.8892,
      "theoretical_loss": 3.3635778229133653,
      "tokens_seen": 2546532352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023226262626262626,
      "loss": 2.6612,
      "theoretical_loss": 3.3635709436721424,
      "tokens_seen": 2546597888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023224242424242425,
      "loss": 2.6752,
      "theoretical_loss": 3.3635640646575213,
      "tokens_seen": 2546663424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023222222222222223,
      "loss": 2.61,
      "theoretical_loss": 3.3635571858694884,
      "tokens_seen": 2546728960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023220202020202022,
      "loss": 2.7056,
      "theoretical_loss": 3.3635503073080306,
      "tokens_seen": 2546794496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002321818181818182,
      "loss": 2.4905,
      "theoretical_loss": 3.3635434289731343,
      "tokens_seen": 2546860032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023216161616161616,
      "loss": 2.4825,
      "theoretical_loss": 3.3635365508647865,
      "tokens_seen": 2546925568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023214141414141414,
      "loss": 2.6878,
      "theoretical_loss": 3.363529672982974,
      "tokens_seen": 2546991104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023212121212121212,
      "loss": 2.5733,
      "theoretical_loss": 3.363522795327683,
      "tokens_seen": 2547056640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002321010101010101,
      "loss": 2.5229,
      "theoretical_loss": 3.3635159178989005,
      "tokens_seen": 2547122176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002320808080808081,
      "loss": 2.365,
      "theoretical_loss": 3.3635090406966133,
      "tokens_seen": 2547187712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023206060606060605,
      "loss": 2.5908,
      "theoretical_loss": 3.363502163720808,
      "tokens_seen": 2547253248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023204040404040406,
      "loss": 2.6327,
      "theoretical_loss": 3.363495286971472,
      "tokens_seen": 2547318784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023202020202020202,
      "loss": 2.5876,
      "theoretical_loss": 3.3634884104485905,
      "tokens_seen": 2547384320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023200000000000003,
      "loss": 2.6502,
      "theoretical_loss": 3.3634815341521516,
      "tokens_seen": 2547449856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023197979797979799,
      "loss": 2.5526,
      "theoretical_loss": 3.3634746580821413,
      "tokens_seen": 2547515392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023195959595959594,
      "loss": 2.4623,
      "theoretical_loss": 3.3634677822385464,
      "tokens_seen": 2547580928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023193939393939395,
      "loss": 2.589,
      "theoretical_loss": 3.363460906621354,
      "tokens_seen": 2547646464
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002319191919191919,
      "loss": 2.5597,
      "theoretical_loss": 3.3634540312305505,
      "tokens_seen": 2547712000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023189898989898992,
      "loss": 2.5301,
      "theoretical_loss": 3.3634471560661225,
      "tokens_seen": 2547777536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023187878787878788,
      "loss": 2.6186,
      "theoretical_loss": 3.3634402811280575,
      "tokens_seen": 2547843072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023185858585858586,
      "loss": 2.7319,
      "theoretical_loss": 3.363433406416341,
      "tokens_seen": 2547908608
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1395695,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3248932361602783,
      "objective/train/theoretical_loss": 3.363426531930961,
      "objective/train/tokens_used": 906833376,
      "theoretical_loss": 3.363426531930961,
      "tokens_seen": 2547974144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023183838383838385,
      "loss": 2.5215,
      "theoretical_loss": 3.363426531930961,
      "tokens_seen": 2547974144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023181818181818183,
      "loss": 2.602,
      "theoretical_loss": 3.363419657671903,
      "tokens_seen": 2548039680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023179797979797981,
      "loss": 2.3246,
      "theoretical_loss": 3.3634127836391547,
      "tokens_seen": 2548105216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023177777777777777,
      "loss": 2.7113,
      "theoretical_loss": 3.3634059098327023,
      "tokens_seen": 2548170752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023175757575757575,
      "loss": 2.6343,
      "theoretical_loss": 3.363399036252533,
      "tokens_seen": 2548236288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023173737373737374,
      "loss": 2.5076,
      "theoretical_loss": 3.363392162898633,
      "tokens_seen": 2548301824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023171717171717172,
      "loss": 2.7902,
      "theoretical_loss": 3.363385289770989,
      "tokens_seen": 2548367360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002316969696969697,
      "loss": 2.7543,
      "theoretical_loss": 3.363378416869588,
      "tokens_seen": 2548432896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002316767676767677,
      "loss": 2.8748,
      "theoretical_loss": 3.3633715441944174,
      "tokens_seen": 2548498432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023165656565656565,
      "loss": 2.529,
      "theoretical_loss": 3.3633646717454626,
      "tokens_seen": 2548563968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023163636363636366,
      "loss": 2.5737,
      "theoretical_loss": 3.3633577995227113,
      "tokens_seen": 2548629504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023161616161616162,
      "loss": 2.6683,
      "theoretical_loss": 3.3633509275261497,
      "tokens_seen": 2548695040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315959595959596,
      "loss": 2.5447,
      "theoretical_loss": 3.363344055755765,
      "tokens_seen": 2548760576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023157575757575758,
      "loss": 2.5594,
      "theoretical_loss": 3.363337184211544,
      "tokens_seen": 2548826112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023155555555555554,
      "loss": 2.6081,
      "theoretical_loss": 3.3633303128934724,
      "tokens_seen": 2548891648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023153535353535355,
      "loss": 2.683,
      "theoretical_loss": 3.3633234418015383,
      "tokens_seen": 2548957184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315151515151515,
      "loss": 2.4854,
      "theoretical_loss": 3.3633165709357273,
      "tokens_seen": 2549022720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023149494949494952,
      "loss": 2.5795,
      "theoretical_loss": 3.3633097002960275,
      "tokens_seen": 2549088256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023147474747474748,
      "loss": 2.5687,
      "theoretical_loss": 3.363302829882424,
      "tokens_seen": 2549153792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023145454545454546,
      "loss": 2.797,
      "theoretical_loss": 3.363295959694905,
      "tokens_seen": 2549219328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023143434343434344,
      "loss": 2.7385,
      "theoretical_loss": 3.363289089733456,
      "tokens_seen": 2549284864
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002314141414141414,
      "loss": 2.8137,
      "theoretical_loss": 3.363282219998065,
      "tokens_seen": 2549350400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002313939393939394,
      "loss": 2.6756,
      "theoretical_loss": 3.363275350488718,
      "tokens_seen": 2549415936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023137373737373737,
      "loss": 2.6287,
      "theoretical_loss": 3.3632684812054015,
      "tokens_seen": 2549481472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023135353535353538,
      "loss": 2.5179,
      "theoretical_loss": 3.363261612148103,
      "tokens_seen": 2549547008
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1396968,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2947990894317627,
      "objective/train/theoretical_loss": 3.3632547433168085,
      "objective/train/tokens_used": 908471776,
      "theoretical_loss": 3.3632547433168085,
      "tokens_seen": 2549612544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023133333333333334,
      "loss": 2.8115,
      "theoretical_loss": 3.3632547433168085,
      "tokens_seen": 2549612544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023131313131313132,
      "loss": 2.6321,
      "theoretical_loss": 3.3632478747115053,
      "tokens_seen": 2549678080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312929292929293,
      "loss": 2.7633,
      "theoretical_loss": 3.3632410063321805,
      "tokens_seen": 2549743616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023127272727272726,
      "loss": 2.4416,
      "theoretical_loss": 3.3632341381788198,
      "tokens_seen": 2549809152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023125252525252527,
      "loss": 2.4922,
      "theoretical_loss": 3.3632272702514108,
      "tokens_seen": 2549874688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023123232323232323,
      "loss": 2.555,
      "theoretical_loss": 3.3632204025499397,
      "tokens_seen": 2549940224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312121212121212,
      "loss": 2.5403,
      "theoretical_loss": 3.3632135350743937,
      "tokens_seen": 2550005760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002311919191919192,
      "loss": 2.4247,
      "theoretical_loss": 3.3632066678247594,
      "tokens_seen": 2550071296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023117171717171718,
      "loss": 2.5022,
      "theoretical_loss": 3.3631998008010235,
      "tokens_seen": 2550136832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023115151515151517,
      "loss": 2.5859,
      "theoretical_loss": 3.3631929340031728,
      "tokens_seen": 2550202368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023113131313131315,
      "loss": 2.8307,
      "theoretical_loss": 3.363186067431194,
      "tokens_seen": 2550267904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002311111111111111,
      "loss": 2.7717,
      "theoretical_loss": 3.3631792010850745,
      "tokens_seen": 2550333440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002310909090909091,
      "loss": 2.6598,
      "theoretical_loss": 3.3631723349648,
      "tokens_seen": 2550398976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023107070707070707,
      "loss": 2.5159,
      "theoretical_loss": 3.363165469070358,
      "tokens_seen": 2550464512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023105050505050506,
      "loss": 2.4401,
      "theoretical_loss": 3.363158603401735,
      "tokens_seen": 2550530048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023103030303030304,
      "loss": 2.4653,
      "theoretical_loss": 3.3631517379589178,
      "tokens_seen": 2550595584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000231010101010101,
      "loss": 2.5412,
      "theoretical_loss": 3.363144872741893,
      "tokens_seen": 2550661120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000230989898989899,
      "loss": 2.4256,
      "theoretical_loss": 3.363138007750648,
      "tokens_seen": 2550726656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023096969696969697,
      "loss": 2.5071,
      "theoretical_loss": 3.3631311429851687,
      "tokens_seen": 2550792192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023094949494949498,
      "loss": 2.3989,
      "theoretical_loss": 3.3631242784454427,
      "tokens_seen": 2550857728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023092929292929293,
      "loss": 2.6109,
      "theoretical_loss": 3.363117414131456,
      "tokens_seen": 2550923264
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002309090909090909,
      "loss": 2.6367,
      "theoretical_loss": 3.3631105500431966,
      "tokens_seen": 2550988800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002308888888888889,
      "loss": 2.5224,
      "theoretical_loss": 3.36310368618065,
      "tokens_seen": 2551054336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023086868686868686,
      "loss": 2.4318,
      "theoretical_loss": 3.363096822543803,
      "tokens_seen": 2551119872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023084848484848487,
      "loss": 2.468,
      "theoretical_loss": 3.363089959132643,
      "tokens_seen": 2551185408
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1397903,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4797961711883545,
      "objective/train/theoretical_loss": 3.363083095947157,
      "objective/train/tokens_used": 910110176,
      "theoretical_loss": 3.363083095947157,
      "tokens_seen": 2551250944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023082828282828283,
      "loss": 2.6271,
      "theoretical_loss": 3.363083095947157,
      "tokens_seen": 2551250944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002308080808080808,
      "loss": 2.6973,
      "theoretical_loss": 3.363076232987331,
      "tokens_seen": 2551316480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307878787878788,
      "loss": 2.7594,
      "theoretical_loss": 3.3630693702531524,
      "tokens_seen": 2551382016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023076767676767675,
      "loss": 2.4957,
      "theoretical_loss": 3.3630625077446075,
      "tokens_seen": 2551447552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023074747474747476,
      "loss": 2.7009,
      "theoretical_loss": 3.3630556454616833,
      "tokens_seen": 2551513088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023072727272727272,
      "loss": 2.6245,
      "theoretical_loss": 3.363048783404367,
      "tokens_seen": 2551578624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307070707070707,
      "loss": 2.4169,
      "theoretical_loss": 3.363041921572645,
      "tokens_seen": 2551644160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002306868686868687,
      "loss": 2.4232,
      "theoretical_loss": 3.3630350599665038,
      "tokens_seen": 2551709696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023066666666666667,
      "loss": 2.6133,
      "theoretical_loss": 3.3630281985859307,
      "tokens_seen": 2551775232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023064646464646466,
      "loss": 2.545,
      "theoretical_loss": 3.3630213374309124,
      "tokens_seen": 2551840768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023062626262626264,
      "loss": 2.419,
      "theoretical_loss": 3.3630144765014354,
      "tokens_seen": 2551906304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002306060606060606,
      "loss": 2.6984,
      "theoretical_loss": 3.3630076157974864,
      "tokens_seen": 2551971840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023058585858585858,
      "loss": 2.7016,
      "theoretical_loss": 3.363000755319053,
      "tokens_seen": 2552037376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023056565656565656,
      "loss": 2.4047,
      "theoretical_loss": 3.362993895066121,
      "tokens_seen": 2552102912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023054545454545455,
      "loss": 2.6434,
      "theoretical_loss": 3.3629870350386777,
      "tokens_seen": 2552168448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023052525252525253,
      "loss": 2.704,
      "theoretical_loss": 3.3629801752367103,
      "tokens_seen": 2552233984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023050505050505052,
      "loss": 2.6878,
      "theoretical_loss": 3.362973315660205,
      "tokens_seen": 2552299520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304848484848485,
      "loss": 2.6135,
      "theoretical_loss": 3.3629664563091484,
      "tokens_seen": 2552365056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023046464646464646,
      "loss": 2.392,
      "theoretical_loss": 3.362959597183528,
      "tokens_seen": 2552430592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023044444444444447,
      "loss": 2.608,
      "theoretical_loss": 3.36295273828333,
      "tokens_seen": 2552496128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023042424242424243,
      "loss": 2.57,
      "theoretical_loss": 3.3629458796085414,
      "tokens_seen": 2552561664
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304040404040404,
      "loss": 2.6565,
      "theoretical_loss": 3.3629390211591494,
      "tokens_seen": 2552627200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002303838383838384,
      "loss": 2.4804,
      "theoretical_loss": 3.3629321629351407,
      "tokens_seen": 2552692736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023036363636363635,
      "loss": 2.5498,
      "theoretical_loss": 3.3629253049365015,
      "tokens_seen": 2552758272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023034343434343436,
      "loss": 2.6348,
      "theoretical_loss": 3.362918447163219,
      "tokens_seen": 2552823808
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1398368,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.575436592102051,
      "objective/train/theoretical_loss": 3.36291158961528,
      "objective/train/tokens_used": 911748576,
      "theoretical_loss": 3.36291158961528,
      "tokens_seen": 2552889344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023032323232323232,
      "loss": 2.7566,
      "theoretical_loss": 3.36291158961528,
      "tokens_seen": 2552889344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023030303030303033,
      "loss": 2.5712,
      "theoretical_loss": 3.3629047322926713,
      "tokens_seen": 2552954880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023028282828282829,
      "loss": 2.6174,
      "theoretical_loss": 3.36289787519538,
      "tokens_seen": 2553020416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023026262626262627,
      "loss": 2.3893,
      "theoretical_loss": 3.3628910183233924,
      "tokens_seen": 2553085952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023024242424242425,
      "loss": 2.4815,
      "theoretical_loss": 3.362884161676696,
      "tokens_seen": 2553151488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002302222222222222,
      "loss": 2.6386,
      "theoretical_loss": 3.3628773052552763,
      "tokens_seen": 2553217024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023020202020202022,
      "loss": 2.4313,
      "theoretical_loss": 3.3628704490591215,
      "tokens_seen": 2553282560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023018181818181818,
      "loss": 2.4207,
      "theoretical_loss": 3.362863593088218,
      "tokens_seen": 2553348096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023016161616161616,
      "loss": 2.7121,
      "theoretical_loss": 3.362856737342552,
      "tokens_seen": 2553413632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023014141414141415,
      "loss": 2.5634,
      "theoretical_loss": 3.3628498818221115,
      "tokens_seen": 2553479168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023012121212121213,
      "loss": 2.7912,
      "theoretical_loss": 3.3628430265268823,
      "tokens_seen": 2553544704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023010101010101011,
      "loss": 2.4155,
      "theoretical_loss": 3.362836171456852,
      "tokens_seen": 2553610240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002300808080808081,
      "loss": 2.5919,
      "theoretical_loss": 3.3628293166120065,
      "tokens_seen": 2553675776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023006060606060606,
      "loss": 2.6233,
      "theoretical_loss": 3.3628224619923333,
      "tokens_seen": 2553741312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023004040404040404,
      "loss": 2.7086,
      "theoretical_loss": 3.362815607597819,
      "tokens_seen": 2553806848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023002020202020202,
      "loss": 2.4947,
      "theoretical_loss": 3.3628087534284505,
      "tokens_seen": 2553872384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023,
      "loss": 2.7466,
      "theoretical_loss": 3.362801899484215,
      "tokens_seen": 2553937920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000229979797979798,
      "loss": 2.5205,
      "theoretical_loss": 3.3627950457650986,
      "tokens_seen": 2554003456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022995959595959595,
      "loss": 2.3375,
      "theoretical_loss": 3.3627881922710885,
      "tokens_seen": 2554068992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022993939393939396,
      "loss": 2.5964,
      "theoretical_loss": 3.3627813390021712,
      "tokens_seen": 2554134528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022991919191919192,
      "loss": 2.5728,
      "theoretical_loss": 3.3627744859583344,
      "tokens_seen": 2554200064
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002298989898989899,
      "loss": 2.4382,
      "theoretical_loss": 3.3627676331395637,
      "tokens_seen": 2554265600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022987878787878788,
      "loss": 2.5344,
      "theoretical_loss": 3.3627607805458473,
      "tokens_seen": 2554331136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022985858585858584,
      "loss": 2.3541,
      "theoretical_loss": 3.3627539281771712,
      "tokens_seen": 2554396672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022983838383838385,
      "loss": 2.6214,
      "theoretical_loss": 3.362747076033522,
      "tokens_seen": 2554462208
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 1399295,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6599607467651367,
      "objective/train/theoretical_loss": 3.362740224114887,
      "objective/train/tokens_used": 913386976,
      "theoretical_loss": 3.362740224114887,
      "tokens_seen": 2554527744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002298181818181818,
      "loss": 2.4687,
      "theoretical_loss": 3.362740224114887,
      "tokens_seen": 2554527744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022979797979797982,
      "loss": 2.5853,
      "theoretical_loss": 3.362733372421253,
      "tokens_seen": 2554593280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022977777777777778,
      "loss": 2.8089,
      "theoretical_loss": 3.3627265209526067,
      "tokens_seen": 2554658816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022975757575757576,
      "loss": 2.577,
      "theoretical_loss": 3.3627196697089357,
      "tokens_seen": 2554724352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022973737373737374,
      "loss": 2.5586,
      "theoretical_loss": 3.362712818690225,
      "tokens_seen": 2554789888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002297171717171717,
      "loss": 2.6406,
      "theoretical_loss": 3.3627059678964635,
      "tokens_seen": 2554855424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002296969696969697,
      "loss": 2.7724,
      "theoretical_loss": 3.362699117327637,
      "tokens_seen": 2554920960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022967676767676767,
      "loss": 2.5388,
      "theoretical_loss": 3.362692266983732,
      "tokens_seen": 2554986496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022965656565656568,
      "loss": 2.5942,
      "theoretical_loss": 3.3626854168647364,
      "tokens_seen": 2555052032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022963636363636364,
      "loss": 2.5285,
      "theoretical_loss": 3.3626785669706365,
      "tokens_seen": 2555117568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022961616161616162,
      "loss": 2.5427,
      "theoretical_loss": 3.362671717301419,
      "tokens_seen": 2555183104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002295959595959596,
      "loss": 2.48,
      "theoretical_loss": 3.3626648678570707,
      "tokens_seen": 2555248640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002295757575757576,
      "loss": 2.4674,
      "theoretical_loss": 3.3626580186375787,
      "tokens_seen": 2555314176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022955555555555557,
      "loss": 2.6681,
      "theoretical_loss": 3.36265116964293,
      "tokens_seen": 2555379712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022953535353535353,
      "loss": 2.5396,
      "theoretical_loss": 3.362644320873111,
      "tokens_seen": 2555445248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022951515151515151,
      "loss": 2.4922,
      "theoretical_loss": 3.362637472328109,
      "tokens_seen": 2555510784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002294949494949495,
      "loss": 2.5783,
      "theoretical_loss": 3.3626306240079105,
      "tokens_seen": 2555576320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022947474747474748,
      "loss": 2.7018,
      "theoretical_loss": 3.362623775912503,
      "tokens_seen": 2555641856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022945454545454547,
      "loss": 2.6705,
      "theoretical_loss": 3.3626169280418723,
      "tokens_seen": 2555707392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022943434343434345,
      "loss": 2.6724,
      "theoretical_loss": 3.362610080396006,
      "tokens_seen": 2555772928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002294141414141414,
      "loss": 2.6912,
      "theoretical_loss": 3.3626032329748905,
      "tokens_seen": 2555838464
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022939393939393942,
      "loss": 2.6626,
      "theoretical_loss": 3.3625963857785135,
      "tokens_seen": 2555904000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022937373737373737,
      "loss": 2.4301,
      "theoretical_loss": 3.362589538806861,
      "tokens_seen": 2555969536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022935353535353536,
      "loss": 2.4563,
      "theoretical_loss": 3.36258269205992,
      "tokens_seen": 2556035072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022933333333333334,
      "loss": 2.588,
      "theoretical_loss": 3.3625758455376777,
      "tokens_seen": 2556100608
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1400625,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.826688289642334,
      "objective/train/theoretical_loss": 3.362568999240121,
      "objective/train/tokens_used": 915025376,
      "theoretical_loss": 3.362568999240121,
      "tokens_seen": 2556166144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002293131313131313,
      "loss": 2.7183,
      "theoretical_loss": 3.362568999240121,
      "tokens_seen": 2556166144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002292929292929293,
      "loss": 2.6652,
      "theoretical_loss": 3.3625621531672367,
      "tokens_seen": 2556231680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022927272727272727,
      "loss": 2.4161,
      "theoretical_loss": 3.362555307319011,
      "tokens_seen": 2556297216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022925252525252528,
      "loss": 2.526,
      "theoretical_loss": 3.3625484616954315,
      "tokens_seen": 2556362752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022923232323232323,
      "loss": 2.714,
      "theoretical_loss": 3.362541616296485,
      "tokens_seen": 2556428288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002292121212121212,
      "loss": 2.6548,
      "theoretical_loss": 3.3625347711221583,
      "tokens_seen": 2556493824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291919191919192,
      "loss": 2.5988,
      "theoretical_loss": 3.362527926172438,
      "tokens_seen": 2556559360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022917171717171716,
      "loss": 2.5166,
      "theoretical_loss": 3.362521081447311,
      "tokens_seen": 2556624896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022915151515151517,
      "loss": 2.7294,
      "theoretical_loss": 3.3625142369467644,
      "tokens_seen": 2556690432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022913131313131313,
      "loss": 2.7967,
      "theoretical_loss": 3.3625073926707856,
      "tokens_seen": 2556755968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291111111111111,
      "loss": 2.5215,
      "theoretical_loss": 3.3625005486193604,
      "tokens_seen": 2556821504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002290909090909091,
      "loss": 2.4801,
      "theoretical_loss": 3.3624937047924766,
      "tokens_seen": 2556887040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022907070707070708,
      "loss": 2.7837,
      "theoretical_loss": 3.36248686119012,
      "tokens_seen": 2556952576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022905050505050506,
      "loss": 2.4681,
      "theoretical_loss": 3.362480017812279,
      "tokens_seen": 2557018112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022903030303030302,
      "loss": 2.6862,
      "theoretical_loss": 3.3624731746589394,
      "tokens_seen": 2557083648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000229010101010101,
      "loss": 2.5918,
      "theoretical_loss": 3.362466331730088,
      "tokens_seen": 2557149184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000228989898989899,
      "loss": 2.488,
      "theoretical_loss": 3.362459489025712,
      "tokens_seen": 2557214720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022896969696969697,
      "loss": 2.4771,
      "theoretical_loss": 3.3624526465457985,
      "tokens_seen": 2557280256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022894949494949496,
      "loss": 2.5262,
      "theoretical_loss": 3.3624458042903336,
      "tokens_seen": 2557345792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022892929292929294,
      "loss": 2.5601,
      "theoretical_loss": 3.362438962259305,
      "tokens_seen": 2557411328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002289090909090909,
      "loss": 2.5782,
      "theoretical_loss": 3.3624321204527,
      "tokens_seen": 2557476864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002288888888888889,
      "loss": 2.5581,
      "theoretical_loss": 3.362425278870504,
      "tokens_seen": 2557542400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022886868686868686,
      "loss": 2.6424,
      "theoretical_loss": 3.362418437512705,
      "tokens_seen": 2557607936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022884848484848485,
      "loss": 2.6295,
      "theoretical_loss": 3.3624115963792898,
      "tokens_seen": 2557673472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022882828282828283,
      "loss": 2.8505,
      "theoretical_loss": 3.3624047554702448,
      "tokens_seen": 2557739008
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1401375,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7073214054107666,
      "objective/train/theoretical_loss": 3.362397914785557,
      "objective/train/tokens_used": 916663776,
      "theoretical_loss": 3.362397914785557,
      "tokens_seen": 2557804544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022880808080808082,
      "loss": 2.7022,
      "theoretical_loss": 3.362397914785557,
      "tokens_seen": 2557804544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002287878787878788,
      "loss": 2.6011,
      "theoretical_loss": 3.362391074325214,
      "tokens_seen": 2557870080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022876767676767676,
      "loss": 2.7017,
      "theoretical_loss": 3.3623842340892023,
      "tokens_seen": 2557935616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022874747474747477,
      "loss": 2.7039,
      "theoretical_loss": 3.3623773940775084,
      "tokens_seen": 2558001152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022872727272727273,
      "loss": 2.5218,
      "theoretical_loss": 3.362370554290119,
      "tokens_seen": 2558066688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022870707070707074,
      "loss": 2.7054,
      "theoretical_loss": 3.3623637147270222,
      "tokens_seen": 2558132224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002286868686868687,
      "loss": 2.6012,
      "theoretical_loss": 3.362356875388204,
      "tokens_seen": 2558197760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022866666666666665,
      "loss": 2.5456,
      "theoretical_loss": 3.362350036273651,
      "tokens_seen": 2558263296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022864646464646466,
      "loss": 2.5671,
      "theoretical_loss": 3.3623431973833506,
      "tokens_seen": 2558328832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022862626262626262,
      "loss": 2.6018,
      "theoretical_loss": 3.36233635871729,
      "tokens_seen": 2558394368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022860606060606063,
      "loss": 2.4401,
      "theoretical_loss": 3.362329520275456,
      "tokens_seen": 2558459904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022858585858585859,
      "loss": 2.5295,
      "theoretical_loss": 3.362322682057835,
      "tokens_seen": 2558525440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022856565656565657,
      "loss": 2.5317,
      "theoretical_loss": 3.3623158440644136,
      "tokens_seen": 2558590976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022854545454545455,
      "loss": 2.6185,
      "theoretical_loss": 3.36230900629518,
      "tokens_seen": 2558656512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022852525252525254,
      "loss": 2.5875,
      "theoretical_loss": 3.3623021687501202,
      "tokens_seen": 2558722048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022850505050505052,
      "loss": 2.5357,
      "theoretical_loss": 3.3622953314292214,
      "tokens_seen": 2558787584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022848484848484848,
      "loss": 2.5946,
      "theoretical_loss": 3.3622884943324705,
      "tokens_seen": 2558853120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022846464646464646,
      "loss": 2.5611,
      "theoretical_loss": 3.362281657459854,
      "tokens_seen": 2558918656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022844444444444445,
      "loss": 2.4251,
      "theoretical_loss": 3.3622748208113595,
      "tokens_seen": 2558984192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022842424242424243,
      "loss": 2.3916,
      "theoretical_loss": 3.3622679843869734,
      "tokens_seen": 2559049728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022840404040404041,
      "loss": 2.6423,
      "theoretical_loss": 3.3622611481866826,
      "tokens_seen": 2559115264
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283838383838384,
      "loss": 2.333,
      "theoretical_loss": 3.3622543122104744,
      "tokens_seen": 2559180800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022836363636363636,
      "loss": 2.7554,
      "theoretical_loss": 3.3622474764583354,
      "tokens_seen": 2559246336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022834343434343437,
      "loss": 2.7019,
      "theoretical_loss": 3.3622406409302528,
      "tokens_seen": 2559311872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022832323232323232,
      "loss": 2.7094,
      "theoretical_loss": 3.3622338056262127,
      "tokens_seen": 2559377408
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1402702,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7075891494750977,
      "objective/train/theoretical_loss": 3.362226970546203,
      "objective/train/tokens_used": 918302176,
      "theoretical_loss": 3.362226970546203,
      "tokens_seen": 2559442944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283030303030303,
      "loss": 2.5418,
      "theoretical_loss": 3.362226970546203,
      "tokens_seen": 2559442944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002282828282828283,
      "loss": 2.6696,
      "theoretical_loss": 3.362220135690211,
      "tokens_seen": 2559508480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022826262626262625,
      "loss": 2.5328,
      "theoretical_loss": 3.362213301058222,
      "tokens_seen": 2559574016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022824242424242426,
      "loss": 2.3914,
      "theoretical_loss": 3.362206466650224,
      "tokens_seen": 2559639552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022822222222222222,
      "loss": 2.5646,
      "theoretical_loss": 3.362199632466204,
      "tokens_seen": 2559705088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022820202020202023,
      "loss": 2.6797,
      "theoretical_loss": 3.3621927985061486,
      "tokens_seen": 2559770624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022818181818181818,
      "loss": 2.6857,
      "theoretical_loss": 3.3621859647700445,
      "tokens_seen": 2559836160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022816161616161614,
      "loss": 2.5767,
      "theoretical_loss": 3.3621791312578795,
      "tokens_seen": 2559901696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022814141414141415,
      "loss": 2.3725,
      "theoretical_loss": 3.3621722979696393,
      "tokens_seen": 2559967232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002281212121212121,
      "loss": 2.6499,
      "theoretical_loss": 3.362165464905312,
      "tokens_seen": 2560032768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022810101010101012,
      "loss": 2.4565,
      "theoretical_loss": 3.362158632064884,
      "tokens_seen": 2560098304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022808080808080808,
      "loss": 2.7,
      "theoretical_loss": 3.362151799448342,
      "tokens_seen": 2560163840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022806060606060606,
      "loss": 2.4886,
      "theoretical_loss": 3.362144967055673,
      "tokens_seen": 2560229376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022804040404040404,
      "loss": 2.5594,
      "theoretical_loss": 3.3621381348868646,
      "tokens_seen": 2560294912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022802020202020203,
      "loss": 2.6457,
      "theoretical_loss": 3.362131302941903,
      "tokens_seen": 2560360448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000228,
      "loss": 2.5623,
      "theoretical_loss": 3.3621244712207754,
      "tokens_seen": 2560425984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022797979797979797,
      "loss": 2.5425,
      "theoretical_loss": 3.3621176397234684,
      "tokens_seen": 2560491520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022795959595959598,
      "loss": 2.4453,
      "theoretical_loss": 3.36211080844997,
      "tokens_seen": 2560557056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022793939393939394,
      "loss": 2.7774,
      "theoretical_loss": 3.362103977400266,
      "tokens_seen": 2560622592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022791919191919192,
      "loss": 2.5743,
      "theoretical_loss": 3.3620971465743437,
      "tokens_seen": 2560688128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002278989898989899,
      "loss": 2.6519,
      "theoretical_loss": 3.36209031597219,
      "tokens_seen": 2560753664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002278787878787879,
      "loss": 2.6161,
      "theoretical_loss": 3.3620834855937924,
      "tokens_seen": 2560819200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022785858585858587,
      "loss": 2.5052,
      "theoretical_loss": 3.362076655439137,
      "tokens_seen": 2560884736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022783838383838386,
      "loss": 2.7581,
      "theoretical_loss": 3.362069825508211,
      "tokens_seen": 2560950272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022781818181818181,
      "loss": 2.4574,
      "theoretical_loss": 3.3620629958010015,
      "tokens_seen": 2561015808
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1403266,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9517369270324707,
      "objective/train/theoretical_loss": 3.3620561663174957,
      "objective/train/tokens_used": 919940576,
      "theoretical_loss": 3.3620561663174957,
      "tokens_seen": 2561081344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002277979797979798,
      "loss": 2.7024,
      "theoretical_loss": 3.3620561663174957,
      "tokens_seen": 2561081344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022777777777777778,
      "loss": 2.5378,
      "theoretical_loss": 3.3620493370576803,
      "tokens_seen": 2561146880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022775757575757577,
      "loss": 2.2478,
      "theoretical_loss": 3.362042508021542,
      "tokens_seen": 2561212416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022773737373737375,
      "loss": 2.5882,
      "theoretical_loss": 3.362035679209068,
      "tokens_seen": 2561277952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002277171717171717,
      "loss": 2.5611,
      "theoretical_loss": 3.3620288506202454,
      "tokens_seen": 2561343488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022769696969696972,
      "loss": 2.7747,
      "theoretical_loss": 3.3620220222550605,
      "tokens_seen": 2561409024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022767676767676767,
      "loss": 2.8544,
      "theoretical_loss": 3.3620151941135012,
      "tokens_seen": 2561474560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022765656565656569,
      "loss": 2.4052,
      "theoretical_loss": 3.3620083661955538,
      "tokens_seen": 2561540096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022763636363636364,
      "loss": 2.3106,
      "theoretical_loss": 3.3620015385012056,
      "tokens_seen": 2561605632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002276161616161616,
      "loss": 2.6862,
      "theoretical_loss": 3.361994711030443,
      "tokens_seen": 2561671168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275959595959596,
      "loss": 2.5657,
      "theoretical_loss": 3.3619878837832537,
      "tokens_seen": 2561736704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022757575757575757,
      "loss": 2.6497,
      "theoretical_loss": 3.361981056759624,
      "tokens_seen": 2561802240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022755555555555558,
      "loss": 2.6136,
      "theoretical_loss": 3.361974229959541,
      "tokens_seen": 2561867776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022753535353535354,
      "loss": 2.6712,
      "theoretical_loss": 3.3619674033829927,
      "tokens_seen": 2561933312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022751515151515152,
      "loss": 2.6558,
      "theoretical_loss": 3.3619605770299645,
      "tokens_seen": 2561998848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274949494949495,
      "loss": 2.474,
      "theoretical_loss": 3.3619537509004442,
      "tokens_seen": 2562064384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022747474747474746,
      "loss": 2.6959,
      "theoretical_loss": 3.361946924994419,
      "tokens_seen": 2562129920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022745454545454547,
      "loss": 2.5167,
      "theoretical_loss": 3.361940099311875,
      "tokens_seen": 2562195456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022743434343434343,
      "loss": 2.6456,
      "theoretical_loss": 3.3619332738527996,
      "tokens_seen": 2562260992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274141414141414,
      "loss": 2.6732,
      "theoretical_loss": 3.3619264486171803,
      "tokens_seen": 2562326528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002273939393939394,
      "loss": 2.6386,
      "theoretical_loss": 3.3619196236050035,
      "tokens_seen": 2562392064
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022737373737373738,
      "loss": 2.6875,
      "theoretical_loss": 3.361912798816256,
      "tokens_seen": 2562457600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022735353535353536,
      "loss": 2.7069,
      "theoretical_loss": 3.361905974250925,
      "tokens_seen": 2562523136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022733333333333335,
      "loss": 2.4783,
      "theoretical_loss": 3.3618991499089974,
      "tokens_seen": 2562588672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002273131313131313,
      "loss": 2.5364,
      "theoretical_loss": 3.3618923257904605,
      "tokens_seen": 2562654208
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1404500,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7659013271331787,
      "objective/train/theoretical_loss": 3.3618855018953013,
      "objective/train/tokens_used": 921578976,
      "theoretical_loss": 3.3618855018953013,
      "tokens_seen": 2562719744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002272929292929293,
      "loss": 2.5743,
      "theoretical_loss": 3.3618855018953013,
      "tokens_seen": 2562719744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022727272727272727,
      "loss": 2.5772,
      "theoretical_loss": 3.3618786782235066,
      "tokens_seen": 2562785280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022725252525252526,
      "loss": 2.4968,
      "theoretical_loss": 3.361871854775063,
      "tokens_seen": 2562850816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022723232323232324,
      "loss": 2.6017,
      "theoretical_loss": 3.3618650315499576,
      "tokens_seen": 2562916352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002272121212121212,
      "loss": 2.5718,
      "theoretical_loss": 3.3618582085481776,
      "tokens_seen": 2562981888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002271919191919192,
      "loss": 2.4662,
      "theoretical_loss": 3.3618513857697105,
      "tokens_seen": 2563047424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022717171717171717,
      "loss": 2.5702,
      "theoretical_loss": 3.361844563214542,
      "tokens_seen": 2563112960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022715151515151518,
      "loss": 2.5992,
      "theoretical_loss": 3.3618377408826605,
      "tokens_seen": 2563178496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022713131313131313,
      "loss": 2.7946,
      "theoretical_loss": 3.361830918774052,
      "tokens_seen": 2563244032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022711111111111112,
      "loss": 2.7117,
      "theoretical_loss": 3.3618240968887037,
      "tokens_seen": 2563309568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002270909090909091,
      "loss": 2.3677,
      "theoretical_loss": 3.3618172752266027,
      "tokens_seen": 2563375104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022707070707070706,
      "loss": 2.4435,
      "theoretical_loss": 3.361810453787736,
      "tokens_seen": 2563440640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022705050505050507,
      "loss": 2.5919,
      "theoretical_loss": 3.361803632572091,
      "tokens_seen": 2563506176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022703030303030303,
      "loss": 2.6822,
      "theoretical_loss": 3.3617968115796533,
      "tokens_seen": 2563571712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022701010101010104,
      "loss": 2.6193,
      "theoretical_loss": 3.3617899908104114,
      "tokens_seen": 2563637248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000226989898989899,
      "loss": 2.435,
      "theoretical_loss": 3.3617831702643515,
      "tokens_seen": 2563702784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022696969696969698,
      "loss": 2.4034,
      "theoretical_loss": 3.361776349941461,
      "tokens_seen": 2563768320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022694949494949496,
      "loss": 2.4265,
      "theoretical_loss": 3.3617695298417267,
      "tokens_seen": 2563833856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022692929292929292,
      "loss": 2.3804,
      "theoretical_loss": 3.3617627099651353,
      "tokens_seen": 2563899392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022690909090909093,
      "loss": 2.721,
      "theoretical_loss": 3.361755890311674,
      "tokens_seen": 2563964928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002268888888888889,
      "loss": 2.5143,
      "theoretical_loss": 3.3617490708813307,
      "tokens_seen": 2564030464
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022686868686868687,
      "loss": 2.3869,
      "theoretical_loss": 3.361742251674091,
      "tokens_seen": 2564096000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022684848484848485,
      "loss": 2.5408,
      "theoretical_loss": 3.3617354326899425,
      "tokens_seen": 2564161536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022682828282828284,
      "loss": 2.5471,
      "theoretical_loss": 3.361728613928872,
      "tokens_seen": 2564227072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022680808080808082,
      "loss": 2.6331,
      "theoretical_loss": 3.361721795390867,
      "tokens_seen": 2564292608
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1405265,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5630226135253906,
      "objective/train/theoretical_loss": 3.361714977075914,
      "objective/train/tokens_used": 923217376,
      "theoretical_loss": 3.361714977075914,
      "tokens_seen": 2564358144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002267878787878788,
      "loss": 2.7071,
      "theoretical_loss": 3.361714977075914,
      "tokens_seen": 2564358144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022676767676767676,
      "loss": 2.5375,
      "theoretical_loss": 3.3617081589840003,
      "tokens_seen": 2564423680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022674747474747475,
      "loss": 2.4482,
      "theoretical_loss": 3.3617013411151127,
      "tokens_seen": 2564489216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022672727272727273,
      "loss": 2.404,
      "theoretical_loss": 3.3616945234692386,
      "tokens_seen": 2564554752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022670707070707072,
      "loss": 2.6495,
      "theoretical_loss": 3.3616877060463644,
      "tokens_seen": 2564620288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266868686868687,
      "loss": 2.5744,
      "theoretical_loss": 3.361680888846477,
      "tokens_seen": 2564685824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022666666666666666,
      "loss": 2.6006,
      "theoretical_loss": 3.3616740718695644,
      "tokens_seen": 2564751360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022664646464646467,
      "loss": 2.6253,
      "theoretical_loss": 3.3616672551156133,
      "tokens_seen": 2564816896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022662626262626262,
      "loss": 2.433,
      "theoretical_loss": 3.36166043858461,
      "tokens_seen": 2564882432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266060606060606,
      "loss": 2.5285,
      "theoretical_loss": 3.361653622276542,
      "tokens_seen": 2564947968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002265858585858586,
      "loss": 2.5386,
      "theoretical_loss": 3.361646806191396,
      "tokens_seen": 2565013504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022656565656565655,
      "loss": 2.4593,
      "theoretical_loss": 3.36163999032916,
      "tokens_seen": 2565079040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022654545454545456,
      "loss": 2.554,
      "theoretical_loss": 3.36163317468982,
      "tokens_seen": 2565144576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022652525252525252,
      "loss": 2.5766,
      "theoretical_loss": 3.361626359273363,
      "tokens_seen": 2565210112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022650505050505053,
      "loss": 2.4994,
      "theoretical_loss": 3.3616195440797765,
      "tokens_seen": 2565275648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022648484848484848,
      "loss": 2.7954,
      "theoretical_loss": 3.3616127291090474,
      "tokens_seen": 2565341184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022646464646464647,
      "loss": 2.478,
      "theoretical_loss": 3.3616059143611623,
      "tokens_seen": 2565406720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022644444444444445,
      "loss": 2.4961,
      "theoretical_loss": 3.361599099836109,
      "tokens_seen": 2565472256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002264242424242424,
      "loss": 2.5967,
      "theoretical_loss": 3.3615922855338742,
      "tokens_seen": 2565537792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022640404040404042,
      "loss": 2.3754,
      "theoretical_loss": 3.3615854714544446,
      "tokens_seen": 2565603328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022638383838383838,
      "loss": 2.6623,
      "theoretical_loss": 3.3615786575978075,
      "tokens_seen": 2565668864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022636363636363636,
      "loss": 2.6872,
      "theoretical_loss": 3.36157184396395,
      "tokens_seen": 2565734400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022634343434343435,
      "loss": 2.6839,
      "theoretical_loss": 3.3615650305528586,
      "tokens_seen": 2565799936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022632323232323233,
      "loss": 2.6944,
      "theoretical_loss": 3.361558217364521,
      "tokens_seen": 2565865472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002263030303030303,
      "loss": 2.5876,
      "theoretical_loss": 3.361551404398924,
      "tokens_seen": 2565931008
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1405984,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.499677896499634,
      "objective/train/theoretical_loss": 3.3615445916560547,
      "objective/train/tokens_used": 924855776,
      "theoretical_loss": 3.3615445916560547,
      "tokens_seen": 2565996544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002262828282828283,
      "loss": 2.4251,
      "theoretical_loss": 3.3615445916560547,
      "tokens_seen": 2565996544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022626262626262628,
      "loss": 2.2787,
      "theoretical_loss": 3.3615377791359,
      "tokens_seen": 2566062080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022624242424242424,
      "loss": 2.6603,
      "theoretical_loss": 3.3615309668384468,
      "tokens_seen": 2566127616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022622222222222222,
      "loss": 2.6231,
      "theoretical_loss": 3.361524154763682,
      "tokens_seen": 2566193152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002262020202020202,
      "loss": 2.6168,
      "theoretical_loss": 3.3615173429115934,
      "tokens_seen": 2566258688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002261818181818182,
      "loss": 2.5521,
      "theoretical_loss": 3.3615105312821676,
      "tokens_seen": 2566324224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022616161616161617,
      "loss": 2.5061,
      "theoretical_loss": 3.361503719875391,
      "tokens_seen": 2566389760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022614141414141416,
      "loss": 2.7799,
      "theoretical_loss": 3.3614969086912523,
      "tokens_seen": 2566455296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022612121212121211,
      "loss": 2.4515,
      "theoretical_loss": 3.3614900977297366,
      "tokens_seen": 2566520832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022610101010101013,
      "loss": 2.622,
      "theoretical_loss": 3.361483286990832,
      "tokens_seen": 2566586368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022608080808080808,
      "loss": 2.5093,
      "theoretical_loss": 3.3614764764745253,
      "tokens_seen": 2566651904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022606060606060607,
      "loss": 2.5789,
      "theoretical_loss": 3.361469666180804,
      "tokens_seen": 2566717440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022604040404040405,
      "loss": 2.7051,
      "theoretical_loss": 3.3614628561096542,
      "tokens_seen": 2566782976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000226020202020202,
      "loss": 2.6851,
      "theoretical_loss": 3.361456046261064,
      "tokens_seen": 2566848512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022600000000000002,
      "loss": 2.5935,
      "theoretical_loss": 3.361449236635019,
      "tokens_seen": 2566914048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022597979797979798,
      "loss": 2.4982,
      "theoretical_loss": 3.3614424272315078,
      "tokens_seen": 2566979584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022595959595959599,
      "loss": 2.5564,
      "theoretical_loss": 3.3614356180505167,
      "tokens_seen": 2567045120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022593939393939394,
      "loss": 2.3668,
      "theoretical_loss": 3.361428809092033,
      "tokens_seen": 2567110656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002259191919191919,
      "loss": 2.6117,
      "theoretical_loss": 3.3614220003560433,
      "tokens_seen": 2567176192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002258989898989899,
      "loss": 2.5992,
      "theoretical_loss": 3.3614151918425357,
      "tokens_seen": 2567241728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022587878787878787,
      "loss": 2.6247,
      "theoretical_loss": 3.3614083835514954,
      "tokens_seen": 2567307264
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022585858585858588,
      "loss": 2.5183,
      "theoretical_loss": 3.3614015754829114,
      "tokens_seen": 2567372800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022583838383838384,
      "loss": 2.562,
      "theoretical_loss": 3.3613947676367695,
      "tokens_seen": 2567438336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022581818181818182,
      "loss": 2.2841,
      "theoretical_loss": 3.361387960013057,
      "tokens_seen": 2567503872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002257979797979798,
      "loss": 2.605,
      "theoretical_loss": 3.3613811526117616,
      "tokens_seen": 2567569408
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1407293,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.837613105773926,
      "objective/train/theoretical_loss": 3.36137434543287,
      "objective/train/tokens_used": 926494176,
      "theoretical_loss": 3.36137434543287,
      "tokens_seen": 2567634944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002257777777777778,
      "loss": 2.7259,
      "theoretical_loss": 3.36137434543287,
      "tokens_seen": 2567634944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022575757575757577,
      "loss": 2.508,
      "theoretical_loss": 3.361367538476369,
      "tokens_seen": 2567700480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022573737373737373,
      "loss": 2.3551,
      "theoretical_loss": 3.3613607317422454,
      "tokens_seen": 2567766016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002257171717171717,
      "loss": 2.6477,
      "theoretical_loss": 3.3613539252304867,
      "tokens_seen": 2567831552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002256969696969697,
      "loss": 2.3382,
      "theoretical_loss": 3.3613471189410804,
      "tokens_seen": 2567897088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022567676767676768,
      "loss": 2.691,
      "theoretical_loss": 3.3613403128740127,
      "tokens_seen": 2567962624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022565656565656566,
      "loss": 2.4764,
      "theoretical_loss": 3.361333507029271,
      "tokens_seen": 2568028160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022563636363636365,
      "loss": 2.5417,
      "theoretical_loss": 3.361326701406843,
      "tokens_seen": 2568093696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002256161616161616,
      "loss": 2.4511,
      "theoretical_loss": 3.361319896006714,
      "tokens_seen": 2568159232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022559595959595962,
      "loss": 2.5368,
      "theoretical_loss": 3.3613130908288733,
      "tokens_seen": 2568224768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022557575757575757,
      "loss": 2.6526,
      "theoretical_loss": 3.3613062858733063,
      "tokens_seen": 2568290304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022555555555555556,
      "loss": 2.5259,
      "theoretical_loss": 3.361299481140001,
      "tokens_seen": 2568355840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022553535353535354,
      "loss": 2.5857,
      "theoretical_loss": 3.361292676628944,
      "tokens_seen": 2568421376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002255151515151515,
      "loss": 2.7697,
      "theoretical_loss": 3.361285872340123,
      "tokens_seen": 2568486912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002254949494949495,
      "loss": 2.5742,
      "theoretical_loss": 3.361279068273524,
      "tokens_seen": 2568552448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022547474747474747,
      "loss": 2.6054,
      "theoretical_loss": 3.3612722644291346,
      "tokens_seen": 2568617984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022545454545454548,
      "loss": 2.6186,
      "theoretical_loss": 3.3612654608069423,
      "tokens_seen": 2568683520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022543434343434343,
      "loss": 2.3318,
      "theoretical_loss": 3.3612586574069336,
      "tokens_seen": 2568749056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022541414141414144,
      "loss": 2.5793,
      "theoretical_loss": 3.3612518542290957,
      "tokens_seen": 2568814592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002253939393939394,
      "loss": 2.4563,
      "theoretical_loss": 3.361245051273416,
      "tokens_seen": 2568880128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022537373737373736,
      "loss": 2.457,
      "theoretical_loss": 3.361238248539881,
      "tokens_seen": 2568945664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022535353535353537,
      "loss": 2.5486,
      "theoretical_loss": 3.3612314460284782,
      "tokens_seen": 2569011200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022533333333333333,
      "loss": 2.4845,
      "theoretical_loss": 3.361224643739195,
      "tokens_seen": 2569076736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022531313131313134,
      "loss": 2.3753,
      "theoretical_loss": 3.3612178416720178,
      "tokens_seen": 2569142272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002252929292929293,
      "loss": 2.325,
      "theoretical_loss": 3.3612110398269337,
      "tokens_seen": 2569207808
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1408033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3271267414093018,
      "objective/train/theoretical_loss": 3.36120423820393,
      "objective/train/tokens_used": 928132576,
      "theoretical_loss": 3.36120423820393,
      "tokens_seen": 2569273344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022527272727272728,
      "loss": 2.3177,
      "theoretical_loss": 3.36120423820393,
      "tokens_seen": 2569273344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022525252525252526,
      "loss": 2.4534,
      "theoretical_loss": 3.3611974368029944,
      "tokens_seen": 2569338880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022523232323232325,
      "loss": 2.3449,
      "theoretical_loss": 3.361190635624113,
      "tokens_seen": 2569404416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022521212121212123,
      "loss": 2.4716,
      "theoretical_loss": 3.3611838346672736,
      "tokens_seen": 2569469952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002251919191919192,
      "loss": 2.4864,
      "theoretical_loss": 3.3611770339324627,
      "tokens_seen": 2569535488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022517171717171717,
      "loss": 2.4013,
      "theoretical_loss": 3.361170233419668,
      "tokens_seen": 2569601024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022515151515151515,
      "loss": 2.6963,
      "theoretical_loss": 3.361163433128876,
      "tokens_seen": 2569666560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022513131313131314,
      "loss": 2.3117,
      "theoretical_loss": 3.361156633060074,
      "tokens_seen": 2569732096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022511111111111112,
      "loss": 2.6837,
      "theoretical_loss": 3.3611498332132497,
      "tokens_seen": 2569797632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002250909090909091,
      "loss": 2.4776,
      "theoretical_loss": 3.361143033588389,
      "tokens_seen": 2569863168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022507070707070706,
      "loss": 2.425,
      "theoretical_loss": 3.36113623418548,
      "tokens_seen": 2569928704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022505050505050507,
      "loss": 2.524,
      "theoretical_loss": 3.361129435004509,
      "tokens_seen": 2569994240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022503030303030303,
      "loss": 2.6779,
      "theoretical_loss": 3.361122636045464,
      "tokens_seen": 2570059776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022501010101010102,
      "loss": 2.613,
      "theoretical_loss": 3.3611158373083314,
      "tokens_seen": 2570125312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000224989898989899,
      "loss": 2.6216,
      "theoretical_loss": 3.3611090387930984,
      "tokens_seen": 2570190848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022496969696969696,
      "loss": 2.6213,
      "theoretical_loss": 3.3611022404997524,
      "tokens_seen": 2570256384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022494949494949497,
      "loss": 2.3933,
      "theoretical_loss": 3.3610954424282804,
      "tokens_seen": 2570321920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022492929292929292,
      "loss": 2.6353,
      "theoretical_loss": 3.361088644578669,
      "tokens_seen": 2570387456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022490909090909094,
      "loss": 2.4455,
      "theoretical_loss": 3.3610818469509063,
      "tokens_seen": 2570452992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002248888888888889,
      "loss": 2.5442,
      "theoretical_loss": 3.3610750495449784,
      "tokens_seen": 2570518528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022486868686868685,
      "loss": 2.5144,
      "theoretical_loss": 3.3610682523608735,
      "tokens_seen": 2570584064
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022484848484848486,
      "loss": 2.6568,
      "theoretical_loss": 3.3610614553985774,
      "tokens_seen": 2570649600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022482828282828282,
      "loss": 2.8062,
      "theoretical_loss": 3.361054658658078,
      "tokens_seen": 2570715136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022480808080808083,
      "loss": 2.6866,
      "theoretical_loss": 3.3610478621393622,
      "tokens_seen": 2570780672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022478787878787878,
      "loss": 2.614,
      "theoretical_loss": 3.3610410658424175,
      "tokens_seen": 2570846208
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1409076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9130380153656006,
      "objective/train/theoretical_loss": 3.3610342697672304,
      "objective/train/tokens_used": 929770976,
      "theoretical_loss": 3.3610342697672304,
      "tokens_seen": 2570911744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022476767676767677,
      "loss": 2.477,
      "theoretical_loss": 3.3610342697672304,
      "tokens_seen": 2570911744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022474747474747475,
      "loss": 2.6016,
      "theoretical_loss": 3.361027473913788,
      "tokens_seen": 2570977280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022472727272727274,
      "loss": 2.3289,
      "theoretical_loss": 3.3610206782820784,
      "tokens_seen": 2571042816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022470707070707072,
      "loss": 2.662,
      "theoretical_loss": 3.3610138828720877,
      "tokens_seen": 2571108352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022468686868686868,
      "loss": 2.4994,
      "theoretical_loss": 3.361007087683803,
      "tokens_seen": 2571173888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022466666666666666,
      "loss": 2.7204,
      "theoretical_loss": 3.361000292717212,
      "tokens_seen": 2571239424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022464646464646465,
      "loss": 2.4349,
      "theoretical_loss": 3.3609934979723017,
      "tokens_seen": 2571304960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022462626262626263,
      "loss": 2.5819,
      "theoretical_loss": 3.360986703449059,
      "tokens_seen": 2571370496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002246060606060606,
      "loss": 2.6861,
      "theoretical_loss": 3.360979909147471,
      "tokens_seen": 2571436032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245858585858586,
      "loss": 2.6242,
      "theoretical_loss": 3.3609731150675253,
      "tokens_seen": 2571501568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022456565656565658,
      "loss": 2.4868,
      "theoretical_loss": 3.3609663212092085,
      "tokens_seen": 2571567104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022454545454545457,
      "loss": 2.4331,
      "theoretical_loss": 3.3609595275725077,
      "tokens_seen": 2571632640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022452525252525252,
      "loss": 2.5809,
      "theoretical_loss": 3.36095273415741,
      "tokens_seen": 2571698176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245050505050505,
      "loss": 2.4557,
      "theoretical_loss": 3.360945940963903,
      "tokens_seen": 2571763712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002244848484848485,
      "loss": 2.5004,
      "theoretical_loss": 3.3609391479919735,
      "tokens_seen": 2571829248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022446464646464647,
      "loss": 2.6773,
      "theoretical_loss": 3.3609323552416086,
      "tokens_seen": 2571894784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022444444444444446,
      "loss": 2.699,
      "theoretical_loss": 3.360925562712796,
      "tokens_seen": 2571960320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022442424242424241,
      "loss": 2.5101,
      "theoretical_loss": 3.3609187704055214,
      "tokens_seen": 2572025856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022440404040404043,
      "loss": 2.4041,
      "theoretical_loss": 3.3609119783197734,
      "tokens_seen": 2572091392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022438383838383838,
      "loss": 2.3493,
      "theoretical_loss": 3.360905186455539,
      "tokens_seen": 2572156928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002243636363636364,
      "loss": 2.6747,
      "theoretical_loss": 3.360898394812804,
      "tokens_seen": 2572222464
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022434343434343435,
      "loss": 2.6986,
      "theoretical_loss": 3.360891603391557,
      "tokens_seen": 2572288000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002243232323232323,
      "loss": 2.4015,
      "theoretical_loss": 3.3608848121917845,
      "tokens_seen": 2572353536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022430303030303032,
      "loss": 2.5643,
      "theoretical_loss": 3.3608780212134737,
      "tokens_seen": 2572419072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022428282828282828,
      "loss": 2.6435,
      "theoretical_loss": 3.3608712304566115,
      "tokens_seen": 2572484608
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1409639,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9053080081939697,
      "objective/train/theoretical_loss": 3.3608644399211856,
      "objective/train/tokens_used": 931409376,
      "theoretical_loss": 3.3608644399211856,
      "tokens_seen": 2572550144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002242626262626263,
      "loss": 2.6655,
      "theoretical_loss": 3.3608644399211856,
      "tokens_seen": 2572550144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022424242424242424,
      "loss": 2.6647,
      "theoretical_loss": 3.360857649607183,
      "tokens_seen": 2572615680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022422222222222223,
      "loss": 2.5757,
      "theoretical_loss": 3.3608508595145903,
      "tokens_seen": 2572681216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002242020202020202,
      "loss": 2.5073,
      "theoretical_loss": 3.360844069643395,
      "tokens_seen": 2572746752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022418181818181817,
      "loss": 2.6572,
      "theoretical_loss": 3.360837279993585,
      "tokens_seen": 2572812288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022416161616161618,
      "loss": 2.518,
      "theoretical_loss": 3.360830490565146,
      "tokens_seen": 2572877824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022414141414141414,
      "loss": 2.3146,
      "theoretical_loss": 3.3608237013580657,
      "tokens_seen": 2572943360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022412121212121212,
      "loss": 2.6971,
      "theoretical_loss": 3.3608169123723317,
      "tokens_seen": 2573008896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241010101010101,
      "loss": 2.5024,
      "theoretical_loss": 3.360810123607931,
      "tokens_seen": 2573074432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002240808080808081,
      "loss": 2.5778,
      "theoretical_loss": 3.3608033350648503,
      "tokens_seen": 2573139968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022406060606060607,
      "loss": 2.4016,
      "theoretical_loss": 3.3607965467430767,
      "tokens_seen": 2573205504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022404040404040406,
      "loss": 2.6393,
      "theoretical_loss": 3.360789758642598,
      "tokens_seen": 2573271040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000224020202020202,
      "loss": 2.5162,
      "theoretical_loss": 3.360782970763401,
      "tokens_seen": 2573336576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000224,
      "loss": 2.6595,
      "theoretical_loss": 3.3607761831054734,
      "tokens_seen": 2573402112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022397979797979798,
      "loss": 2.3669,
      "theoretical_loss": 3.360769395668801,
      "tokens_seen": 2573467648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022395959595959596,
      "loss": 2.3576,
      "theoretical_loss": 3.3607626084533724,
      "tokens_seen": 2573533184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022393939393939395,
      "loss": 2.602,
      "theoretical_loss": 3.3607558214591737,
      "tokens_seen": 2573598720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002239191919191919,
      "loss": 2.7358,
      "theoretical_loss": 3.360749034686193,
      "tokens_seen": 2573664256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022389898989898992,
      "loss": 2.4063,
      "theoretical_loss": 3.3607422481344167,
      "tokens_seen": 2573729792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022387878787878787,
      "loss": 2.508,
      "theoretical_loss": 3.360735461803832,
      "tokens_seen": 2573795328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022385858585858588,
      "loss": 2.5933,
      "theoretical_loss": 3.3607286756944266,
      "tokens_seen": 2573860864
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022383838383838384,
      "loss": 2.6648,
      "theoretical_loss": 3.360721889806187,
      "tokens_seen": 2573926400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002238181818181818,
      "loss": 2.5607,
      "theoretical_loss": 3.360715104139101,
      "tokens_seen": 2573991936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237979797979798,
      "loss": 2.5683,
      "theoretical_loss": 3.3607083186931552,
      "tokens_seen": 2574057472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022377777777777777,
      "loss": 2.5328,
      "theoretical_loss": 3.3607015334683368,
      "tokens_seen": 2574123008
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1410544,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5232367515563965,
      "objective/train/theoretical_loss": 3.360694748464634,
      "objective/train/tokens_used": 933047776,
      "theoretical_loss": 3.360694748464634,
      "tokens_seen": 2574188544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022375757575757578,
      "loss": 2.523,
      "theoretical_loss": 3.360694748464634,
      "tokens_seen": 2574188544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022373737373737373,
      "loss": 2.4974,
      "theoretical_loss": 3.3606879636820324,
      "tokens_seen": 2574254080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022371717171717172,
      "loss": 2.8157,
      "theoretical_loss": 3.3606811791205202,
      "tokens_seen": 2574319616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002236969696969697,
      "loss": 2.736,
      "theoretical_loss": 3.360674394780084,
      "tokens_seen": 2574385152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022367676767676769,
      "loss": 2.4203,
      "theoretical_loss": 3.3606676106607116,
      "tokens_seen": 2574450688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022365656565656567,
      "loss": 2.5239,
      "theoretical_loss": 3.36066082676239,
      "tokens_seen": 2574516224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022363636363636363,
      "loss": 2.2266,
      "theoretical_loss": 3.3606540430851055,
      "tokens_seen": 2574581760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022361616161616164,
      "loss": 2.439,
      "theoretical_loss": 3.3606472596288466,
      "tokens_seen": 2574647296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002235959595959596,
      "loss": 2.5288,
      "theoretical_loss": 3.3606404763935993,
      "tokens_seen": 2574712832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022357575757575758,
      "loss": 2.5587,
      "theoretical_loss": 3.3606336933793517,
      "tokens_seen": 2574778368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022355555555555556,
      "loss": 2.5504,
      "theoretical_loss": 3.3606269105860904,
      "tokens_seen": 2574843904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022353535353535355,
      "loss": 2.57,
      "theoretical_loss": 3.360620128013803,
      "tokens_seen": 2574909440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022351515151515153,
      "loss": 2.6005,
      "theoretical_loss": 3.3606133456624763,
      "tokens_seen": 2574974976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022349494949494951,
      "loss": 2.6424,
      "theoretical_loss": 3.360606563532097,
      "tokens_seen": 2575040512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022347474747474747,
      "loss": 2.526,
      "theoretical_loss": 3.3605997816226534,
      "tokens_seen": 2575106048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022345454545454546,
      "loss": 2.6213,
      "theoretical_loss": 3.3605929999341324,
      "tokens_seen": 2575171584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022343434343434344,
      "loss": 2.5482,
      "theoretical_loss": 3.360586218466521,
      "tokens_seen": 2575237120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022341414141414142,
      "loss": 2.6305,
      "theoretical_loss": 3.3605794372198057,
      "tokens_seen": 2575302656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002233939393939394,
      "loss": 2.691,
      "theoretical_loss": 3.360572656193975,
      "tokens_seen": 2575368192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022337373737373736,
      "loss": 2.3053,
      "theoretical_loss": 3.360565875389015,
      "tokens_seen": 2575433728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022335353535353537,
      "loss": 2.2463,
      "theoretical_loss": 3.3605590948049135,
      "tokens_seen": 2575499264
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022333333333333333,
      "loss": 2.6453,
      "theoretical_loss": 3.3605523144416574,
      "tokens_seen": 2575564800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022331313131313132,
      "loss": 2.722,
      "theoretical_loss": 3.360545534299234,
      "tokens_seen": 2575630336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002232929292929293,
      "loss": 2.782,
      "theoretical_loss": 3.36053875437763,
      "tokens_seen": 2575695872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022327272727272726,
      "loss": 2.558,
      "theoretical_loss": 3.3605319746768334,
      "tokens_seen": 2575761408
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1411184,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.559873342514038,
      "objective/train/theoretical_loss": 3.360525195196831,
      "objective/train/tokens_used": 934686176,
      "theoretical_loss": 3.360525195196831,
      "tokens_seen": 2575826944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022325252525252527,
      "loss": 2.3536,
      "theoretical_loss": 3.360525195196831,
      "tokens_seen": 2575826944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022323232323232322,
      "loss": 2.5187,
      "theoretical_loss": 3.3605184159376105,
      "tokens_seen": 2575892480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022321212121212124,
      "loss": 2.7863,
      "theoretical_loss": 3.360511636899158,
      "tokens_seen": 2575958016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002231919191919192,
      "loss": 2.8348,
      "theoretical_loss": 3.3605048580814616,
      "tokens_seen": 2576023552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022317171717171718,
      "loss": 2.5115,
      "theoretical_loss": 3.360498079484508,
      "tokens_seen": 2576089088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022315151515151516,
      "loss": 2.61,
      "theoretical_loss": 3.3604913011082846,
      "tokens_seen": 2576154624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022313131313131312,
      "loss": 2.6702,
      "theoretical_loss": 3.360484522952779,
      "tokens_seen": 2576220160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022311111111111113,
      "loss": 2.7061,
      "theoretical_loss": 3.3604777450179775,
      "tokens_seen": 2576285696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022309090909090909,
      "loss": 2.5649,
      "theoretical_loss": 3.3604709673038684,
      "tokens_seen": 2576351232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022307070707070707,
      "loss": 2.708,
      "theoretical_loss": 3.360464189810438,
      "tokens_seen": 2576416768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022305050505050505,
      "loss": 2.8263,
      "theoretical_loss": 3.360457412537673,
      "tokens_seen": 2576482304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022303030303030304,
      "loss": 2.5526,
      "theoretical_loss": 3.3604506354855626,
      "tokens_seen": 2576547840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022301010101010102,
      "loss": 2.5622,
      "theoretical_loss": 3.3604438586540923,
      "tokens_seen": 2576613376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000222989898989899,
      "loss": 2.7016,
      "theoretical_loss": 3.3604370820432496,
      "tokens_seen": 2576678912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022296969696969696,
      "loss": 2.5369,
      "theoretical_loss": 3.3604303056530225,
      "tokens_seen": 2576744448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022294949494949495,
      "loss": 2.7631,
      "theoretical_loss": 3.360423529483397,
      "tokens_seen": 2576809984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022292929292929293,
      "loss": 2.5487,
      "theoretical_loss": 3.3604167535343614,
      "tokens_seen": 2576875520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022290909090909091,
      "loss": 2.5263,
      "theoretical_loss": 3.3604099778059022,
      "tokens_seen": 2576941056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002228888888888889,
      "loss": 2.7057,
      "theoretical_loss": 3.360403202298007,
      "tokens_seen": 2577006592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022286868686868685,
      "loss": 2.6283,
      "theoretical_loss": 3.3603964270106625,
      "tokens_seen": 2577072128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022284848484848487,
      "loss": 2.5462,
      "theoretical_loss": 3.3603896519438567,
      "tokens_seen": 2577137664
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022282828282828282,
      "loss": 2.3673,
      "theoretical_loss": 3.3603828770975763,
      "tokens_seen": 2577203200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022280808080808083,
      "loss": 2.7053,
      "theoretical_loss": 3.3603761024718084,
      "tokens_seen": 2577268736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002227878787878788,
      "loss": 2.5939,
      "theoretical_loss": 3.3603693280665405,
      "tokens_seen": 2577334272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022276767676767677,
      "loss": 2.5805,
      "theoretical_loss": 3.3603625538817603,
      "tokens_seen": 2577399808
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1412369,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5855770111083984,
      "objective/train/theoretical_loss": 3.3603557799174535,
      "objective/train/tokens_used": 936324576,
      "theoretical_loss": 3.3603557799174535,
      "tokens_seen": 2577465344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022274747474747476,
      "loss": 2.7335,
      "theoretical_loss": 3.3603557799174535,
      "tokens_seen": 2577465344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022272727272727272,
      "loss": 2.5779,
      "theoretical_loss": 3.360349006173609,
      "tokens_seen": 2577530880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022270707070707073,
      "loss": 2.6307,
      "theoretical_loss": 3.360342232650213,
      "tokens_seen": 2577596416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022268686868686868,
      "loss": 2.393,
      "theoretical_loss": 3.360335459347253,
      "tokens_seen": 2577661952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002226666666666667,
      "loss": 2.7563,
      "theoretical_loss": 3.360328686264716,
      "tokens_seen": 2577727488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022264646464646465,
      "loss": 2.4388,
      "theoretical_loss": 3.36032191340259,
      "tokens_seen": 2577793024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002226262626262626,
      "loss": 2.7172,
      "theoretical_loss": 3.3603151407608616,
      "tokens_seen": 2577858560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022260606060606062,
      "loss": 2.5949,
      "theoretical_loss": 3.360308368339518,
      "tokens_seen": 2577924096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022258585858585858,
      "loss": 2.5263,
      "theoretical_loss": 3.3603015961385463,
      "tokens_seen": 2577989632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002225656565656566,
      "loss": 2.5724,
      "theoretical_loss": 3.360294824157934,
      "tokens_seen": 2578055168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022254545454545454,
      "loss": 2.6198,
      "theoretical_loss": 3.3602880523976686,
      "tokens_seen": 2578120704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022252525252525253,
      "loss": 2.4131,
      "theoretical_loss": 3.3602812808577367,
      "tokens_seen": 2578186240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002225050505050505,
      "loss": 2.4091,
      "theoretical_loss": 3.360274509538126,
      "tokens_seen": 2578251776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224848484848485,
      "loss": 2.4155,
      "theoretical_loss": 3.3602677384388233,
      "tokens_seen": 2578317312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022246464646464648,
      "loss": 2.5602,
      "theoretical_loss": 3.3602609675598165,
      "tokens_seen": 2578382848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022244444444444444,
      "loss": 2.6259,
      "theoretical_loss": 3.3602541969010926,
      "tokens_seen": 2578448384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022242424242424242,
      "loss": 2.5302,
      "theoretical_loss": 3.3602474264626383,
      "tokens_seen": 2578513920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224040404040404,
      "loss": 2.7642,
      "theoretical_loss": 3.3602406562444416,
      "tokens_seen": 2578579456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002223838383838384,
      "loss": 2.4968,
      "theoretical_loss": 3.3602338862464887,
      "tokens_seen": 2578644992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022236363636363637,
      "loss": 2.7516,
      "theoretical_loss": 3.360227116468768,
      "tokens_seen": 2578710528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022234343434343436,
      "loss": 2.3713,
      "theoretical_loss": 3.360220346911266,
      "tokens_seen": 2578776064
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002223232323232323,
      "loss": 2.4961,
      "theoretical_loss": 3.3602135775739708,
      "tokens_seen": 2578841600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022230303030303032,
      "loss": 2.5778,
      "theoretical_loss": 3.3602068084568684,
      "tokens_seen": 2578907136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022228282828282828,
      "loss": 2.5252,
      "theoretical_loss": 3.3602000395599463,
      "tokens_seen": 2578972672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022226262626262626,
      "loss": 2.6237,
      "theoretical_loss": 3.360193270883193,
      "tokens_seen": 2579038208
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1412775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.703599452972412,
      "objective/train/theoretical_loss": 3.3601865024265942,
      "objective/train/tokens_used": 937962976,
      "theoretical_loss": 3.3601865024265942,
      "tokens_seen": 2579103744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022224242424242425,
      "loss": 2.522,
      "theoretical_loss": 3.3601865024265942,
      "tokens_seen": 2579103744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002222222222222222,
      "loss": 2.4549,
      "theoretical_loss": 3.360179734190138,
      "tokens_seen": 2579169280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022220202020202022,
      "loss": 2.5503,
      "theoretical_loss": 3.3601729661738116,
      "tokens_seen": 2579234816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022218181818181817,
      "loss": 2.7587,
      "theoretical_loss": 3.360166198377602,
      "tokens_seen": 2579300352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022216161616161618,
      "loss": 2.5149,
      "theoretical_loss": 3.360159430801496,
      "tokens_seen": 2579365888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022214141414141414,
      "loss": 2.4234,
      "theoretical_loss": 3.360152663445482,
      "tokens_seen": 2579431424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022212121212121213,
      "loss": 2.6957,
      "theoretical_loss": 3.3601458963095463,
      "tokens_seen": 2579496960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002221010101010101,
      "loss": 2.7463,
      "theoretical_loss": 3.3601391293936764,
      "tokens_seen": 2579562496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022208080808080807,
      "loss": 2.7041,
      "theoretical_loss": 3.3601323626978603,
      "tokens_seen": 2579628032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022206060606060608,
      "loss": 2.6664,
      "theoretical_loss": 3.360125596222084,
      "tokens_seen": 2579693568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022204040404040403,
      "loss": 2.3405,
      "theoretical_loss": 3.3601188299663356,
      "tokens_seen": 2579759104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022202020202020202,
      "loss": 2.662,
      "theoretical_loss": 3.360112063930602,
      "tokens_seen": 2579824640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000222,
      "loss": 2.5805,
      "theoretical_loss": 3.3601052981148705,
      "tokens_seen": 2579890176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022197979797979799,
      "loss": 2.7325,
      "theoretical_loss": 3.3600985325191286,
      "tokens_seen": 2579955712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022195959595959597,
      "loss": 2.5754,
      "theoretical_loss": 3.360091767143363,
      "tokens_seen": 2580021248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022193939393939395,
      "loss": 2.6249,
      "theoretical_loss": 3.360085001987562,
      "tokens_seen": 2580086784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022191919191919194,
      "loss": 2.4187,
      "theoretical_loss": 3.3600782370517113,
      "tokens_seen": 2580152320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002218989898989899,
      "loss": 2.671,
      "theoretical_loss": 3.3600714723358,
      "tokens_seen": 2580217856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022187878787878788,
      "loss": 2.4087,
      "theoretical_loss": 3.360064707839814,
      "tokens_seen": 2580283392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022185858585858586,
      "loss": 2.4344,
      "theoretical_loss": 3.3600579435637408,
      "tokens_seen": 2580348928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022183838383838385,
      "loss": 2.5499,
      "theoretical_loss": 3.360051179507568,
      "tokens_seen": 2580414464
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022181818181818183,
      "loss": 2.5075,
      "theoretical_loss": 3.3600444156712825,
      "tokens_seen": 2580480000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022179797979797981,
      "loss": 2.7245,
      "theoretical_loss": 3.360037652054872,
      "tokens_seen": 2580545536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022177777777777777,
      "loss": 2.8689,
      "theoretical_loss": 3.3600308886583234,
      "tokens_seen": 2580611072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022175757575757576,
      "loss": 2.4853,
      "theoretical_loss": 3.3600241254816243,
      "tokens_seen": 2580676608
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1413675,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5682265758514404,
      "objective/train/theoretical_loss": 3.3600173625247622,
      "objective/train/tokens_used": 939601376,
      "theoretical_loss": 3.3600173625247622,
      "tokens_seen": 2580742144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022173737373737374,
      "loss": 2.6656,
      "theoretical_loss": 3.3600173625247622,
      "tokens_seen": 2580742144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022171717171717172,
      "loss": 2.5645,
      "theoretical_loss": 3.3600105997877234,
      "tokens_seen": 2580807680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002216969696969697,
      "loss": 2.5198,
      "theoretical_loss": 3.360003837270496,
      "tokens_seen": 2580873216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022167676767676766,
      "loss": 2.4892,
      "theoretical_loss": 3.3599970749730668,
      "tokens_seen": 2580938752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022165656565656568,
      "loss": 2.2658,
      "theoretical_loss": 3.359990312895423,
      "tokens_seen": 2581004288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022163636363636363,
      "loss": 2.5639,
      "theoretical_loss": 3.359983551037553,
      "tokens_seen": 2581069824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022161616161616164,
      "loss": 2.7028,
      "theoretical_loss": 3.3599767893994428,
      "tokens_seen": 2581135360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002215959595959596,
      "loss": 2.5408,
      "theoretical_loss": 3.3599700279810802,
      "tokens_seen": 2581200896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022157575757575756,
      "loss": 2.5146,
      "theoretical_loss": 3.359963266782452,
      "tokens_seen": 2581266432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022155555555555557,
      "loss": 2.5869,
      "theoretical_loss": 3.359956505803546,
      "tokens_seen": 2581331968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022153535353535352,
      "loss": 2.4845,
      "theoretical_loss": 3.35994974504435,
      "tokens_seen": 2581397504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022151515151515154,
      "loss": 2.9334,
      "theoretical_loss": 3.3599429845048503,
      "tokens_seen": 2581463040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002214949494949495,
      "loss": 2.7303,
      "theoretical_loss": 3.3599362241850343,
      "tokens_seen": 2581528576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022147474747474748,
      "loss": 2.6198,
      "theoretical_loss": 3.35992946408489,
      "tokens_seen": 2581594112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022145454545454546,
      "loss": 2.7762,
      "theoretical_loss": 3.3599227042044033,
      "tokens_seen": 2581659648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022143434343434344,
      "loss": 2.5838,
      "theoretical_loss": 3.3599159445435633,
      "tokens_seen": 2581725184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022141414141414143,
      "loss": 2.5541,
      "theoretical_loss": 3.3599091851023557,
      "tokens_seen": 2581790720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022139393939393939,
      "loss": 2.5901,
      "theoretical_loss": 3.359902425880769,
      "tokens_seen": 2581856256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022137373737373737,
      "loss": 2.6997,
      "theoretical_loss": 3.35989566687879,
      "tokens_seen": 2581921792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022135353535353535,
      "loss": 2.5625,
      "theoretical_loss": 3.3598889080964054,
      "tokens_seen": 2581987328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022133333333333334,
      "loss": 2.5755,
      "theoretical_loss": 3.359882149533603,
      "tokens_seen": 2582052864
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022131313131313132,
      "loss": 2.5887,
      "theoretical_loss": 3.3598753911903705,
      "tokens_seen": 2582118400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002212929292929293,
      "loss": 2.531,
      "theoretical_loss": 3.359868633066695,
      "tokens_seen": 2582183936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022127272727272726,
      "loss": 2.6268,
      "theoretical_loss": 3.3598618751625633,
      "tokens_seen": 2582249472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022125252525252527,
      "loss": 2.6622,
      "theoretical_loss": 3.359855117477963,
      "tokens_seen": 2582315008
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1414348,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.053363561630249,
      "objective/train/theoretical_loss": 3.3598483600128817,
      "objective/train/tokens_used": 941239776,
      "theoretical_loss": 3.3598483600128817,
      "tokens_seen": 2582380544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022123232323232323,
      "loss": 2.4699,
      "theoretical_loss": 3.3598483600128817,
      "tokens_seen": 2582380544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022121212121212121,
      "loss": 2.5715,
      "theoretical_loss": 3.359841602767306,
      "tokens_seen": 2582446080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002211919191919192,
      "loss": 2.3848,
      "theoretical_loss": 3.359834845741224,
      "tokens_seen": 2582511616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022117171717171715,
      "loss": 2.6419,
      "theoretical_loss": 3.3598280889346226,
      "tokens_seen": 2582577152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022115151515151517,
      "loss": 2.2869,
      "theoretical_loss": 3.359821332347489,
      "tokens_seen": 2582642688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022113131313131312,
      "loss": 2.6196,
      "theoretical_loss": 3.3598145759798106,
      "tokens_seen": 2582708224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022111111111111113,
      "loss": 2.592,
      "theoretical_loss": 3.3598078198315746,
      "tokens_seen": 2582773760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002210909090909091,
      "loss": 2.548,
      "theoretical_loss": 3.3598010639027684,
      "tokens_seen": 2582839296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002210707070707071,
      "loss": 2.6218,
      "theoretical_loss": 3.3597943081933797,
      "tokens_seen": 2582904832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022105050505050506,
      "loss": 2.701,
      "theoretical_loss": 3.359787552703395,
      "tokens_seen": 2582970368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022103030303030302,
      "loss": 2.5702,
      "theoretical_loss": 3.3597807974328022,
      "tokens_seen": 2583035904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022101010101010103,
      "loss": 2.5625,
      "theoretical_loss": 3.3597740423815883,
      "tokens_seen": 2583101440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022098989898989898,
      "loss": 2.5406,
      "theoretical_loss": 3.359767287549741,
      "tokens_seen": 2583166976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220969696969697,
      "loss": 2.7338,
      "theoretical_loss": 3.3597605329372473,
      "tokens_seen": 2583232512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022094949494949495,
      "loss": 2.3598,
      "theoretical_loss": 3.3597537785440945,
      "tokens_seen": 2583298048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022092929292929294,
      "loss": 2.7676,
      "theoretical_loss": 3.3597470243702703,
      "tokens_seen": 2583363584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022090909090909092,
      "loss": 2.4892,
      "theoretical_loss": 3.3597402704157613,
      "tokens_seen": 2583429120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022088888888888888,
      "loss": 2.5141,
      "theoretical_loss": 3.3597335166805555,
      "tokens_seen": 2583494656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002208686868686869,
      "loss": 2.6098,
      "theoretical_loss": 3.35972676316464,
      "tokens_seen": 2583560192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022084848484848484,
      "loss": 2.6547,
      "theoretical_loss": 3.359720009868002,
      "tokens_seen": 2583625728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022082828282828283,
      "loss": 2.5265,
      "theoretical_loss": 3.3597132567906285,
      "tokens_seen": 2583691264
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002208080808080808,
      "loss": 2.821,
      "theoretical_loss": 3.3597065039325074,
      "tokens_seen": 2583756800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002207878787878788,
      "loss": 2.5841,
      "theoretical_loss": 3.3596997512936264,
      "tokens_seen": 2583822336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022076767676767678,
      "loss": 2.7251,
      "theoretical_loss": 3.3596929988739714,
      "tokens_seen": 2583887872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022074747474747476,
      "loss": 2.4899,
      "theoretical_loss": 3.359686246673531,
      "tokens_seen": 2583953408
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1415593,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.620751142501831,
      "objective/train/theoretical_loss": 3.3596794946922923,
      "objective/train/tokens_used": 942878176,
      "theoretical_loss": 3.3596794946922923,
      "tokens_seen": 2584018944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022072727272727272,
      "loss": 2.3497,
      "theoretical_loss": 3.3596794946922923,
      "tokens_seen": 2584018944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002207070707070707,
      "loss": 2.5823,
      "theoretical_loss": 3.359672742930242,
      "tokens_seen": 2584084480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002206868686868687,
      "loss": 2.5812,
      "theoretical_loss": 3.359665991387368,
      "tokens_seen": 2584150016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022066666666666667,
      "loss": 2.5143,
      "theoretical_loss": 3.3596592400636576,
      "tokens_seen": 2584215552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022064646464646466,
      "loss": 2.5914,
      "theoretical_loss": 3.3596524889590977,
      "tokens_seen": 2584281088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002206262626262626,
      "loss": 2.6124,
      "theoretical_loss": 3.359645738073676,
      "tokens_seen": 2584346624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022060606060606062,
      "loss": 2.4529,
      "theoretical_loss": 3.3596389874073798,
      "tokens_seen": 2584412160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022058585858585858,
      "loss": 2.4931,
      "theoretical_loss": 3.3596322369601963,
      "tokens_seen": 2584477696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002205656565656566,
      "loss": 2.5786,
      "theoretical_loss": 3.359625486732113,
      "tokens_seen": 2584543232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022054545454545455,
      "loss": 2.5751,
      "theoretical_loss": 3.3596187367231174,
      "tokens_seen": 2584608768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002205252525252525,
      "loss": 2.5049,
      "theoretical_loss": 3.359611986933196,
      "tokens_seen": 2584674304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022050505050505052,
      "loss": 2.5752,
      "theoretical_loss": 3.359605237362337,
      "tokens_seen": 2584739840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022048484848484847,
      "loss": 2.6426,
      "theoretical_loss": 3.3595984880105276,
      "tokens_seen": 2584805376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022046464646464649,
      "loss": 2.5223,
      "theoretical_loss": 3.359591738877755,
      "tokens_seen": 2584870912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022044444444444444,
      "loss": 2.5745,
      "theoretical_loss": 3.3595849899640067,
      "tokens_seen": 2584936448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022042424242424243,
      "loss": 2.5532,
      "theoretical_loss": 3.3595782412692694,
      "tokens_seen": 2585001984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002204040404040404,
      "loss": 2.5212,
      "theoretical_loss": 3.359571492793531,
      "tokens_seen": 2585067520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002203838383838384,
      "loss": 2.423,
      "theoretical_loss": 3.359564744536779,
      "tokens_seen": 2585133056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022036363636363638,
      "loss": 2.565,
      "theoretical_loss": 3.3595579964990003,
      "tokens_seen": 2585198592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022034343434343433,
      "loss": 2.5455,
      "theoretical_loss": 3.359551248680183,
      "tokens_seen": 2585264128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022032323232323232,
      "loss": 2.5565,
      "theoretical_loss": 3.359544501080313,
      "tokens_seen": 2585329664
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002203030303030303,
      "loss": 2.6657,
      "theoretical_loss": 3.359537753699379,
      "tokens_seen": 2585395200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002202828282828283,
      "loss": 2.7,
      "theoretical_loss": 3.3595310065373676,
      "tokens_seen": 2585460736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022026262626262627,
      "loss": 2.399,
      "theoretical_loss": 3.3595242595942665,
      "tokens_seen": 2585526272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022024242424242425,
      "loss": 2.6826,
      "theoretical_loss": 3.359517512870063,
      "tokens_seen": 2585591808
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1416631,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3755147457122803,
      "objective/train/theoretical_loss": 3.3595107663647448,
      "objective/train/tokens_used": 944516576,
      "theoretical_loss": 3.3595107663647448,
      "tokens_seen": 2585657344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022022222222222224,
      "loss": 2.4763,
      "theoretical_loss": 3.3595107663647448,
      "tokens_seen": 2585657344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022020202020202022,
      "loss": 2.5341,
      "theoretical_loss": 3.3595040200782984,
      "tokens_seen": 2585722880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022018181818181818,
      "loss": 2.6848,
      "theoretical_loss": 3.359497274010712,
      "tokens_seen": 2585788416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022016161616161616,
      "loss": 2.5841,
      "theoretical_loss": 3.3594905281619725,
      "tokens_seen": 2585853952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022014141414141415,
      "loss": 2.4764,
      "theoretical_loss": 3.359483782532067,
      "tokens_seen": 2585919488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022012121212121213,
      "loss": 2.534,
      "theoretical_loss": 3.3594770371209837,
      "tokens_seen": 2585985024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022010101010101012,
      "loss": 2.4264,
      "theoretical_loss": 3.359470291928709,
      "tokens_seen": 2586050560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022008080808080807,
      "loss": 2.7399,
      "theoretical_loss": 3.359463546955231,
      "tokens_seen": 2586116096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022006060606060608,
      "loss": 2.3414,
      "theoretical_loss": 3.3594568022005364,
      "tokens_seen": 2586181632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022004040404040404,
      "loss": 2.5416,
      "theoretical_loss": 3.3594500576646134,
      "tokens_seen": 2586247168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022002020202020202,
      "loss": 2.5454,
      "theoretical_loss": 3.3594433133474486,
      "tokens_seen": 2586312704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022,
      "loss": 2.5902,
      "theoretical_loss": 3.35943656924903,
      "tokens_seen": 2586378240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021997979797979796,
      "loss": 2.4311,
      "theoretical_loss": 3.359429825369344,
      "tokens_seen": 2586443776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021995959595959598,
      "loss": 2.4764,
      "theoretical_loss": 3.3594230817083788,
      "tokens_seen": 2586509312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021993939393939393,
      "loss": 2.6504,
      "theoretical_loss": 3.3594163382661217,
      "tokens_seen": 2586574848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021991919191919194,
      "loss": 2.5732,
      "theoretical_loss": 3.3594095950425595,
      "tokens_seen": 2586640384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002198989898989899,
      "loss": 2.5213,
      "theoretical_loss": 3.3594028520376806,
      "tokens_seen": 2586705920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021987878787878788,
      "loss": 2.5253,
      "theoretical_loss": 3.3593961092514713,
      "tokens_seen": 2586771456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021985858585858587,
      "loss": 2.4615,
      "theoretical_loss": 3.3593893666839194,
      "tokens_seen": 2586836992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021983838383838383,
      "loss": 2.498,
      "theoretical_loss": 3.3593826243350127,
      "tokens_seen": 2586902528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021981818181818184,
      "loss": 2.503,
      "theoretical_loss": 3.3593758822047377,
      "tokens_seen": 2586968064
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002197979797979798,
      "loss": 2.459,
      "theoretical_loss": 3.359369140293082,
      "tokens_seen": 2587033600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021977777777777778,
      "loss": 2.5173,
      "theoretical_loss": 3.359362398600034,
      "tokens_seen": 2587099136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021975757575757576,
      "loss": 2.5156,
      "theoretical_loss": 3.3593556571255796,
      "tokens_seen": 2587164672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021973737373737375,
      "loss": 2.3776,
      "theoretical_loss": 3.359348915869707,
      "tokens_seen": 2587230208
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1417164,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6016485691070557,
      "objective/train/theoretical_loss": 3.3593421748324035,
      "objective/train/tokens_used": 946154976,
      "theoretical_loss": 3.3593421748324035,
      "tokens_seen": 2587295744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021971717171717173,
      "loss": 2.6303,
      "theoretical_loss": 3.3593421748324035,
      "tokens_seen": 2587295744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002196969696969697,
      "loss": 2.5608,
      "theoretical_loss": 3.3593354340136563,
      "tokens_seen": 2587361280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021967676767676767,
      "loss": 2.4379,
      "theoretical_loss": 3.359328693413453,
      "tokens_seen": 2587426816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021965656565656565,
      "loss": 2.3855,
      "theoretical_loss": 3.3593219530317806,
      "tokens_seen": 2587492352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021963636363636364,
      "loss": 2.6761,
      "theoretical_loss": 3.359315212868627,
      "tokens_seen": 2587557888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021961616161616162,
      "loss": 2.4719,
      "theoretical_loss": 3.3593084729239795,
      "tokens_seen": 2587623424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002195959595959596,
      "loss": 2.5267,
      "theoretical_loss": 3.359301733197825,
      "tokens_seen": 2587688960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021957575757575756,
      "loss": 2.564,
      "theoretical_loss": 3.359294993690151,
      "tokens_seen": 2587754496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021955555555555557,
      "loss": 2.5549,
      "theoretical_loss": 3.3592882544009455,
      "tokens_seen": 2587820032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021953535353535353,
      "loss": 2.3944,
      "theoretical_loss": 3.359281515330195,
      "tokens_seen": 2587885568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021951515151515154,
      "loss": 2.465,
      "theoretical_loss": 3.3592747764778874,
      "tokens_seen": 2587951104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194949494949495,
      "loss": 2.6697,
      "theoretical_loss": 3.3592680378440103,
      "tokens_seen": 2588016640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021947474747474746,
      "loss": 2.5515,
      "theoretical_loss": 3.3592612994285505,
      "tokens_seen": 2588082176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021945454545454547,
      "loss": 2.4575,
      "theoretical_loss": 3.359254561231496,
      "tokens_seen": 2588147712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021943434343434342,
      "loss": 2.6306,
      "theoretical_loss": 3.3592478232528338,
      "tokens_seen": 2588213248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021941414141414143,
      "loss": 2.6545,
      "theoretical_loss": 3.3592410854925516,
      "tokens_seen": 2588278784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002193939393939394,
      "loss": 2.6376,
      "theoretical_loss": 3.359234347950636,
      "tokens_seen": 2588344320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002193737373737374,
      "loss": 2.556,
      "theoretical_loss": 3.3592276106270758,
      "tokens_seen": 2588409856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021935353535353536,
      "loss": 2.6016,
      "theoretical_loss": 3.3592208735218567,
      "tokens_seen": 2588475392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021933333333333332,
      "loss": 2.7429,
      "theoretical_loss": 3.3592141366349675,
      "tokens_seen": 2588540928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021931313131313133,
      "loss": 2.4559,
      "theoretical_loss": 3.3592073999663947,
      "tokens_seen": 2588606464
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021929292929292928,
      "loss": 2.5043,
      "theoretical_loss": 3.3592006635161265,
      "tokens_seen": 2588672000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002192727272727273,
      "loss": 2.4034,
      "theoretical_loss": 3.3591939272841493,
      "tokens_seen": 2588737536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021925252525252525,
      "loss": 2.4723,
      "theoretical_loss": 3.359187191270452,
      "tokens_seen": 2588803072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021923232323232324,
      "loss": 2.7292,
      "theoretical_loss": 3.35918045547502,
      "tokens_seen": 2588868608
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1417967,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.8440449237823486,
      "objective/train/theoretical_loss": 3.3591737198978424,
      "objective/train/tokens_used": 947793376,
      "theoretical_loss": 3.3591737198978424,
      "tokens_seen": 2588934144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021921212121212122,
      "loss": 2.4915,
      "theoretical_loss": 3.3591737198978424,
      "tokens_seen": 2588934144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191919191919192,
      "loss": 2.4884,
      "theoretical_loss": 3.359166984538906,
      "tokens_seen": 2588999680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191717171717172,
      "loss": 2.5033,
      "theoretical_loss": 3.359160249398198,
      "tokens_seen": 2589065216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021915151515151514,
      "loss": 2.6362,
      "theoretical_loss": 3.3591535144757056,
      "tokens_seen": 2589130752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021913131313131313,
      "loss": 2.4171,
      "theoretical_loss": 3.359146779771417,
      "tokens_seen": 2589196288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191111111111111,
      "loss": 2.4259,
      "theoretical_loss": 3.3591400452853195,
      "tokens_seen": 2589261824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002190909090909091,
      "loss": 2.5817,
      "theoretical_loss": 3.3591333110173998,
      "tokens_seen": 2589327360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021907070707070708,
      "loss": 2.7101,
      "theoretical_loss": 3.3591265769676455,
      "tokens_seen": 2589392896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021905050505050506,
      "loss": 2.4445,
      "theoretical_loss": 3.3591198431360443,
      "tokens_seen": 2589458432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021903030303030302,
      "loss": 2.4888,
      "theoretical_loss": 3.359113109522584,
      "tokens_seen": 2589523968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021901010101010103,
      "loss": 2.6211,
      "theoretical_loss": 3.3591063761272513,
      "tokens_seen": 2589589504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218989898989899,
      "loss": 2.6563,
      "theoretical_loss": 3.3590996429500337,
      "tokens_seen": 2589655040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021896969696969697,
      "loss": 2.6382,
      "theoretical_loss": 3.3590929099909186,
      "tokens_seen": 2589720576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021894949494949496,
      "loss": 2.3845,
      "theoretical_loss": 3.3590861772498943,
      "tokens_seen": 2589786112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021892929292929291,
      "loss": 2.5932,
      "theoretical_loss": 3.359079444726947,
      "tokens_seen": 2589851648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021890909090909092,
      "loss": 2.4212,
      "theoretical_loss": 3.3590727124220647,
      "tokens_seen": 2589917184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021888888888888888,
      "loss": 2.3237,
      "theoretical_loss": 3.3590659803352345,
      "tokens_seen": 2589982720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002188686868686869,
      "loss": 2.739,
      "theoretical_loss": 3.3590592484664445,
      "tokens_seen": 2590048256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021884848484848485,
      "loss": 2.6274,
      "theoretical_loss": 3.3590525168156815,
      "tokens_seen": 2590113792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021882828282828283,
      "loss": 2.5006,
      "theoretical_loss": 3.359045785382933,
      "tokens_seen": 2590179328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021880808080808082,
      "loss": 2.4492,
      "theoretical_loss": 3.359039054168187,
      "tokens_seen": 2590244864
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021878787878787877,
      "loss": 2.5897,
      "theoretical_loss": 3.35903232317143,
      "tokens_seen": 2590310400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021876767676767679,
      "loss": 2.59,
      "theoretical_loss": 3.35902559239265,
      "tokens_seen": 2590375936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021874747474747474,
      "loss": 2.295,
      "theoretical_loss": 3.359018861831834,
      "tokens_seen": 2590441472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021872727272727273,
      "loss": 2.4984,
      "theoretical_loss": 3.3590121314889703,
      "tokens_seen": 2590507008
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1418580,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.70296573638916,
      "objective/train/theoretical_loss": 3.359005401364046,
      "objective/train/tokens_used": 949431776,
      "theoretical_loss": 3.359005401364046,
      "tokens_seen": 2590572544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002187070707070707,
      "loss": 2.5571,
      "theoretical_loss": 3.359005401364046,
      "tokens_seen": 2590572544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186868686868687,
      "loss": 2.4321,
      "theoretical_loss": 3.3589986714570474,
      "tokens_seen": 2590638080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021866666666666668,
      "loss": 2.5474,
      "theoretical_loss": 3.3589919417679632,
      "tokens_seen": 2590703616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021864646464646466,
      "loss": 2.4746,
      "theoretical_loss": 3.3589852122967807,
      "tokens_seen": 2590769152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021862626262626262,
      "loss": 2.6257,
      "theoretical_loss": 3.3589784830434866,
      "tokens_seen": 2590834688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186060606060606,
      "loss": 2.6667,
      "theoretical_loss": 3.3589717540080692,
      "tokens_seen": 2590900224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002185858585858586,
      "loss": 2.5003,
      "theoretical_loss": 3.3589650251905154,
      "tokens_seen": 2590965760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021856565656565657,
      "loss": 2.4967,
      "theoretical_loss": 3.358958296590813,
      "tokens_seen": 2591031296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021854545454545455,
      "loss": 2.4113,
      "theoretical_loss": 3.358951568208949,
      "tokens_seen": 2591096832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021852525252525254,
      "loss": 2.4337,
      "theoretical_loss": 3.358944840044911,
      "tokens_seen": 2591162368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021850505050505052,
      "loss": 2.486,
      "theoretical_loss": 3.358938112098687,
      "tokens_seen": 2591227904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021848484848484848,
      "loss": 2.5468,
      "theoretical_loss": 3.3589313843702637,
      "tokens_seen": 2591293440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021846464646464646,
      "loss": 2.5335,
      "theoretical_loss": 3.358924656859629,
      "tokens_seen": 2591358976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021844444444444445,
      "loss": 2.5353,
      "theoretical_loss": 3.3589179295667697,
      "tokens_seen": 2591424512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021842424242424243,
      "loss": 2.4998,
      "theoretical_loss": 3.3589112024916736,
      "tokens_seen": 2591490048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021840404040404042,
      "loss": 2.5761,
      "theoretical_loss": 3.3589044756343287,
      "tokens_seen": 2591555584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021838383838383837,
      "loss": 2.5652,
      "theoretical_loss": 3.3588977489947216,
      "tokens_seen": 2591621120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021836363636363638,
      "loss": 2.5676,
      "theoretical_loss": 3.3588910225728403,
      "tokens_seen": 2591686656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021834343434343434,
      "loss": 2.5555,
      "theoretical_loss": 3.358884296368672,
      "tokens_seen": 2591752192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021832323232323235,
      "loss": 2.4154,
      "theoretical_loss": 3.3588775703822042,
      "tokens_seen": 2591817728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002183030303030303,
      "loss": 2.6286,
      "theoretical_loss": 3.358870844613424,
      "tokens_seen": 2591883264
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021828282828282827,
      "loss": 2.3768,
      "theoretical_loss": 3.35886411906232,
      "tokens_seen": 2591948800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021826262626262628,
      "loss": 2.5213,
      "theoretical_loss": 3.3588573937288784,
      "tokens_seen": 2592014336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021824242424242423,
      "loss": 2.4374,
      "theoretical_loss": 3.358850668613087,
      "tokens_seen": 2592079872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021822222222222224,
      "loss": 2.6197,
      "theoretical_loss": 3.3588439437149336,
      "tokens_seen": 2592145408
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1419695,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1405866146087646,
      "objective/train/theoretical_loss": 3.3588372190344056,
      "objective/train/tokens_used": 951070176,
      "theoretical_loss": 3.3588372190344056,
      "tokens_seen": 2592210944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002182020202020202,
      "loss": 2.2491,
      "theoretical_loss": 3.3588372190344056,
      "tokens_seen": 2592210944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021818181818181818,
      "loss": 2.3255,
      "theoretical_loss": 3.3588304945714897,
      "tokens_seen": 2592276480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021816161616161617,
      "loss": 2.5569,
      "theoretical_loss": 3.3588237703261745,
      "tokens_seen": 2592342016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021814141414141415,
      "loss": 2.4683,
      "theoretical_loss": 3.3588170462984466,
      "tokens_seen": 2592407552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021812121212121214,
      "loss": 2.5846,
      "theoretical_loss": 3.358810322488294,
      "tokens_seen": 2592473088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002181010101010101,
      "loss": 2.3852,
      "theoretical_loss": 3.3588035988957032,
      "tokens_seen": 2592538624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021808080808080808,
      "loss": 2.4947,
      "theoretical_loss": 3.358796875520663,
      "tokens_seen": 2592604160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021806060606060606,
      "loss": 2.5856,
      "theoretical_loss": 3.3587901523631603,
      "tokens_seen": 2592669696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021804040404040405,
      "loss": 2.6675,
      "theoretical_loss": 3.358783429423182,
      "tokens_seen": 2592735232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021802020202020203,
      "loss": 2.5032,
      "theoretical_loss": 3.3587767067007164,
      "tokens_seen": 2592800768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218,
      "loss": 2.469,
      "theoretical_loss": 3.358769984195751,
      "tokens_seen": 2592866304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021797979797979797,
      "loss": 2.3924,
      "theoretical_loss": 3.358763261908272,
      "tokens_seen": 2592931840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021795959595959598,
      "loss": 2.3487,
      "theoretical_loss": 3.3587565398382684,
      "tokens_seen": 2592997376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021793939393939394,
      "loss": 2.4935,
      "theoretical_loss": 3.3587498179857267,
      "tokens_seen": 2593062912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021791919191919192,
      "loss": 2.5652,
      "theoretical_loss": 3.358743096350635,
      "tokens_seen": 2593128448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002178989898989899,
      "loss": 2.572,
      "theoretical_loss": 3.35873637493298,
      "tokens_seen": 2593193984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021787878787878786,
      "loss": 2.1529,
      "theoretical_loss": 3.3587296537327505,
      "tokens_seen": 2593259520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021785858585858587,
      "loss": 2.5112,
      "theoretical_loss": 3.3587229327499326,
      "tokens_seen": 2593325056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021783838383838383,
      "loss": 2.5555,
      "theoretical_loss": 3.358716211984514,
      "tokens_seen": 2593390592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021781818181818184,
      "loss": 2.6006,
      "theoretical_loss": 3.358709491436483,
      "tokens_seen": 2593456128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002177979797979798,
      "loss": 2.3527,
      "theoretical_loss": 3.358702771105826,
      "tokens_seen": 2593521664
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021777777777777776,
      "loss": 2.506,
      "theoretical_loss": 3.358696050992531,
      "tokens_seen": 2593587200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021775757575757577,
      "loss": 2.3082,
      "theoretical_loss": 3.3586893310965857,
      "tokens_seen": 2593652736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021773737373737372,
      "loss": 2.488,
      "theoretical_loss": 3.3586826114179775,
      "tokens_seen": 2593718272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021771717171717173,
      "loss": 2.4448,
      "theoretical_loss": 3.3586758919566937,
      "tokens_seen": 2593783808
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1420179,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4729344844818115,
      "objective/train/theoretical_loss": 3.358669172712722,
      "objective/train/tokens_used": 952708576,
      "theoretical_loss": 3.358669172712722,
      "tokens_seen": 2593849344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002176969696969697,
      "loss": 2.3976,
      "theoretical_loss": 3.358669172712722,
      "tokens_seen": 2593849344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021767676767676768,
      "loss": 2.2915,
      "theoretical_loss": 3.3586624536860494,
      "tokens_seen": 2593914880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021765656565656566,
      "loss": 2.4951,
      "theoretical_loss": 3.3586557348766637,
      "tokens_seen": 2593980416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021763636363636364,
      "loss": 2.359,
      "theoretical_loss": 3.358649016284552,
      "tokens_seen": 2594045952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021761616161616163,
      "loss": 2.4568,
      "theoretical_loss": 3.358642297909703,
      "tokens_seen": 2594111488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021761616161616163,
      "loss": 2.2921,
      "theoretical_loss": 3.358635579752103,
      "tokens_seen": 2594177024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021759595959595958,
      "loss": 2.6008,
      "theoretical_loss": 3.3586288618117397,
      "tokens_seen": 2594242560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002175757575757576,
      "loss": 2.396,
      "theoretical_loss": 3.3586221440886006,
      "tokens_seen": 2594308096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021755555555555555,
      "loss": 2.5289,
      "theoretical_loss": 3.3586154265826735,
      "tokens_seen": 2594373632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021753535353535354,
      "loss": 2.4621,
      "theoretical_loss": 3.3586087092939456,
      "tokens_seen": 2594439168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021751515151515152,
      "loss": 2.4745,
      "theoretical_loss": 3.358601992222405,
      "tokens_seen": 2594504704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174949494949495,
      "loss": 2.6116,
      "theoretical_loss": 3.358595275368038,
      "tokens_seen": 2594570240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174747474747475,
      "loss": 2.7344,
      "theoretical_loss": 3.3585885587308333,
      "tokens_seen": 2594635776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021745454545454547,
      "loss": 2.5481,
      "theoretical_loss": 3.3585818423107776,
      "tokens_seen": 2594701312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021743434343434343,
      "loss": 2.5348,
      "theoretical_loss": 3.3585751261078585,
      "tokens_seen": 2594766848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174141414141414,
      "loss": 2.6566,
      "theoretical_loss": 3.358568410122064,
      "tokens_seen": 2594832384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002173939393939394,
      "loss": 2.619,
      "theoretical_loss": 3.358561694353381,
      "tokens_seen": 2594897920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021737373737373738,
      "loss": 2.3552,
      "theoretical_loss": 3.358554978801797,
      "tokens_seen": 2594963456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021735353535353536,
      "loss": 2.539,
      "theoretical_loss": 3.3585482634673003,
      "tokens_seen": 2595028992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021733333333333332,
      "loss": 2.437,
      "theoretical_loss": 3.3585415483498777,
      "tokens_seen": 2595094528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021731313131313133,
      "loss": 2.4495,
      "theoretical_loss": 3.3585348334495166,
      "tokens_seen": 2595160064
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002172929292929293,
      "loss": 2.3601,
      "theoretical_loss": 3.358528118766205,
      "tokens_seen": 2595225600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002172727272727273,
      "loss": 2.6459,
      "theoretical_loss": 3.35852140429993,
      "tokens_seen": 2595291136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021725252525252526,
      "loss": 2.3839,
      "theoretical_loss": 3.3585146900506793,
      "tokens_seen": 2595356672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021723232323232321,
      "loss": 2.2672,
      "theoretical_loss": 3.3585079760184406,
      "tokens_seen": 2595422208
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1421066,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4548287391662598,
      "objective/train/theoretical_loss": 3.358501262203201,
      "objective/train/tokens_used": 954346976,
      "theoretical_loss": 3.358501262203201,
      "tokens_seen": 2595487744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021721212121212123,
      "loss": 2.596,
      "theoretical_loss": 3.358501262203201,
      "tokens_seen": 2595487744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021719191919191918,
      "loss": 2.5514,
      "theoretical_loss": 3.3584945486049484,
      "tokens_seen": 2595553280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002171717171717172,
      "loss": 2.3528,
      "theoretical_loss": 3.35848783522367,
      "tokens_seen": 2595618816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021715151515151515,
      "loss": 2.3445,
      "theoretical_loss": 3.358481122059353,
      "tokens_seen": 2595684352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021713131313131313,
      "loss": 2.563,
      "theoretical_loss": 3.358474409111986,
      "tokens_seen": 2595749888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021711111111111112,
      "loss": 2.6417,
      "theoretical_loss": 3.358467696381555,
      "tokens_seen": 2595815424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002170909090909091,
      "loss": 2.5867,
      "theoretical_loss": 3.358460983868049,
      "tokens_seen": 2595880960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021707070707070709,
      "loss": 2.4192,
      "theoretical_loss": 3.3584542715714543,
      "tokens_seen": 2595946496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021705050505050504,
      "loss": 2.6282,
      "theoretical_loss": 3.3584475594917595,
      "tokens_seen": 2596012032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021703030303030303,
      "loss": 2.7238,
      "theoretical_loss": 3.3584408476289513,
      "tokens_seen": 2596077568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000217010101010101,
      "loss": 2.6259,
      "theoretical_loss": 3.3584341359830177,
      "tokens_seen": 2596143104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000216989898989899,
      "loss": 2.3881,
      "theoretical_loss": 3.358427424553946,
      "tokens_seen": 2596208640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021696969696969698,
      "loss": 2.4984,
      "theoretical_loss": 3.3584207133417237,
      "tokens_seen": 2596274176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021694949494949496,
      "loss": 2.5496,
      "theoretical_loss": 3.358414002346338,
      "tokens_seen": 2596339712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021692929292929292,
      "loss": 2.5617,
      "theoretical_loss": 3.358407291567777,
      "tokens_seen": 2596405248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021690909090909093,
      "loss": 2.5783,
      "theoretical_loss": 3.3584005810060282,
      "tokens_seen": 2596470784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002168888888888889,
      "loss": 2.6994,
      "theoretical_loss": 3.3583938706610788,
      "tokens_seen": 2596536320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021686868686868687,
      "loss": 2.5369,
      "theoretical_loss": 3.3583871605329163,
      "tokens_seen": 2596601856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021684848484848486,
      "loss": 2.5248,
      "theoretical_loss": 3.358380450621529,
      "tokens_seen": 2596667392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002168282828282828,
      "loss": 2.7868,
      "theoretical_loss": 3.358373740926903,
      "tokens_seen": 2596732928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021680808080808082,
      "loss": 2.6456,
      "theoretical_loss": 3.358367031449027,
      "tokens_seen": 2596798464
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021678787878787878,
      "loss": 2.5403,
      "theoretical_loss": 3.358360322187888,
      "tokens_seen": 2596864000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002167676767676768,
      "loss": 2.2078,
      "theoretical_loss": 3.358353613143474,
      "tokens_seen": 2596929536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021674747474747475,
      "loss": 2.5189,
      "theoretical_loss": 3.3583469043157717,
      "tokens_seen": 2596995072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021672727272727273,
      "loss": 2.5678,
      "theoretical_loss": 3.35834019570477,
      "tokens_seen": 2597060608
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1421641,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6630356311798096,
      "objective/train/theoretical_loss": 3.3583334873104547,
      "objective/train/tokens_used": 955985376,
      "theoretical_loss": 3.3583334873104547,
      "tokens_seen": 2597126144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021670707070707072,
      "loss": 2.5555,
      "theoretical_loss": 3.3583334873104547,
      "tokens_seen": 2597126144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021668686868686867,
      "loss": 2.2591,
      "theoretical_loss": 3.358326779132814,
      "tokens_seen": 2597191680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021666666666666668,
      "loss": 2.4252,
      "theoretical_loss": 3.3583200711718364,
      "tokens_seen": 2597257216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021664646464646464,
      "loss": 2.4429,
      "theoretical_loss": 3.3583133634275084,
      "tokens_seen": 2597322752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021662626262626265,
      "loss": 2.5285,
      "theoretical_loss": 3.358306655899818,
      "tokens_seen": 2597388288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002166060606060606,
      "loss": 2.6184,
      "theoretical_loss": 3.3582999485887526,
      "tokens_seen": 2597453824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002165858585858586,
      "loss": 2.6832,
      "theoretical_loss": 3.3582932414942994,
      "tokens_seen": 2597519360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021656565656565658,
      "loss": 2.8435,
      "theoretical_loss": 3.3582865346164463,
      "tokens_seen": 2597584896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021654545454545453,
      "loss": 2.4983,
      "theoretical_loss": 3.358279827955181,
      "tokens_seen": 2597650432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021652525252525254,
      "loss": 2.4359,
      "theoretical_loss": 3.35827312151049,
      "tokens_seen": 2597715968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002165050505050505,
      "loss": 2.5759,
      "theoretical_loss": 3.3582664152823627,
      "tokens_seen": 2597781504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021648484848484849,
      "loss": 2.4554,
      "theoretical_loss": 3.358259709270785,
      "tokens_seen": 2597847040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021646464646464647,
      "loss": 2.6763,
      "theoretical_loss": 3.3582530034757454,
      "tokens_seen": 2597912576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021644444444444445,
      "loss": 2.5601,
      "theoretical_loss": 3.358246297897231,
      "tokens_seen": 2597978112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021642424242424244,
      "loss": 2.7189,
      "theoretical_loss": 3.3582395925352295,
      "tokens_seen": 2598043648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021640404040404042,
      "loss": 2.5977,
      "theoretical_loss": 3.358232887389728,
      "tokens_seen": 2598109184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021638383838383838,
      "loss": 2.6685,
      "theoretical_loss": 3.3582261824607147,
      "tokens_seen": 2598174720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021636363636363636,
      "loss": 2.4319,
      "theoretical_loss": 3.358219477748177,
      "tokens_seen": 2598240256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021634343434343435,
      "loss": 2.5106,
      "theoretical_loss": 3.3582127732521023,
      "tokens_seen": 2598305792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021632323232323233,
      "loss": 2.4512,
      "theoretical_loss": 3.358206068972478,
      "tokens_seen": 2598371328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021630303030303031,
      "loss": 2.4567,
      "theoretical_loss": 3.358199364909292,
      "tokens_seen": 2598436864
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021628282828282827,
      "loss": 2.5224,
      "theoretical_loss": 3.3581926610625317,
      "tokens_seen": 2598502400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021626262626262628,
      "loss": 2.276,
      "theoretical_loss": 3.3581859574321844,
      "tokens_seen": 2598567936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021624242424242424,
      "loss": 2.7024,
      "theoretical_loss": 3.3581792540182382,
      "tokens_seen": 2598633472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021622222222222225,
      "loss": 2.5148,
      "theoretical_loss": 3.35817255082068,
      "tokens_seen": 2598699008
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1422678,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7150049209594727,
      "objective/train/theoretical_loss": 3.358165847839498,
      "objective/train/tokens_used": 957623776,
      "theoretical_loss": 3.358165847839498,
      "tokens_seen": 2598764544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002162020202020202,
      "loss": 2.4448,
      "theoretical_loss": 3.358165847839498,
      "tokens_seen": 2598764544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021618181818181816,
      "loss": 2.466,
      "theoretical_loss": 3.3581591450746795,
      "tokens_seen": 2598830080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021616161616161617,
      "loss": 2.6349,
      "theoretical_loss": 3.3581524425262117,
      "tokens_seen": 2598895616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021614141414141413,
      "loss": 2.4254,
      "theoretical_loss": 3.358145740194083,
      "tokens_seen": 2598961152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021612121212121214,
      "loss": 2.4,
      "theoretical_loss": 3.3581390380782805,
      "tokens_seen": 2599026688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002161010101010101,
      "loss": 2.5153,
      "theoretical_loss": 3.3581323361787914,
      "tokens_seen": 2599092224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021608080808080808,
      "loss": 2.4718,
      "theoretical_loss": 3.3581256344956034,
      "tokens_seen": 2599157760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021606060606060607,
      "loss": 2.5701,
      "theoretical_loss": 3.3581189330287047,
      "tokens_seen": 2599223296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021604040404040402,
      "loss": 2.4253,
      "theoretical_loss": 3.3581122317780823,
      "tokens_seen": 2599288832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021602020202020204,
      "loss": 2.491,
      "theoretical_loss": 3.3581055307437238,
      "tokens_seen": 2599354368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000216,
      "loss": 2.6047,
      "theoretical_loss": 3.3580988299256167,
      "tokens_seen": 2599419904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021597979797979798,
      "loss": 2.3913,
      "theoretical_loss": 3.358092129323749,
      "tokens_seen": 2599485440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021595959595959596,
      "loss": 2.5345,
      "theoretical_loss": 3.358085428938108,
      "tokens_seen": 2599550976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021593939393939394,
      "loss": 2.5214,
      "theoretical_loss": 3.358078728768681,
      "tokens_seen": 2599616512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021591919191919193,
      "loss": 2.6337,
      "theoretical_loss": 3.358072028815456,
      "tokens_seen": 2599682048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002158989898989899,
      "loss": 2.574,
      "theoretical_loss": 3.358065329078421,
      "tokens_seen": 2599747584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002158787878787879,
      "loss": 2.6739,
      "theoretical_loss": 3.3580586295575623,
      "tokens_seen": 2599813120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021585858585858585,
      "loss": 2.5347,
      "theoretical_loss": 3.358051930252868,
      "tokens_seen": 2599878656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021583838383838384,
      "loss": 2.5135,
      "theoretical_loss": 3.3580452311643265,
      "tokens_seen": 2599944192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021581818181818182,
      "loss": 2.0731,
      "theoretical_loss": 3.358038532291924,
      "tokens_seen": 2600009728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157979797979798,
      "loss": 2.5628,
      "theoretical_loss": 3.3580318336356494,
      "tokens_seen": 2600075264
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157777777777778,
      "loss": 2.6217,
      "theoretical_loss": 3.358025135195489,
      "tokens_seen": 2600140800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021575757575757577,
      "loss": 2.5824,
      "theoretical_loss": 3.358018436971432,
      "tokens_seen": 2600206336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021573737373737373,
      "loss": 2.4306,
      "theoretical_loss": 3.3580117389634645,
      "tokens_seen": 2600271872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021571717171717174,
      "loss": 2.4561,
      "theoretical_loss": 3.3580050411715745,
      "tokens_seen": 2600337408
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1423150,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.528973340988159,
      "objective/train/theoretical_loss": 3.35799834359575,
      "objective/train/tokens_used": 959262176,
      "theoretical_loss": 3.35799834359575,
      "tokens_seen": 2600402944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002156969696969697,
      "loss": 2.4335,
      "theoretical_loss": 3.35799834359575,
      "tokens_seen": 2600402944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021567676767676768,
      "loss": 2.4746,
      "theoretical_loss": 3.3579916462359782,
      "tokens_seen": 2600468480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021565656565656567,
      "loss": 2.3377,
      "theoretical_loss": 3.3579849490922467,
      "tokens_seen": 2600534016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021563636363636362,
      "loss": 2.3646,
      "theoretical_loss": 3.357978252164543,
      "tokens_seen": 2600599552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021561616161616163,
      "loss": 2.4497,
      "theoretical_loss": 3.3579715554528553,
      "tokens_seen": 2600665088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155959595959596,
      "loss": 2.7239,
      "theoretical_loss": 3.3579648589571707,
      "tokens_seen": 2600730624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155757575757576,
      "loss": 2.3426,
      "theoretical_loss": 3.3579581626774764,
      "tokens_seen": 2600796160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021555555555555556,
      "loss": 2.4124,
      "theoretical_loss": 3.357951466613761,
      "tokens_seen": 2600861696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021553535353535354,
      "loss": 2.6073,
      "theoretical_loss": 3.357944770766011,
      "tokens_seen": 2600927232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021551515151515153,
      "loss": 2.3444,
      "theoretical_loss": 3.3579380751342147,
      "tokens_seen": 2600992768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021549494949494948,
      "loss": 2.4205,
      "theoretical_loss": 3.3579313797183596,
      "tokens_seen": 2601058304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002154747474747475,
      "loss": 2.6766,
      "theoretical_loss": 3.3579246845184327,
      "tokens_seen": 2601123840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021545454545454545,
      "loss": 2.4486,
      "theoretical_loss": 3.357917989534423,
      "tokens_seen": 2601189376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021543434343434343,
      "loss": 2.4147,
      "theoretical_loss": 3.3579112947663163,
      "tokens_seen": 2601254912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021541414141414142,
      "loss": 2.6322,
      "theoretical_loss": 3.3579046002141015,
      "tokens_seen": 2601320448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153939393939394,
      "loss": 2.5092,
      "theoretical_loss": 3.357897905877766,
      "tokens_seen": 2601385984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021537373737373739,
      "loss": 2.3349,
      "theoretical_loss": 3.3578912117572965,
      "tokens_seen": 2601451520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021535353535353537,
      "loss": 2.51,
      "theoretical_loss": 3.357884517852682,
      "tokens_seen": 2601517056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021533333333333333,
      "loss": 2.3547,
      "theoretical_loss": 3.3578778241639093,
      "tokens_seen": 2601582592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153131313131313,
      "loss": 2.62,
      "theoretical_loss": 3.357871130690966,
      "tokens_seen": 2601648128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002152929292929293,
      "loss": 2.6658,
      "theoretical_loss": 3.3578644374338396,
      "tokens_seen": 2601713664
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021527272727272728,
      "loss": 2.6041,
      "theoretical_loss": 3.357857744392518,
      "tokens_seen": 2601779200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021525252525252526,
      "loss": 2.3339,
      "theoretical_loss": 3.357851051566989,
      "tokens_seen": 2601844736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021523232323232322,
      "loss": 2.6773,
      "theoretical_loss": 3.3578443589572395,
      "tokens_seen": 2601910272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021521212121212123,
      "loss": 2.6162,
      "theoretical_loss": 3.357837666563258,
      "tokens_seen": 2601975808
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1424283,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8036410808563232,
      "objective/train/theoretical_loss": 3.357830974385031,
      "objective/train/tokens_used": 960900576,
      "theoretical_loss": 3.357830974385031,
      "tokens_seen": 2602041344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002151919191919192,
      "loss": 2.629,
      "theoretical_loss": 3.357830974385031,
      "tokens_seen": 2602041344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021517171717171717,
      "loss": 2.6666,
      "theoretical_loss": 3.357824282422547,
      "tokens_seen": 2602106880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021515151515151516,
      "loss": 2.5084,
      "theoretical_loss": 3.3578175906757934,
      "tokens_seen": 2602172416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002151313131313131,
      "loss": 2.473,
      "theoretical_loss": 3.357810899144758,
      "tokens_seen": 2602237952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021511111111111112,
      "loss": 2.5214,
      "theoretical_loss": 3.357804207829428,
      "tokens_seen": 2602303488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021509090909090908,
      "loss": 2.6761,
      "theoretical_loss": 3.357797516729791,
      "tokens_seen": 2602369024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002150707070707071,
      "loss": 2.4958,
      "theoretical_loss": 3.3577908258458353,
      "tokens_seen": 2602434560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021505050505050505,
      "loss": 2.5877,
      "theoretical_loss": 3.357784135177548,
      "tokens_seen": 2602500096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021503030303030306,
      "loss": 2.4545,
      "theoretical_loss": 3.357777444724916,
      "tokens_seen": 2602565632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021501010101010102,
      "loss": 2.562,
      "theoretical_loss": 3.3577707544879285,
      "tokens_seen": 2602631168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021498989898989897,
      "loss": 2.4617,
      "theoretical_loss": 3.3577640644665716,
      "tokens_seen": 2602696704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021496969696969698,
      "loss": 2.3442,
      "theoretical_loss": 3.3577573746608342,
      "tokens_seen": 2602762240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021494949494949494,
      "loss": 2.4284,
      "theoretical_loss": 3.357750685070703,
      "tokens_seen": 2602827776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021492929292929295,
      "loss": 2.4988,
      "theoretical_loss": 3.357743995696166,
      "tokens_seen": 2602893312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149090909090909,
      "loss": 2.3649,
      "theoretical_loss": 3.357737306537211,
      "tokens_seen": 2602958848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148888888888889,
      "loss": 2.4496,
      "theoretical_loss": 3.357730617593825,
      "tokens_seen": 2603024384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021486868686868688,
      "loss": 2.6176,
      "theoretical_loss": 3.3577239288659966,
      "tokens_seen": 2603089920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021484848484848486,
      "loss": 2.4863,
      "theoretical_loss": 3.357717240353712,
      "tokens_seen": 2603155456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021482828282828284,
      "loss": 2.4509,
      "theoretical_loss": 3.3577105520569606,
      "tokens_seen": 2603220992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148080808080808,
      "loss": 2.531,
      "theoretical_loss": 3.3577038639757286,
      "tokens_seen": 2603286528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021478787878787879,
      "loss": 2.4638,
      "theoretical_loss": 3.3576971761100043,
      "tokens_seen": 2603352064
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021476767676767677,
      "loss": 2.4935,
      "theoretical_loss": 3.357690488459775,
      "tokens_seen": 2603417600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021474747474747475,
      "loss": 2.6634,
      "theoretical_loss": 3.3576838010250287,
      "tokens_seen": 2603483136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021472727272727274,
      "loss": 2.5092,
      "theoretical_loss": 3.3576771138057526,
      "tokens_seen": 2603548672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021470707070707072,
      "loss": 2.5816,
      "theoretical_loss": 3.357670426801935,
      "tokens_seen": 2603614208
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1425394,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.416898250579834,
      "objective/train/theoretical_loss": 3.357663740013563,
      "objective/train/tokens_used": 962538976,
      "theoretical_loss": 3.357663740013563,
      "tokens_seen": 2603679744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021468686868686868,
      "loss": 2.6167,
      "theoretical_loss": 3.357663740013563,
      "tokens_seen": 2603679744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002146666666666667,
      "loss": 2.3749,
      "theoretical_loss": 3.357657053440624,
      "tokens_seen": 2603745280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021464646464646465,
      "loss": 2.5893,
      "theoretical_loss": 3.3576503670831057,
      "tokens_seen": 2603810816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021462626262626263,
      "loss": 2.3739,
      "theoretical_loss": 3.357643680940997,
      "tokens_seen": 2603876352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021460606060606061,
      "loss": 2.5018,
      "theoretical_loss": 3.3576369950142837,
      "tokens_seen": 2603941888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021458585858585857,
      "loss": 2.452,
      "theoretical_loss": 3.3576303093029547,
      "tokens_seen": 2604007424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021456565656565658,
      "loss": 2.6996,
      "theoretical_loss": 3.3576236238069974,
      "tokens_seen": 2604072960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021454545454545454,
      "loss": 2.5707,
      "theoretical_loss": 3.357616938526399,
      "tokens_seen": 2604138496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021452525252525255,
      "loss": 2.549,
      "theoretical_loss": 3.3576102534611474,
      "tokens_seen": 2604204032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002145050505050505,
      "loss": 2.7563,
      "theoretical_loss": 3.35760356861123,
      "tokens_seen": 2604269568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021448484848484846,
      "loss": 2.6428,
      "theoretical_loss": 3.357596883976635,
      "tokens_seen": 2604335104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021446464646464647,
      "loss": 2.5993,
      "theoretical_loss": 3.3575901995573503,
      "tokens_seen": 2604400640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021444444444444443,
      "loss": 2.5414,
      "theoretical_loss": 3.357583515353362,
      "tokens_seen": 2604466176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021442424242424244,
      "loss": 2.5438,
      "theoretical_loss": 3.3575768313646597,
      "tokens_seen": 2604531712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002144040404040404,
      "loss": 2.4132,
      "theoretical_loss": 3.3575701475912294,
      "tokens_seen": 2604597248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021438383838383838,
      "loss": 2.5396,
      "theoretical_loss": 3.35756346403306,
      "tokens_seen": 2604662784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021436363636363637,
      "loss": 2.2879,
      "theoretical_loss": 3.357556780690138,
      "tokens_seen": 2604728320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021434343434343435,
      "loss": 2.7272,
      "theoretical_loss": 3.357550097562452,
      "tokens_seen": 2604793856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021432323232323234,
      "loss": 2.5865,
      "theoretical_loss": 3.357543414649989,
      "tokens_seen": 2604859392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002143030303030303,
      "loss": 2.5205,
      "theoretical_loss": 3.3575367319527376,
      "tokens_seen": 2604924928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021428282828282828,
      "loss": 2.757,
      "theoretical_loss": 3.3575300494706846,
      "tokens_seen": 2604990464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021426262626262626,
      "loss": 2.8165,
      "theoretical_loss": 3.3575233672038176,
      "tokens_seen": 2605056000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021424242424242424,
      "loss": 2.3575,
      "theoretical_loss": 3.3575166851521248,
      "tokens_seen": 2605121536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021422222222222223,
      "loss": 2.4721,
      "theoretical_loss": 3.3575100033155936,
      "tokens_seen": 2605187072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002142020202020202,
      "loss": 2.4847,
      "theoretical_loss": 3.357503321694211,
      "tokens_seen": 2605252608
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1425905,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4751393795013428,
      "objective/train/theoretical_loss": 3.357496640287966,
      "objective/train/tokens_used": 964177376,
      "theoretical_loss": 3.357496640287966,
      "tokens_seen": 2605318144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141818181818182,
      "loss": 2.2784,
      "theoretical_loss": 3.357496640287966,
      "tokens_seen": 2605318144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021416161616161618,
      "loss": 2.5948,
      "theoretical_loss": 3.3574899590968457,
      "tokens_seen": 2605383680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021414141414141414,
      "loss": 2.5617,
      "theoretical_loss": 3.3574832781208372,
      "tokens_seen": 2605449216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021412121212121212,
      "loss": 2.6004,
      "theoretical_loss": 3.3574765973599288,
      "tokens_seen": 2605514752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141010101010101,
      "loss": 2.5505,
      "theoretical_loss": 3.357469916814108,
      "tokens_seen": 2605580288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002140808080808081,
      "loss": 2.7487,
      "theoretical_loss": 3.3574632364833623,
      "tokens_seen": 2605645824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021406060606060607,
      "loss": 2.55,
      "theoretical_loss": 3.3574565563676795,
      "tokens_seen": 2605711360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021404040404040403,
      "loss": 2.5279,
      "theoretical_loss": 3.3574498764670473,
      "tokens_seen": 2605776896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021402020202020204,
      "loss": 2.6366,
      "theoretical_loss": 3.3574431967814533,
      "tokens_seen": 2605842432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000214,
      "loss": 2.702,
      "theoretical_loss": 3.3574365173108855,
      "tokens_seen": 2605907968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000213979797979798,
      "loss": 2.6411,
      "theoretical_loss": 3.3574298380553307,
      "tokens_seen": 2605973504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021395959595959597,
      "loss": 2.3604,
      "theoretical_loss": 3.3574231590147776,
      "tokens_seen": 2606039040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021393939393939392,
      "loss": 2.3543,
      "theoretical_loss": 3.3574164801892135,
      "tokens_seen": 2606104576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021391919191919193,
      "loss": 2.6456,
      "theoretical_loss": 3.3574098015786253,
      "tokens_seen": 2606170112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002138989898989899,
      "loss": 2.6066,
      "theoretical_loss": 3.357403123183002,
      "tokens_seen": 2606235648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002138787878787879,
      "loss": 2.6323,
      "theoretical_loss": 3.3573964450023306,
      "tokens_seen": 2606301184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021385858585858586,
      "loss": 2.3246,
      "theoretical_loss": 3.3573897670365986,
      "tokens_seen": 2606366720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021383838383838384,
      "loss": 2.4493,
      "theoretical_loss": 3.3573830892857943,
      "tokens_seen": 2606432256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021381818181818183,
      "loss": 2.4728,
      "theoretical_loss": 3.3573764117499043,
      "tokens_seen": 2606497792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137979797979798,
      "loss": 2.6322,
      "theoretical_loss": 3.3573697344289175,
      "tokens_seen": 2606563328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137777777777778,
      "loss": 2.572,
      "theoretical_loss": 3.3573630573228206,
      "tokens_seen": 2606628864
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021375757575757575,
      "loss": 2.8425,
      "theoretical_loss": 3.357356380431602,
      "tokens_seen": 2606694400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021373737373737373,
      "loss": 2.7532,
      "theoretical_loss": 3.3573497037552493,
      "tokens_seen": 2606759936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021371717171717172,
      "loss": 2.5353,
      "theoretical_loss": 3.3573430272937497,
      "tokens_seen": 2606825472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002136969696969697,
      "loss": 2.4463,
      "theoretical_loss": 3.357336351047091,
      "tokens_seen": 2606891008
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1426905,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.029381275177002,
      "objective/train/theoretical_loss": 3.3573296750152615,
      "objective/train/tokens_used": 965815776,
      "theoretical_loss": 3.3573296750152615,
      "tokens_seen": 2606956544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002136767676767677,
      "loss": 2.7504,
      "theoretical_loss": 3.3573296750152615,
      "tokens_seen": 2606956544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021365656565656567,
      "loss": 2.5439,
      "theoretical_loss": 3.3573229991982485,
      "tokens_seen": 2607022080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021363636363636363,
      "loss": 2.4821,
      "theoretical_loss": 3.3573163235960393,
      "tokens_seen": 2607087616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021361616161616164,
      "loss": 2.5,
      "theoretical_loss": 3.3573096482086218,
      "tokens_seen": 2607153152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002135959595959596,
      "loss": 2.5273,
      "theoretical_loss": 3.357302973035984,
      "tokens_seen": 2607218688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021357575757575758,
      "loss": 2.4311,
      "theoretical_loss": 3.357296298078113,
      "tokens_seen": 2607284224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021355555555555556,
      "loss": 2.5382,
      "theoretical_loss": 3.3572896233349976,
      "tokens_seen": 2607349760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021353535353535352,
      "loss": 2.4258,
      "theoretical_loss": 3.3572829488066245,
      "tokens_seen": 2607415296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021351515151515153,
      "loss": 2.5728,
      "theoretical_loss": 3.357276274492982,
      "tokens_seen": 2607480832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134949494949495,
      "loss": 2.8131,
      "theoretical_loss": 3.3572696003940568,
      "tokens_seen": 2607546368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134747474747475,
      "loss": 2.6259,
      "theoretical_loss": 3.3572629265098377,
      "tokens_seen": 2607611904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021345454545454546,
      "loss": 2.501,
      "theoretical_loss": 3.357256252840312,
      "tokens_seen": 2607677440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134343434343434,
      "loss": 2.5461,
      "theoretical_loss": 3.357249579385467,
      "tokens_seen": 2607742976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021341414141414142,
      "loss": 2.4399,
      "theoretical_loss": 3.357242906145291,
      "tokens_seen": 2607808512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021339393939393938,
      "loss": 2.4862,
      "theoretical_loss": 3.3572362331197714,
      "tokens_seen": 2607874048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002133737373737374,
      "loss": 2.461,
      "theoretical_loss": 3.357229560308896,
      "tokens_seen": 2607939584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021335353535353535,
      "loss": 2.5046,
      "theoretical_loss": 3.3572228877126524,
      "tokens_seen": 2608005120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021333333333333336,
      "loss": 2.4246,
      "theoretical_loss": 3.3572162153310288,
      "tokens_seen": 2608070656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021331313131313132,
      "loss": 2.5559,
      "theoretical_loss": 3.3572095431640117,
      "tokens_seen": 2608136192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132929292929293,
      "loss": 2.557,
      "theoretical_loss": 3.35720287121159,
      "tokens_seen": 2608201728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021327272727272728,
      "loss": 2.4461,
      "theoretical_loss": 3.3571961994737514,
      "tokens_seen": 2608267264
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021325252525252524,
      "loss": 2.5888,
      "theoretical_loss": 3.3571895279504824,
      "tokens_seen": 2608332800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021323232323232325,
      "loss": 2.5643,
      "theoretical_loss": 3.357182856641772,
      "tokens_seen": 2608398336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132121212121212,
      "loss": 2.6168,
      "theoretical_loss": 3.3571761855476074,
      "tokens_seen": 2608463872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002131919191919192,
      "loss": 2.6193,
      "theoretical_loss": 3.357169514667976,
      "tokens_seen": 2608529408
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1427573,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6366658210754395,
      "objective/train/theoretical_loss": 3.3571628440028665,
      "objective/train/tokens_used": 967454176,
      "theoretical_loss": 3.3571628440028665,
      "tokens_seen": 2608594944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021317171717171718,
      "loss": 2.6232,
      "theoretical_loss": 3.3571628440028665,
      "tokens_seen": 2608594944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021315151515151516,
      "loss": 2.371,
      "theoretical_loss": 3.3571561735522653,
      "tokens_seen": 2608660480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021313131313131315,
      "loss": 2.5256,
      "theoretical_loss": 3.357149503316161,
      "tokens_seen": 2608726016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021311111111111113,
      "loss": 2.317,
      "theoretical_loss": 3.357142833294541,
      "tokens_seen": 2608791552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021309090909090909,
      "loss": 2.3651,
      "theoretical_loss": 3.357136163487393,
      "tokens_seen": 2608857088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021307070707070707,
      "loss": 2.5298,
      "theoretical_loss": 3.357129493894705,
      "tokens_seen": 2608922624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021305050505050505,
      "loss": 2.3274,
      "theoretical_loss": 3.3571228245164644,
      "tokens_seen": 2608988160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021303030303030304,
      "loss": 2.3741,
      "theoretical_loss": 3.357116155352659,
      "tokens_seen": 2609053696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021301010101010102,
      "loss": 2.9016,
      "theoretical_loss": 3.357109486403277,
      "tokens_seen": 2609119232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021298989898989898,
      "loss": 2.4458,
      "theoretical_loss": 3.3571028176683053,
      "tokens_seen": 2609184768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000212969696969697,
      "loss": 2.5049,
      "theoretical_loss": 3.357096149147732,
      "tokens_seen": 2609250304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021294949494949495,
      "loss": 2.5534,
      "theoretical_loss": 3.3570894808415446,
      "tokens_seen": 2609315840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021292929292929296,
      "loss": 2.4275,
      "theoretical_loss": 3.3570828127497316,
      "tokens_seen": 2609381376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021290909090909091,
      "loss": 2.6579,
      "theoretical_loss": 3.3570761448722797,
      "tokens_seen": 2609446912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021288888888888887,
      "loss": 2.5843,
      "theoretical_loss": 3.3570694772091776,
      "tokens_seen": 2609512448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021286868686868688,
      "loss": 2.6099,
      "theoretical_loss": 3.357062809760412,
      "tokens_seen": 2609577984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021284848484848484,
      "loss": 2.6261,
      "theoretical_loss": 3.3570561425259715,
      "tokens_seen": 2609643520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021282828282828285,
      "loss": 2.3038,
      "theoretical_loss": 3.357049475505843,
      "tokens_seen": 2609709056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002128080808080808,
      "loss": 2.35,
      "theoretical_loss": 3.3570428087000153,
      "tokens_seen": 2609774592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127878787878788,
      "loss": 2.4055,
      "theoretical_loss": 3.357036142108475,
      "tokens_seen": 2609840128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021276767676767678,
      "loss": 2.6026,
      "theoretical_loss": 3.3570294757312107,
      "tokens_seen": 2609905664
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021274747474747473,
      "loss": 2.447,
      "theoretical_loss": 3.35702280956821,
      "tokens_seen": 2609971200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021272727272727274,
      "loss": 2.6633,
      "theoretical_loss": 3.35701614361946,
      "tokens_seen": 2610036736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127070707070707,
      "loss": 2.5256,
      "theoretical_loss": 3.357009477884949,
      "tokens_seen": 2610102272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021268686868686868,
      "loss": 2.6561,
      "theoretical_loss": 3.357002812364665,
      "tokens_seen": 2610167808
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1428899,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.595594882965088,
      "objective/train/theoretical_loss": 3.3569961470585947,
      "objective/train/tokens_used": 969092576,
      "theoretical_loss": 3.3569961470585947,
      "tokens_seen": 2610233344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021266666666666667,
      "loss": 2.6376,
      "theoretical_loss": 3.3569961470585947,
      "tokens_seen": 2610233344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021264646464646465,
      "loss": 2.6591,
      "theoretical_loss": 3.356989481966727,
      "tokens_seen": 2610298880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021262626262626264,
      "loss": 2.6441,
      "theoretical_loss": 3.3569828170890488,
      "tokens_seen": 2610364416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021260606060606062,
      "loss": 2.4531,
      "theoretical_loss": 3.356976152425548,
      "tokens_seen": 2610429952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021258585858585858,
      "loss": 2.4429,
      "theoretical_loss": 3.356969487976213,
      "tokens_seen": 2610495488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021256565656565656,
      "loss": 2.5106,
      "theoretical_loss": 3.3569628237410307,
      "tokens_seen": 2610561024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021254545454545454,
      "loss": 2.6624,
      "theoretical_loss": 3.3569561597199895,
      "tokens_seen": 2610626560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021252525252525253,
      "loss": 2.4727,
      "theoretical_loss": 3.3569494959130766,
      "tokens_seen": 2610692096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002125050505050505,
      "loss": 2.4208,
      "theoretical_loss": 3.35694283232028,
      "tokens_seen": 2610757632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124848484848485,
      "loss": 2.5561,
      "theoretical_loss": 3.356936168941587,
      "tokens_seen": 2610823168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021246464646464648,
      "loss": 2.6174,
      "theoretical_loss": 3.3569295057769866,
      "tokens_seen": 2610888704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021244444444444444,
      "loss": 2.5465,
      "theoretical_loss": 3.356922842826465,
      "tokens_seen": 2610954240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021242424242424245,
      "loss": 2.4753,
      "theoretical_loss": 3.3569161800900114,
      "tokens_seen": 2611019776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124040404040404,
      "loss": 2.567,
      "theoretical_loss": 3.3569095175676122,
      "tokens_seen": 2611085312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002123838383838384,
      "loss": 2.537,
      "theoretical_loss": 3.3569028552592557,
      "tokens_seen": 2611150848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021236363636363637,
      "loss": 2.4309,
      "theoretical_loss": 3.35689619316493,
      "tokens_seen": 2611216384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021234343434343433,
      "loss": 2.4974,
      "theoretical_loss": 3.3568895312846223,
      "tokens_seen": 2611281920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021232323232323234,
      "loss": 2.569,
      "theoretical_loss": 3.356882869618321,
      "tokens_seen": 2611347456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002123030303030303,
      "loss": 2.5171,
      "theoretical_loss": 3.3568762081660135,
      "tokens_seen": 2611412992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002122828282828283,
      "loss": 2.514,
      "theoretical_loss": 3.3568695469276872,
      "tokens_seen": 2611478528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021226262626262627,
      "loss": 2.4508,
      "theoretical_loss": 3.3568628859033303,
      "tokens_seen": 2611544064
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021224242424242425,
      "loss": 2.5712,
      "theoretical_loss": 3.3568562250929306,
      "tokens_seen": 2611609600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021222222222222223,
      "loss": 2.4007,
      "theoretical_loss": 3.3568495644964753,
      "tokens_seen": 2611675136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002122020202020202,
      "loss": 2.4889,
      "theoretical_loss": 3.356842904113953,
      "tokens_seen": 2611740672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121818181818182,
      "loss": 2.5711,
      "theoretical_loss": 3.3568362439453505,
      "tokens_seen": 2611806208
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1430125,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6226844787597656,
      "objective/train/theoretical_loss": 3.3568295839906566,
      "objective/train/tokens_used": 970730976,
      "theoretical_loss": 3.3568295839906566,
      "tokens_seen": 2611871744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021216161616161616,
      "loss": 2.613,
      "theoretical_loss": 3.3568295839906566,
      "tokens_seen": 2611871744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021214141414141414,
      "loss": 2.5875,
      "theoretical_loss": 3.356822924249858,
      "tokens_seen": 2611937280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021212121212121213,
      "loss": 2.5992,
      "theoretical_loss": 3.356816264722944,
      "tokens_seen": 2612002816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121010101010101,
      "loss": 2.5432,
      "theoretical_loss": 3.3568096054099006,
      "tokens_seen": 2612068352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120808080808081,
      "loss": 2.5754,
      "theoretical_loss": 3.3568029463107165,
      "tokens_seen": 2612133888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021206060606060608,
      "loss": 2.5232,
      "theoretical_loss": 3.356796287425379,
      "tokens_seen": 2612199424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021204040404040404,
      "loss": 2.3733,
      "theoretical_loss": 3.3567896287538765,
      "tokens_seen": 2612264960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021202020202020202,
      "loss": 2.4386,
      "theoretical_loss": 3.3567829702961967,
      "tokens_seen": 2612330496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000212,
      "loss": 2.6223,
      "theoretical_loss": 3.3567763120523266,
      "tokens_seen": 2612396032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000211979797979798,
      "loss": 2.5512,
      "theoretical_loss": 3.356769654022255,
      "tokens_seen": 2612461568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021195959595959597,
      "loss": 2.428,
      "theoretical_loss": 3.356762996205969,
      "tokens_seen": 2612527104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021193939393939393,
      "loss": 2.3907,
      "theoretical_loss": 3.3567563386034562,
      "tokens_seen": 2612592640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021191919191919194,
      "loss": 2.6224,
      "theoretical_loss": 3.356749681214705,
      "tokens_seen": 2612658176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002118989898989899,
      "loss": 2.4833,
      "theoretical_loss": 3.356743024039703,
      "tokens_seen": 2612723712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021187878787878788,
      "loss": 2.5694,
      "theoretical_loss": 3.3567363670784376,
      "tokens_seen": 2612789248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021185858585858586,
      "loss": 2.4991,
      "theoretical_loss": 3.356729710330897,
      "tokens_seen": 2612854784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021183838383838382,
      "loss": 2.4517,
      "theoretical_loss": 3.3567230537970683,
      "tokens_seen": 2612920320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021181818181818183,
      "loss": 2.589,
      "theoretical_loss": 3.3567163974769403,
      "tokens_seen": 2612985856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002117979797979798,
      "loss": 2.6619,
      "theoretical_loss": 3.3567097413705,
      "tokens_seen": 2613051392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002117777777777778,
      "loss": 2.4908,
      "theoretical_loss": 3.356703085477736,
      "tokens_seen": 2613116928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021175757575757576,
      "loss": 2.5084,
      "theoretical_loss": 3.3566964297986353,
      "tokens_seen": 2613182464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021173737373737374,
      "loss": 2.4844,
      "theoretical_loss": 3.3566897743331854,
      "tokens_seen": 2613248000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021171717171717172,
      "loss": 2.6649,
      "theoretical_loss": 3.356683119081375,
      "tokens_seen": 2613313536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021169696969696968,
      "loss": 2.8031,
      "theoretical_loss": 3.3566764640431916,
      "tokens_seen": 2613379072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002116767676767677,
      "loss": 2.6248,
      "theoretical_loss": 3.3566698092186225,
      "tokens_seen": 2613444608
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1430667,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.676116943359375,
      "objective/train/theoretical_loss": 3.356663154607656,
      "objective/train/tokens_used": 972369376,
      "theoretical_loss": 3.356663154607656,
      "tokens_seen": 2613510144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021165656565656565,
      "loss": 2.7086,
      "theoretical_loss": 3.356663154607656,
      "tokens_seen": 2613510144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021163636363636366,
      "loss": 2.6487,
      "theoretical_loss": 3.35665650021028,
      "tokens_seen": 2613575680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021161616161616162,
      "loss": 2.4514,
      "theoretical_loss": 3.3566498460264818,
      "tokens_seen": 2613641216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002115959595959596,
      "loss": 2.3919,
      "theoretical_loss": 3.356643192056249,
      "tokens_seen": 2613706752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021157575757575758,
      "loss": 2.4953,
      "theoretical_loss": 3.3566365382995706,
      "tokens_seen": 2613772288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021155555555555557,
      "loss": 2.708,
      "theoretical_loss": 3.3566298847564333,
      "tokens_seen": 2613837824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021153535353535355,
      "loss": 2.3042,
      "theoretical_loss": 3.356623231426825,
      "tokens_seen": 2613903360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002115151515151515,
      "loss": 2.3695,
      "theoretical_loss": 3.3566165783107342,
      "tokens_seen": 2613968896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002114949494949495,
      "loss": 2.6424,
      "theoretical_loss": 3.3566099254081476,
      "tokens_seen": 2614034432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021147474747474748,
      "loss": 2.5789,
      "theoretical_loss": 3.3566032727190542,
      "tokens_seen": 2614099968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021145454545454546,
      "loss": 2.4884,
      "theoretical_loss": 3.3565966202434407,
      "tokens_seen": 2614165504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021143434343434345,
      "loss": 2.4687,
      "theoretical_loss": 3.3565899679812956,
      "tokens_seen": 2614231040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021141414141414143,
      "loss": 2.5493,
      "theoretical_loss": 3.356583315932606,
      "tokens_seen": 2614296576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021139393939393939,
      "loss": 2.4455,
      "theoretical_loss": 3.3565766640973607,
      "tokens_seen": 2614362112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002113737373737374,
      "loss": 2.4711,
      "theoretical_loss": 3.356570012475547,
      "tokens_seen": 2614427648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021135353535353535,
      "loss": 2.7244,
      "theoretical_loss": 3.3565633610671526,
      "tokens_seen": 2614493184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021133333333333334,
      "loss": 2.4952,
      "theoretical_loss": 3.356556709872165,
      "tokens_seen": 2614558720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021131313131313132,
      "loss": 2.5953,
      "theoretical_loss": 3.356550058890573,
      "tokens_seen": 2614624256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021129292929292928,
      "loss": 2.4256,
      "theoretical_loss": 3.3565434081223633,
      "tokens_seen": 2614689792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002112727272727273,
      "loss": 2.4189,
      "theoretical_loss": 3.3565367575675245,
      "tokens_seen": 2614755328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021125252525252525,
      "loss": 2.6432,
      "theoretical_loss": 3.356530107226044,
      "tokens_seen": 2614820864
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021123232323232326,
      "loss": 2.6864,
      "theoretical_loss": 3.3565234570979094,
      "tokens_seen": 2614886400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021121212121212121,
      "loss": 2.8216,
      "theoretical_loss": 3.3565168071831093,
      "tokens_seen": 2614951936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021119191919191917,
      "loss": 2.5928,
      "theoretical_loss": 3.356510157481631,
      "tokens_seen": 2615017472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021117171717171718,
      "loss": 2.4133,
      "theoretical_loss": 3.356503507993462,
      "tokens_seen": 2615083008
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1431756,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4766592979431152,
      "objective/train/theoretical_loss": 3.3564968587185904,
      "objective/train/tokens_used": 974007776,
      "theoretical_loss": 3.3564968587185904,
      "tokens_seen": 2615148544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021115151515151514,
      "loss": 2.6954,
      "theoretical_loss": 3.3564968587185904,
      "tokens_seen": 2615148544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021113131313131315,
      "loss": 2.4157,
      "theoretical_loss": 3.3564902096570046,
      "tokens_seen": 2615214080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002111111111111111,
      "loss": 2.578,
      "theoretical_loss": 3.356483560808692,
      "tokens_seen": 2615279616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002110909090909091,
      "loss": 2.5581,
      "theoretical_loss": 3.3564769121736395,
      "tokens_seen": 2615345152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021107070707070708,
      "loss": 2.4454,
      "theoretical_loss": 3.356470263751836,
      "tokens_seen": 2615410688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021105050505050506,
      "loss": 2.5662,
      "theoretical_loss": 3.3564636155432694,
      "tokens_seen": 2615476224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021103030303030304,
      "loss": 2.5524,
      "theoretical_loss": 3.3564569675479268,
      "tokens_seen": 2615541760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000211010101010101,
      "loss": 2.5799,
      "theoretical_loss": 3.356450319765796,
      "tokens_seen": 2615607296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021098989898989898,
      "loss": 2.2762,
      "theoretical_loss": 3.356443672196866,
      "tokens_seen": 2615672832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021096969696969697,
      "loss": 2.574,
      "theoretical_loss": 3.3564370248411235,
      "tokens_seen": 2615738368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021094949494949495,
      "loss": 2.8441,
      "theoretical_loss": 3.3564303776985565,
      "tokens_seen": 2615803904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021092929292929294,
      "loss": 2.722,
      "theoretical_loss": 3.356423730769153,
      "tokens_seen": 2615869440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021090909090909092,
      "loss": 2.5251,
      "theoretical_loss": 3.356417084052901,
      "tokens_seen": 2615934976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021088888888888888,
      "loss": 2.3438,
      "theoretical_loss": 3.356410437549788,
      "tokens_seen": 2616000512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002108686868686869,
      "loss": 2.7276,
      "theoretical_loss": 3.3564037912598015,
      "tokens_seen": 2616066048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021084848484848484,
      "loss": 2.5053,
      "theoretical_loss": 3.35639714518293,
      "tokens_seen": 2616131584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021082828282828283,
      "loss": 2.4028,
      "theoretical_loss": 3.3563904993191613,
      "tokens_seen": 2616197120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002108080808080808,
      "loss": 2.4793,
      "theoretical_loss": 3.3563838536684827,
      "tokens_seen": 2616262656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021078787878787877,
      "loss": 2.3259,
      "theoretical_loss": 3.356377208230883,
      "tokens_seen": 2616328192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021076767676767678,
      "loss": 2.4687,
      "theoretical_loss": 3.3563705630063487,
      "tokens_seen": 2616393728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021074747474747474,
      "loss": 2.4625,
      "theoretical_loss": 3.3563639179948686,
      "tokens_seen": 2616459264
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021072727272727275,
      "loss": 2.6138,
      "theoretical_loss": 3.35635727319643,
      "tokens_seen": 2616524800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002107070707070707,
      "loss": 2.7129,
      "theoretical_loss": 3.356350628611021,
      "tokens_seen": 2616590336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021068686868686872,
      "loss": 2.5521,
      "theoretical_loss": 3.3563439842386296,
      "tokens_seen": 2616655872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021066666666666667,
      "loss": 2.5632,
      "theoretical_loss": 3.3563373400792433,
      "tokens_seen": 2616721408
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1432270,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.063832998275757,
      "objective/train/theoretical_loss": 3.3563306961328503,
      "objective/train/tokens_used": 975646176,
      "theoretical_loss": 3.3563306961328503,
      "tokens_seen": 2616786944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021064646464646463,
      "loss": 2.4609,
      "theoretical_loss": 3.3563306961328503,
      "tokens_seen": 2616786944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021062626262626264,
      "loss": 2.4333,
      "theoretical_loss": 3.356324052399438,
      "tokens_seen": 2616852480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002106060606060606,
      "loss": 2.2694,
      "theoretical_loss": 3.3563174088789944,
      "tokens_seen": 2616918016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002105858585858586,
      "loss": 2.5277,
      "theoretical_loss": 3.3563107655715076,
      "tokens_seen": 2616983552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021056565656565657,
      "loss": 2.6887,
      "theoretical_loss": 3.356304122476965,
      "tokens_seen": 2617049088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021054545454545455,
      "loss": 2.5384,
      "theoretical_loss": 3.356297479595355,
      "tokens_seen": 2617114624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021052525252525253,
      "loss": 2.7132,
      "theoretical_loss": 3.356290836926665,
      "tokens_seen": 2617180160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021050505050505052,
      "loss": 2.7574,
      "theoretical_loss": 3.356284194470883,
      "tokens_seen": 2617245696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002104848484848485,
      "loss": 2.5398,
      "theoretical_loss": 3.3562775522279966,
      "tokens_seen": 2617311232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021046464646464646,
      "loss": 2.5217,
      "theoretical_loss": 3.356270910197994,
      "tokens_seen": 2617376768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021044444444444444,
      "loss": 2.57,
      "theoretical_loss": 3.356264268380863,
      "tokens_seen": 2617442304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021042424242424243,
      "loss": 2.659,
      "theoretical_loss": 3.3562576267765913,
      "tokens_seen": 2617507840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002104040404040404,
      "loss": 2.8339,
      "theoretical_loss": 3.3562509853851665,
      "tokens_seen": 2617573376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002103838383838384,
      "loss": 2.6079,
      "theoretical_loss": 3.3562443442065772,
      "tokens_seen": 2617638912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021036363636363638,
      "loss": 2.5655,
      "theoretical_loss": 3.3562377032408106,
      "tokens_seen": 2617704448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021034343434343434,
      "loss": 2.3997,
      "theoretical_loss": 3.3562310624878546,
      "tokens_seen": 2617769984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021032323232323232,
      "loss": 2.7091,
      "theoretical_loss": 3.3562244219476973,
      "tokens_seen": 2617835520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002103030303030303,
      "loss": 2.6427,
      "theoretical_loss": 3.3562177816203267,
      "tokens_seen": 2617901056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002102828282828283,
      "loss": 2.6266,
      "theoretical_loss": 3.3562111415057303,
      "tokens_seen": 2617966592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021026262626262627,
      "loss": 2.6516,
      "theoretical_loss": 3.356204501603896,
      "tokens_seen": 2618032128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021024242424242423,
      "loss": 2.6616,
      "theoretical_loss": 3.356197861914812,
      "tokens_seen": 2618097664
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021022222222222224,
      "loss": 2.3793,
      "theoretical_loss": 3.356191222438466,
      "tokens_seen": 2618163200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002102020202020202,
      "loss": 2.6618,
      "theoretical_loss": 3.356184583174845,
      "tokens_seen": 2618228736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002101818181818182,
      "loss": 2.5448,
      "theoretical_loss": 3.356177944123938,
      "tokens_seen": 2618294272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021016161616161616,
      "loss": 2.5587,
      "theoretical_loss": 3.3561713052857325,
      "tokens_seen": 2618359808
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1433381,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.570761203765869,
      "objective/train/theoretical_loss": 3.356164666660216,
      "objective/train/tokens_used": 977284576,
      "theoretical_loss": 3.356164666660216,
      "tokens_seen": 2618425344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021014141414141412,
      "loss": 2.7156,
      "theoretical_loss": 3.356164666660216,
      "tokens_seen": 2618425344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021012121212121213,
      "loss": 2.3835,
      "theoretical_loss": 3.356158028247377,
      "tokens_seen": 2618490880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002101010101010101,
      "loss": 2.4358,
      "theoretical_loss": 3.356151390047203,
      "tokens_seen": 2618556416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002100808080808081,
      "loss": 2.434,
      "theoretical_loss": 3.356144752059682,
      "tokens_seen": 2618621952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021006060606060606,
      "loss": 2.7625,
      "theoretical_loss": 3.3561381142848017,
      "tokens_seen": 2618687488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021004040404040404,
      "loss": 2.5602,
      "theoretical_loss": 3.3561314767225503,
      "tokens_seen": 2618753024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021002020202020202,
      "loss": 2.3882,
      "theoretical_loss": 3.356124839372915,
      "tokens_seen": 2618818560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021,
      "loss": 2.4798,
      "theoretical_loss": 3.3561182022358844,
      "tokens_seen": 2618884096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000209979797979798,
      "loss": 2.632,
      "theoretical_loss": 3.356111565311446,
      "tokens_seen": 2618949632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020995959595959595,
      "loss": 2.485,
      "theoretical_loss": 3.3561049285995876,
      "tokens_seen": 2619015168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020993939393939393,
      "loss": 2.6499,
      "theoretical_loss": 3.3560982921002975,
      "tokens_seen": 2619080704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020991919191919192,
      "loss": 2.6056,
      "theoretical_loss": 3.356091655813563,
      "tokens_seen": 2619146240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002098989898989899,
      "loss": 2.6201,
      "theoretical_loss": 3.3560850197393726,
      "tokens_seen": 2619211776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020987878787878789,
      "loss": 2.5483,
      "theoretical_loss": 3.3560783838777137,
      "tokens_seen": 2619277312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020985858585858587,
      "loss": 2.4559,
      "theoretical_loss": 3.356071748228574,
      "tokens_seen": 2619342848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020983838383838385,
      "loss": 2.7005,
      "theoretical_loss": 3.3560651127919416,
      "tokens_seen": 2619408384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020981818181818184,
      "loss": 2.6068,
      "theoretical_loss": 3.356058477567805,
      "tokens_seen": 2619473920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002097979797979798,
      "loss": 2.7769,
      "theoretical_loss": 3.356051842556151,
      "tokens_seen": 2619539456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020977777777777778,
      "loss": 2.5373,
      "theoretical_loss": 3.3560452077569685,
      "tokens_seen": 2619604992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020975757575757576,
      "loss": 2.3276,
      "theoretical_loss": 3.3560385731702445,
      "tokens_seen": 2619670528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020973737373737375,
      "loss": 2.5158,
      "theoretical_loss": 3.3560319387959674,
      "tokens_seen": 2619736064
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020971717171717173,
      "loss": 2.4936,
      "theoretical_loss": 3.356025304634125,
      "tokens_seen": 2619801600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002096969696969697,
      "loss": 2.5438,
      "theoretical_loss": 3.3560186706847053,
      "tokens_seen": 2619867136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002096767676767677,
      "loss": 2.4555,
      "theoretical_loss": 3.356012036947696,
      "tokens_seen": 2619932672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020965656565656565,
      "loss": 2.6299,
      "theoretical_loss": 3.356005403423085,
      "tokens_seen": 2619998208
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1434670,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.737292766571045,
      "objective/train/theoretical_loss": 3.3559987701108596,
      "objective/train/tokens_used": 978922976,
      "theoretical_loss": 3.3559987701108596,
      "tokens_seen": 2620063744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020963636363636367,
      "loss": 2.5781,
      "theoretical_loss": 3.3559987701108596,
      "tokens_seen": 2620063744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020961616161616162,
      "loss": 2.6748,
      "theoretical_loss": 3.355992137011009,
      "tokens_seen": 2620129280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020959595959595958,
      "loss": 2.5884,
      "theoretical_loss": 3.3559855041235203,
      "tokens_seen": 2620194816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002095757575757576,
      "loss": 2.6832,
      "theoretical_loss": 3.355978871448381,
      "tokens_seen": 2620260352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020955555555555555,
      "loss": 2.5558,
      "theoretical_loss": 3.3559722389855797,
      "tokens_seen": 2620325888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020953535353535356,
      "loss": 2.5155,
      "theoretical_loss": 3.3559656067351042,
      "tokens_seen": 2620391424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020951515151515152,
      "loss": 2.6279,
      "theoretical_loss": 3.355958974696942,
      "tokens_seen": 2620456960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002094949494949495,
      "loss": 2.7061,
      "theoretical_loss": 3.3559523428710816,
      "tokens_seen": 2620522496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020947474747474748,
      "loss": 2.3612,
      "theoretical_loss": 3.3559457112575104,
      "tokens_seen": 2620588032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020945454545454544,
      "loss": 2.4493,
      "theoretical_loss": 3.355939079856216,
      "tokens_seen": 2620653568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020943434343434345,
      "loss": 2.3312,
      "theoretical_loss": 3.355932448667187,
      "tokens_seen": 2620719104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002094141414141414,
      "loss": 2.349,
      "theoretical_loss": 3.3559258176904114,
      "tokens_seen": 2620784640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002093939393939394,
      "loss": 2.3716,
      "theoretical_loss": 3.3559191869258767,
      "tokens_seen": 2620850176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020937373737373738,
      "loss": 2.4964,
      "theoretical_loss": 3.3559125563735703,
      "tokens_seen": 2620915712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020935353535353536,
      "loss": 2.4622,
      "theoretical_loss": 3.355905926033481,
      "tokens_seen": 2620981248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020933333333333334,
      "loss": 2.3406,
      "theoretical_loss": 3.355899295905596,
      "tokens_seen": 2621046784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020931313131313133,
      "loss": 2.4408,
      "theoretical_loss": 3.355892665989904,
      "tokens_seen": 2621112320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020929292929292928,
      "loss": 2.4602,
      "theoretical_loss": 3.355886036286392,
      "tokens_seen": 2621177856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020927272727272727,
      "loss": 2.591,
      "theoretical_loss": 3.3558794067950486,
      "tokens_seen": 2621243392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020925252525252525,
      "loss": 2.6486,
      "theoretical_loss": 3.355872777515861,
      "tokens_seen": 2621308928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020923232323232324,
      "loss": 2.5668,
      "theoretical_loss": 3.3558661484488184,
      "tokens_seen": 2621374464
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020921212121212122,
      "loss": 2.7602,
      "theoretical_loss": 3.355859519593907,
      "tokens_seen": 2621440000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020919191919191918,
      "loss": 2.8709,
      "theoretical_loss": 3.355852890951116,
      "tokens_seen": 2621505536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091717171717172,
      "loss": 2.4494,
      "theoretical_loss": 3.3558462625204326,
      "tokens_seen": 2621571072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020915151515151515,
      "loss": 2.6668,
      "theoretical_loss": 3.3558396343018453,
      "tokens_seen": 2621636608
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1435251,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4519340991973877,
      "objective/train/theoretical_loss": 3.3558330062953416,
      "objective/train/tokens_used": 980561376,
      "theoretical_loss": 3.3558330062953416,
      "tokens_seen": 2621702144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020913131313131316,
      "loss": 2.3416,
      "theoretical_loss": 3.3558330062953416,
      "tokens_seen": 2621702144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091111111111111,
      "loss": 2.6353,
      "theoretical_loss": 3.3558263785009093,
      "tokens_seen": 2621767680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020909090909090907,
      "loss": 2.4991,
      "theoretical_loss": 3.3558197509185366,
      "tokens_seen": 2621833216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020907070707070708,
      "loss": 2.2038,
      "theoretical_loss": 3.3558131235482116,
      "tokens_seen": 2621898752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020905050505050504,
      "loss": 2.316,
      "theoretical_loss": 3.3558064963899215,
      "tokens_seen": 2621964288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020903030303030305,
      "loss": 2.3229,
      "theoretical_loss": 3.355799869443655,
      "tokens_seen": 2622029824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000209010101010101,
      "loss": 2.483,
      "theoretical_loss": 3.3557932427093995,
      "tokens_seen": 2622095360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020898989898989902,
      "loss": 2.517,
      "theoretical_loss": 3.3557866161871432,
      "tokens_seen": 2622160896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020896969696969697,
      "loss": 2.4839,
      "theoretical_loss": 3.3557799898768743,
      "tokens_seen": 2622226432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020894949494949496,
      "loss": 2.5892,
      "theoretical_loss": 3.35577336377858,
      "tokens_seen": 2622291968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020892929292929294,
      "loss": 2.3282,
      "theoretical_loss": 3.3557667378922487,
      "tokens_seen": 2622357504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002089090909090909,
      "loss": 2.7857,
      "theoretical_loss": 3.355760112217868,
      "tokens_seen": 2622423040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002088888888888889,
      "loss": 2.6895,
      "theoretical_loss": 3.355753486755426,
      "tokens_seen": 2622488576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020886868686868687,
      "loss": 2.4773,
      "theoretical_loss": 3.3557468615049104,
      "tokens_seen": 2622554112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020884848484848485,
      "loss": 2.8306,
      "theoretical_loss": 3.3557402364663096,
      "tokens_seen": 2622619648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020882828282828283,
      "loss": 2.3545,
      "theoretical_loss": 3.3557336116396117,
      "tokens_seen": 2622685184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020880808080808082,
      "loss": 2.7187,
      "theoretical_loss": 3.355726987024804,
      "tokens_seen": 2622750720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002087878787878788,
      "loss": 2.6083,
      "theoretical_loss": 3.355720362621874,
      "tokens_seen": 2622816256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020876767676767679,
      "loss": 2.5964,
      "theoretical_loss": 3.355713738430811,
      "tokens_seen": 2622881792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020874747474747474,
      "loss": 2.5436,
      "theoretical_loss": 3.3557071144516017,
      "tokens_seen": 2622947328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020872727272727273,
      "loss": 2.53,
      "theoretical_loss": 3.355700490684235,
      "tokens_seen": 2623012864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002087070707070707,
      "loss": 2.5755,
      "theoretical_loss": 3.355693867128698,
      "tokens_seen": 2623078400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086868686868687,
      "loss": 2.6648,
      "theoretical_loss": 3.3556872437849794,
      "tokens_seen": 2623143936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020866666666666668,
      "loss": 2.7903,
      "theoretical_loss": 3.3556806206530667,
      "tokens_seen": 2623209472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020864646464646464,
      "loss": 2.6338,
      "theoretical_loss": 3.3556739977329477,
      "tokens_seen": 2623275008
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1436460,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.362396478652954,
      "objective/train/theoretical_loss": 3.3556673750246104,
      "objective/train/tokens_used": 982199776,
      "theoretical_loss": 3.3556673750246104,
      "tokens_seen": 2623340544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020862626262626265,
      "loss": 2.5939,
      "theoretical_loss": 3.3556673750246104,
      "tokens_seen": 2623340544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086060606060606,
      "loss": 2.58,
      "theoretical_loss": 3.355660752528043,
      "tokens_seen": 2623406080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002085858585858586,
      "loss": 2.3428,
      "theoretical_loss": 3.3556541302432334,
      "tokens_seen": 2623471616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020856565656565657,
      "loss": 2.4109,
      "theoretical_loss": 3.3556475081701693,
      "tokens_seen": 2623537152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020854545454545453,
      "loss": 2.5077,
      "theoretical_loss": 3.3556408863088385,
      "tokens_seen": 2623602688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020852525252525254,
      "loss": 2.4767,
      "theoretical_loss": 3.3556342646592294,
      "tokens_seen": 2623668224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002085050505050505,
      "loss": 2.4522,
      "theoretical_loss": 3.3556276432213297,
      "tokens_seen": 2623733760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084848484848485,
      "loss": 2.4326,
      "theoretical_loss": 3.3556210219951277,
      "tokens_seen": 2623799296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020846464646464646,
      "loss": 2.7576,
      "theoretical_loss": 3.3556144009806106,
      "tokens_seen": 2623864832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020844444444444445,
      "loss": 2.586,
      "theoretical_loss": 3.355607780177767,
      "tokens_seen": 2623930368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020842424242424243,
      "loss": 2.4097,
      "theoretical_loss": 3.355601159586585,
      "tokens_seen": 2623995904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084040404040404,
      "loss": 2.5562,
      "theoretical_loss": 3.3555945392070514,
      "tokens_seen": 2624061440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002083838383838384,
      "loss": 2.556,
      "theoretical_loss": 3.3555879190391553,
      "tokens_seen": 2624126976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020836363636363636,
      "loss": 2.4892,
      "theoretical_loss": 3.3555812990828846,
      "tokens_seen": 2624192512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020834343434343434,
      "loss": 2.3191,
      "theoretical_loss": 3.3555746793382264,
      "tokens_seen": 2624258048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020832323232323233,
      "loss": 2.7068,
      "theoretical_loss": 3.355568059805169,
      "tokens_seen": 2624323584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002083030303030303,
      "loss": 2.5889,
      "theoretical_loss": 3.3555614404837013,
      "tokens_seen": 2624389120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082828282828283,
      "loss": 2.3777,
      "theoretical_loss": 3.35555482137381,
      "tokens_seen": 2624454656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020826262626262628,
      "loss": 2.545,
      "theoretical_loss": 3.3555482024754832,
      "tokens_seen": 2624520192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020824242424242423,
      "loss": 2.2774,
      "theoretical_loss": 3.35554158378871,
      "tokens_seen": 2624585728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020822222222222222,
      "loss": 2.4767,
      "theoretical_loss": 3.355534965313477,
      "tokens_seen": 2624651264
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082020202020202,
      "loss": 2.5686,
      "theoretical_loss": 3.355528347049773,
      "tokens_seen": 2624716800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020818181818181819,
      "loss": 2.518,
      "theoretical_loss": 3.3555217289975854,
      "tokens_seen": 2624782336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020816161616161617,
      "loss": 2.6216,
      "theoretical_loss": 3.3555151111569024,
      "tokens_seen": 2624847872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020814141414141415,
      "loss": 2.3936,
      "theoretical_loss": 3.3555084935277124,
      "tokens_seen": 2624913408
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1436953,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.385594606399536,
      "objective/train/theoretical_loss": 3.3555018761100026,
      "objective/train/tokens_used": 983838176,
      "theoretical_loss": 3.3555018761100026,
      "tokens_seen": 2624978944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020812121212121214,
      "loss": 2.589,
      "theoretical_loss": 3.3555018761100026,
      "tokens_seen": 2624978944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002081010101010101,
      "loss": 2.4644,
      "theoretical_loss": 3.355495258903761,
      "tokens_seen": 2625044480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002080808080808081,
      "loss": 2.5098,
      "theoretical_loss": 3.3554886419089764,
      "tokens_seen": 2625110016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020806060606060606,
      "loss": 2.4511,
      "theoretical_loss": 3.355482025125636,
      "tokens_seen": 2625175552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020804040404040405,
      "loss": 2.625,
      "theoretical_loss": 3.355475408553728,
      "tokens_seen": 2625241088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020802020202020203,
      "loss": 2.8077,
      "theoretical_loss": 3.35546879219324,
      "tokens_seen": 2625306624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000208,
      "loss": 2.6063,
      "theoretical_loss": 3.3554621760441607,
      "tokens_seen": 2625372160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000207979797979798,
      "loss": 2.4482,
      "theoretical_loss": 3.3554555601064777,
      "tokens_seen": 2625437696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020795959595959596,
      "loss": 2.4554,
      "theoretical_loss": 3.3554489443801794,
      "tokens_seen": 2625503232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020793939393939397,
      "loss": 2.3918,
      "theoretical_loss": 3.3554423288652524,
      "tokens_seen": 2625568768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020791919191919192,
      "loss": 2.6476,
      "theoretical_loss": 3.355435713561686,
      "tokens_seen": 2625634304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020789898989898988,
      "loss": 2.596,
      "theoretical_loss": 3.355429098469468,
      "tokens_seen": 2625699840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002078787878787879,
      "loss": 2.6903,
      "theoretical_loss": 3.3554224835885864,
      "tokens_seen": 2625765376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020785858585858585,
      "loss": 2.5218,
      "theoretical_loss": 3.3554158689190285,
      "tokens_seen": 2625830912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020783838383838386,
      "loss": 2.5325,
      "theoretical_loss": 3.3554092544607825,
      "tokens_seen": 2625896448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020781818181818182,
      "loss": 2.4557,
      "theoretical_loss": 3.355402640213837,
      "tokens_seen": 2625961984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002077979797979798,
      "loss": 2.6001,
      "theoretical_loss": 3.3553960261781794,
      "tokens_seen": 2626027520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020777777777777778,
      "loss": 2.6986,
      "theoretical_loss": 3.355389412353798,
      "tokens_seen": 2626093056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020775757575757577,
      "loss": 2.5085,
      "theoretical_loss": 3.35538279874068,
      "tokens_seen": 2626158592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020773737373737375,
      "loss": 2.6162,
      "theoretical_loss": 3.355376185338814,
      "tokens_seen": 2626224128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002077171717171717,
      "loss": 2.1777,
      "theoretical_loss": 3.355369572148189,
      "tokens_seen": 2626289664
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002076969696969697,
      "loss": 2.364,
      "theoretical_loss": 3.355362959168791,
      "tokens_seen": 2626355200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020767676767676768,
      "loss": 2.6279,
      "theoretical_loss": 3.355356346400609,
      "tokens_seen": 2626420736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020765656565656566,
      "loss": 2.3316,
      "theoretical_loss": 3.3553497338436316,
      "tokens_seen": 2626486272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020763636363636364,
      "loss": 2.7823,
      "theoretical_loss": 3.3553431214978455,
      "tokens_seen": 2626551808
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1438355,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3611056804656982,
      "objective/train/theoretical_loss": 3.3553365093632395,
      "objective/train/tokens_used": 985476576,
      "theoretical_loss": 3.3553365093632395,
      "tokens_seen": 2626617344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020761616161616163,
      "loss": 2.3885,
      "theoretical_loss": 3.3553365093632395,
      "tokens_seen": 2626617344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020759595959595959,
      "loss": 2.7687,
      "theoretical_loss": 3.355329897439801,
      "tokens_seen": 2626682880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075757575757576,
      "loss": 2.5705,
      "theoretical_loss": 3.355323285727519,
      "tokens_seen": 2626748416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020755555555555555,
      "loss": 2.6344,
      "theoretical_loss": 3.3553166742263802,
      "tokens_seen": 2626813952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020753535353535354,
      "loss": 2.3722,
      "theoretical_loss": 3.3553100629363737,
      "tokens_seen": 2626879488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020751515151515152,
      "loss": 2.4149,
      "theoretical_loss": 3.3553034518574867,
      "tokens_seen": 2626945024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020749494949494948,
      "loss": 2.6055,
      "theoretical_loss": 3.3552968409897077,
      "tokens_seen": 2627010560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074747474747475,
      "loss": 2.5863,
      "theoretical_loss": 3.3552902303330248,
      "tokens_seen": 2627076096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020745454545454545,
      "loss": 2.4718,
      "theoretical_loss": 3.355283619887425,
      "tokens_seen": 2627141632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020743434343434346,
      "loss": 2.5338,
      "theoretical_loss": 3.3552770096528977,
      "tokens_seen": 2627207168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020741414141414141,
      "loss": 2.3503,
      "theoretical_loss": 3.35527039962943,
      "tokens_seen": 2627272704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002073939393939394,
      "loss": 2.5626,
      "theoretical_loss": 3.35526378981701,
      "tokens_seen": 2627338240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020737373737373738,
      "loss": 2.3726,
      "theoretical_loss": 3.3552571802156255,
      "tokens_seen": 2627403776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020735353535353534,
      "loss": 2.309,
      "theoretical_loss": 3.355250570825265,
      "tokens_seen": 2627469312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020733333333333335,
      "loss": 2.6262,
      "theoretical_loss": 3.3552439616459164,
      "tokens_seen": 2627534848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002073131313131313,
      "loss": 2.6787,
      "theoretical_loss": 3.355237352677568,
      "tokens_seen": 2627600384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020729292929292932,
      "loss": 2.753,
      "theoretical_loss": 3.355230743920207,
      "tokens_seen": 2627665920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020727272727272727,
      "loss": 2.3945,
      "theoretical_loss": 3.3552241353738212,
      "tokens_seen": 2627731456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020725252525252526,
      "loss": 2.3233,
      "theoretical_loss": 3.3552175270383997,
      "tokens_seen": 2627796992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020723232323232324,
      "loss": 2.515,
      "theoretical_loss": 3.3552109189139303,
      "tokens_seen": 2627862528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020721212121212123,
      "loss": 2.4881,
      "theoretical_loss": 3.3552043110004,
      "tokens_seen": 2627928064
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002071919191919192,
      "loss": 2.7409,
      "theoretical_loss": 3.355197703297798,
      "tokens_seen": 2627993600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020717171717171717,
      "loss": 2.2419,
      "theoretical_loss": 3.355191095806112,
      "tokens_seen": 2628059136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020715151515151515,
      "loss": 2.8155,
      "theoretical_loss": 3.35518448852533,
      "tokens_seen": 2628124672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020713131313131313,
      "loss": 2.4631,
      "theoretical_loss": 3.355177881455439,
      "tokens_seen": 2628190208
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1438882,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7245101928710938,
      "objective/train/theoretical_loss": 3.3551712745964286,
      "objective/train/tokens_used": 987114976,
      "theoretical_loss": 3.3551712745964286,
      "tokens_seen": 2628255744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020711111111111112,
      "loss": 2.488,
      "theoretical_loss": 3.3551712745964286,
      "tokens_seen": 2628255744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070909090909091,
      "loss": 2.647,
      "theoretical_loss": 3.3551646679482854,
      "tokens_seen": 2628321280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070707070707071,
      "loss": 2.5554,
      "theoretical_loss": 3.3551580615109984,
      "tokens_seen": 2628386816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020705050505050504,
      "loss": 2.4623,
      "theoretical_loss": 3.355151455284555,
      "tokens_seen": 2628452352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020703030303030303,
      "loss": 2.3514,
      "theoretical_loss": 3.355144849268944,
      "tokens_seen": 2628517888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000207010101010101,
      "loss": 2.5696,
      "theoretical_loss": 3.355138243464153,
      "tokens_seen": 2628583424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000206989898989899,
      "loss": 2.628,
      "theoretical_loss": 3.3551316378701697,
      "tokens_seen": 2628648960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020696969696969698,
      "loss": 2.6639,
      "theoretical_loss": 3.355125032486982,
      "tokens_seen": 2628714496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020694949494949494,
      "loss": 2.4404,
      "theoretical_loss": 3.3551184273145784,
      "tokens_seen": 2628780032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020692929292929295,
      "loss": 2.5456,
      "theoretical_loss": 3.355111822352947,
      "tokens_seen": 2628845568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002069090909090909,
      "loss": 2.6819,
      "theoretical_loss": 3.355105217602076,
      "tokens_seen": 2628911104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020688888888888892,
      "loss": 2.5105,
      "theoretical_loss": 3.3550986130619522,
      "tokens_seen": 2628976640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020686868686868687,
      "loss": 2.5711,
      "theoretical_loss": 3.355092008732565,
      "tokens_seen": 2629042176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020684848484848483,
      "loss": 2.5188,
      "theoretical_loss": 3.3550854046139014,
      "tokens_seen": 2629107712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020682828282828284,
      "loss": 2.526,
      "theoretical_loss": 3.35507880070595,
      "tokens_seen": 2629173248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002068080808080808,
      "loss": 2.4181,
      "theoretical_loss": 3.355072197008699,
      "tokens_seen": 2629238784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002067878787878788,
      "loss": 2.6339,
      "theoretical_loss": 3.355065593522136,
      "tokens_seen": 2629304320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020676767676767676,
      "loss": 2.4246,
      "theoretical_loss": 3.3550589902462495,
      "tokens_seen": 2629369856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020674747474747475,
      "loss": 2.4179,
      "theoretical_loss": 3.355052387181027,
      "tokens_seen": 2629435392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020672727272727273,
      "loss": 2.5745,
      "theoretical_loss": 3.3550457843264563,
      "tokens_seen": 2629500928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020670707070707072,
      "loss": 2.4752,
      "theoretical_loss": 3.3550391816825265,
      "tokens_seen": 2629566464
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002066868686868687,
      "loss": 2.7124,
      "theoretical_loss": 3.355032579249224,
      "tokens_seen": 2629632000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020666666666666666,
      "loss": 2.5756,
      "theoretical_loss": 3.3550259770265387,
      "tokens_seen": 2629697536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020664646464646464,
      "loss": 2.6481,
      "theoretical_loss": 3.3550193750144577,
      "tokens_seen": 2629763072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020662626262626263,
      "loss": 2.5412,
      "theoretical_loss": 3.355012773212969,
      "tokens_seen": 2629828608
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1440194,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7252695560455322,
      "objective/train/theoretical_loss": 3.3550061716220605,
      "objective/train/tokens_used": 988753376,
      "theoretical_loss": 3.3550061716220605,
      "tokens_seen": 2629894144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002066060606060606,
      "loss": 2.6699,
      "theoretical_loss": 3.3550061716220605,
      "tokens_seen": 2629894144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002065858585858586,
      "loss": 2.498,
      "theoretical_loss": 3.354999570241721,
      "tokens_seen": 2629959680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020656565656565658,
      "loss": 2.5242,
      "theoretical_loss": 3.3549929690719376,
      "tokens_seen": 2630025216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020654545454545453,
      "loss": 2.8215,
      "theoretical_loss": 3.3549863681126983,
      "tokens_seen": 2630090752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020652525252525255,
      "loss": 2.5494,
      "theoretical_loss": 3.354979767363992,
      "tokens_seen": 2630156288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002065050505050505,
      "loss": 2.6365,
      "theoretical_loss": 3.3549731668258063,
      "tokens_seen": 2630221824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020648484848484849,
      "loss": 2.5794,
      "theoretical_loss": 3.3549665664981294,
      "tokens_seen": 2630287360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020646464646464647,
      "loss": 2.4213,
      "theoretical_loss": 3.354959966380949,
      "tokens_seen": 2630352896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020644444444444445,
      "loss": 2.5761,
      "theoretical_loss": 3.3549533664742532,
      "tokens_seen": 2630418432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020642424242424244,
      "loss": 2.3287,
      "theoretical_loss": 3.3549467667780304,
      "tokens_seen": 2630483968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002064040404040404,
      "loss": 2.6289,
      "theoretical_loss": 3.3549401672922685,
      "tokens_seen": 2630549504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002063838383838384,
      "loss": 2.6217,
      "theoretical_loss": 3.354933568016955,
      "tokens_seen": 2630615040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020636363636363636,
      "loss": 2.4354,
      "theoretical_loss": 3.3549269689520793,
      "tokens_seen": 2630680576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020634343434343437,
      "loss": 2.6467,
      "theoretical_loss": 3.354920370097628,
      "tokens_seen": 2630746112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020632323232323233,
      "loss": 2.4941,
      "theoretical_loss": 3.3549137714535897,
      "tokens_seen": 2630811648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002063030303030303,
      "loss": 2.3973,
      "theoretical_loss": 3.354907173019952,
      "tokens_seen": 2630877184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062828282828283,
      "loss": 2.4841,
      "theoretical_loss": 3.3549005747967042,
      "tokens_seen": 2630942720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020626262626262626,
      "loss": 2.5298,
      "theoretical_loss": 3.354893976783833,
      "tokens_seen": 2631008256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020624242424242427,
      "loss": 2.7488,
      "theoretical_loss": 3.3548873789813274,
      "tokens_seen": 2631073792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020622222222222222,
      "loss": 2.6124,
      "theoretical_loss": 3.3548807813891752,
      "tokens_seen": 2631139328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062020202020202,
      "loss": 2.6101,
      "theoretical_loss": 3.3548741840073637,
      "tokens_seen": 2631204864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061818181818182,
      "loss": 2.5685,
      "theoretical_loss": 3.354867586835882,
      "tokens_seen": 2631270400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020616161616161615,
      "loss": 2.4826,
      "theoretical_loss": 3.3548609898747177,
      "tokens_seen": 2631335936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020614141414141416,
      "loss": 2.4373,
      "theoretical_loss": 3.3548543931238592,
      "tokens_seen": 2631401472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020612121212121212,
      "loss": 2.1639,
      "theoretical_loss": 3.354847796583294,
      "tokens_seen": 2631467008
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1441640,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7747178077697754,
      "objective/train/theoretical_loss": 3.35484120025301,
      "objective/train/tokens_used": 990391776,
      "theoretical_loss": 3.35484120025301,
      "tokens_seen": 2631532544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061010101010101,
      "loss": 2.7235,
      "theoretical_loss": 3.35484120025301,
      "tokens_seen": 2631532544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020608080808080808,
      "loss": 2.5871,
      "theoretical_loss": 3.354834604132996,
      "tokens_seen": 2631598080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020606060606060607,
      "loss": 2.478,
      "theoretical_loss": 3.35482800822324,
      "tokens_seen": 2631663616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020604040404040405,
      "loss": 2.6366,
      "theoretical_loss": 3.3548214125237292,
      "tokens_seen": 2631729152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020602020202020204,
      "loss": 2.4885,
      "theoretical_loss": 3.3548148170344523,
      "tokens_seen": 2631794688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000206,
      "loss": 2.5617,
      "theoretical_loss": 3.354808221755398,
      "tokens_seen": 2631860224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020597979797979798,
      "loss": 2.513,
      "theoretical_loss": 3.354801626686553,
      "tokens_seen": 2631925760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020595959595959596,
      "loss": 2.7004,
      "theoretical_loss": 3.3547950318279063,
      "tokens_seen": 2631991296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020593939393939394,
      "loss": 2.6097,
      "theoretical_loss": 3.3547884371794456,
      "tokens_seen": 2632056832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020591919191919193,
      "loss": 2.7247,
      "theoretical_loss": 3.3547818427411595,
      "tokens_seen": 2632122368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020589898989898989,
      "loss": 2.4718,
      "theoretical_loss": 3.354775248513035,
      "tokens_seen": 2632187904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002058787878787879,
      "loss": 2.5828,
      "theoretical_loss": 3.3547686544950612,
      "tokens_seen": 2632253440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020585858585858585,
      "loss": 2.6441,
      "theoretical_loss": 3.354762060687226,
      "tokens_seen": 2632318976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020583838383838386,
      "loss": 2.4276,
      "theoretical_loss": 3.3547554670895168,
      "tokens_seen": 2632384512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020581818181818182,
      "loss": 2.6191,
      "theoretical_loss": 3.354748873701922,
      "tokens_seen": 2632450048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020579797979797978,
      "loss": 2.544,
      "theoretical_loss": 3.3547422805244307,
      "tokens_seen": 2632515584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002057777777777778,
      "loss": 2.6791,
      "theoretical_loss": 3.3547356875570293,
      "tokens_seen": 2632581120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020575757575757575,
      "loss": 2.3915,
      "theoretical_loss": 3.354729094799707,
      "tokens_seen": 2632646656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020573737373737376,
      "loss": 2.4431,
      "theoretical_loss": 3.3547225022524514,
      "tokens_seen": 2632712192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020571717171717171,
      "loss": 2.7013,
      "theoretical_loss": 3.354715909915251,
      "tokens_seen": 2632777728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002056969696969697,
      "loss": 2.5844,
      "theoretical_loss": 3.3547093177880933,
      "tokens_seen": 2632843264
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020567676767676768,
      "loss": 2.5503,
      "theoretical_loss": 3.3547027258709665,
      "tokens_seen": 2632908800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020565656565656567,
      "loss": 2.6289,
      "theoretical_loss": 3.3546961341638593,
      "tokens_seen": 2632974336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020563636363636365,
      "loss": 2.614,
      "theoretical_loss": 3.354689542666759,
      "tokens_seen": 2633039872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002056161616161616,
      "loss": 2.4605,
      "theoretical_loss": 3.354682951379654,
      "tokens_seen": 2633105408
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1442322,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3413121700286865,
      "objective/train/theoretical_loss": 3.3546763603025327,
      "objective/train/tokens_used": 992030176,
      "theoretical_loss": 3.3546763603025327,
      "tokens_seen": 2633170944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020559595959595962,
      "loss": 2.5204,
      "theoretical_loss": 3.3546763603025327,
      "tokens_seen": 2633170944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020557575757575757,
      "loss": 2.6001,
      "theoretical_loss": 3.354669769435383,
      "tokens_seen": 2633236480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020555555555555556,
      "loss": 2.5761,
      "theoretical_loss": 3.354663178778192,
      "tokens_seen": 2633302016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020553535353535354,
      "loss": 2.4491,
      "theoretical_loss": 3.3546565883309496,
      "tokens_seen": 2633367552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020551515151515153,
      "loss": 2.5183,
      "theoretical_loss": 3.3546499980936426,
      "tokens_seen": 2633433088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054949494949495,
      "loss": 2.6907,
      "theoretical_loss": 3.3546434080662593,
      "tokens_seen": 2633498624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054747474747475,
      "loss": 2.305,
      "theoretical_loss": 3.354636818248788,
      "tokens_seen": 2633564160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020545454545454545,
      "loss": 2.3771,
      "theoretical_loss": 3.354630228641217,
      "tokens_seen": 2633629696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020543434343434344,
      "loss": 2.5314,
      "theoretical_loss": 3.354623639243534,
      "tokens_seen": 2633695232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020541414141414142,
      "loss": 2.5414,
      "theoretical_loss": 3.354617050055727,
      "tokens_seen": 2633760768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053939393939394,
      "loss": 2.5888,
      "theoretical_loss": 3.354610461077784,
      "tokens_seen": 2633826304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053737373737374,
      "loss": 2.4271,
      "theoretical_loss": 3.354603872309694,
      "tokens_seen": 2633891840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020535353535353534,
      "loss": 2.5103,
      "theoretical_loss": 3.354597283751444,
      "tokens_seen": 2633957376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020533333333333336,
      "loss": 2.37,
      "theoretical_loss": 3.354590695403023,
      "tokens_seen": 2634022912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053131313131313,
      "loss": 2.6337,
      "theoretical_loss": 3.3545841072644187,
      "tokens_seen": 2634088448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002052929292929293,
      "loss": 2.4478,
      "theoretical_loss": 3.354577519335619,
      "tokens_seen": 2634153984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020527272727272728,
      "loss": 2.4778,
      "theoretical_loss": 3.3545709316166117,
      "tokens_seen": 2634219520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020525252525252524,
      "loss": 2.4299,
      "theoretical_loss": 3.354564344107386,
      "tokens_seen": 2634285056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020523232323232325,
      "loss": 2.3935,
      "theoretical_loss": 3.354557756807929,
      "tokens_seen": 2634350592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002052121212121212,
      "loss": 2.6921,
      "theoretical_loss": 3.3545511697182295,
      "tokens_seen": 2634416128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020519191919191922,
      "loss": 2.6472,
      "theoretical_loss": 3.354544582838275,
      "tokens_seen": 2634481664
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020517171717171717,
      "loss": 2.6854,
      "theoretical_loss": 3.354537996168054,
      "tokens_seen": 2634547200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020515151515151516,
      "loss": 2.5316,
      "theoretical_loss": 3.3545314097075547,
      "tokens_seen": 2634612736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020513131313131314,
      "loss": 2.6331,
      "theoretical_loss": 3.3545248234567646,
      "tokens_seen": 2634678272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002051111111111111,
      "loss": 2.6417,
      "theoretical_loss": 3.354518237415672,
      "tokens_seen": 2634743808
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1443625,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.782468795776367,
      "objective/train/theoretical_loss": 3.3545116515842657,
      "objective/train/tokens_used": 993668576,
      "theoretical_loss": 3.3545116515842657,
      "tokens_seen": 2634809344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002050909090909091,
      "loss": 2.6177,
      "theoretical_loss": 3.3545116515842657,
      "tokens_seen": 2634809344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020507070707070707,
      "loss": 2.5574,
      "theoretical_loss": 3.3545050659625333,
      "tokens_seen": 2634874880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020505050505050505,
      "loss": 2.5084,
      "theoretical_loss": 3.354498480550463,
      "tokens_seen": 2634940416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020503030303030303,
      "loss": 2.6944,
      "theoretical_loss": 3.3544918953480423,
      "tokens_seen": 2635005952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020501010101010102,
      "loss": 2.3932,
      "theoretical_loss": 3.35448531035526,
      "tokens_seen": 2635071488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000204989898989899,
      "loss": 2.4214,
      "theoretical_loss": 3.3544787255721045,
      "tokens_seen": 2635137024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020496969696969699,
      "loss": 2.8083,
      "theoretical_loss": 3.354472140998563,
      "tokens_seen": 2635202560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020494949494949494,
      "loss": 2.483,
      "theoretical_loss": 3.3544655566346244,
      "tokens_seen": 2635268096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020492929292929293,
      "loss": 2.6708,
      "theoretical_loss": 3.3544589724802765,
      "tokens_seen": 2635333632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002049090909090909,
      "loss": 2.3091,
      "theoretical_loss": 3.354452388535507,
      "tokens_seen": 2635399168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002048888888888889,
      "loss": 2.3091,
      "theoretical_loss": 3.3544458048003047,
      "tokens_seen": 2635464704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020486868686868688,
      "loss": 2.4276,
      "theoretical_loss": 3.3544392212746574,
      "tokens_seen": 2635530240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020484848484848483,
      "loss": 2.4445,
      "theoretical_loss": 3.3544326379585536,
      "tokens_seen": 2635595776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020482828282828285,
      "loss": 2.5868,
      "theoretical_loss": 3.354426054851981,
      "tokens_seen": 2635661312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002048080808080808,
      "loss": 2.4394,
      "theoretical_loss": 3.3544194719549276,
      "tokens_seen": 2635726848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020478787878787881,
      "loss": 2.6389,
      "theoretical_loss": 3.354412889267382,
      "tokens_seen": 2635792384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020476767676767677,
      "loss": 2.3571,
      "theoretical_loss": 3.3544063067893317,
      "tokens_seen": 2635857920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020474747474747473,
      "loss": 2.5287,
      "theoretical_loss": 3.354399724520765,
      "tokens_seen": 2635923456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020472727272727274,
      "loss": 2.4407,
      "theoretical_loss": 3.354393142461671,
      "tokens_seen": 2635988992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002047070707070707,
      "loss": 2.4813,
      "theoretical_loss": 3.3543865606120367,
      "tokens_seen": 2636054528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046868686868687,
      "loss": 2.5608,
      "theoretical_loss": 3.3543799789718505,
      "tokens_seen": 2636120064
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020466666666666666,
      "loss": 2.4228,
      "theoretical_loss": 3.3543733975411003,
      "tokens_seen": 2636185600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020464646464646467,
      "loss": 2.9393,
      "theoretical_loss": 3.3543668163197746,
      "tokens_seen": 2636251136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020462626262626263,
      "loss": 2.6287,
      "theoretical_loss": 3.354360235307862,
      "tokens_seen": 2636316672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046060606060606,
      "loss": 2.523,
      "theoretical_loss": 3.3543536545053496,
      "tokens_seen": 2636382208
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1444226,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1333506107330322,
      "objective/train/theoretical_loss": 3.3543470739122263,
      "objective/train/tokens_used": 995306976,
      "theoretical_loss": 3.3543470739122263,
      "tokens_seen": 2636447744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002045858585858586,
      "loss": 2.4369,
      "theoretical_loss": 3.3543470739122263,
      "tokens_seen": 2636447744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020456565656565656,
      "loss": 2.4366,
      "theoretical_loss": 3.3543404935284795,
      "tokens_seen": 2636513280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020454545454545457,
      "loss": 2.4689,
      "theoretical_loss": 3.354333913354098,
      "tokens_seen": 2636578816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020452525252525252,
      "loss": 2.7063,
      "theoretical_loss": 3.3543273333890697,
      "tokens_seen": 2636644352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002045050505050505,
      "loss": 2.513,
      "theoretical_loss": 3.354320753633383,
      "tokens_seen": 2636709888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044848484848485,
      "loss": 2.618,
      "theoretical_loss": 3.3543141740870257,
      "tokens_seen": 2636775424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020446464646464648,
      "loss": 2.6782,
      "theoretical_loss": 3.354307594749986,
      "tokens_seen": 2636840960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020444444444444446,
      "loss": 2.4553,
      "theoretical_loss": 3.354301015622252,
      "tokens_seen": 2636906496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020442424242424242,
      "loss": 2.5374,
      "theoretical_loss": 3.3542944367038117,
      "tokens_seen": 2636972032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044040404040404,
      "loss": 2.535,
      "theoretical_loss": 3.354287857994654,
      "tokens_seen": 2637037568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020438383838383838,
      "loss": 2.4309,
      "theoretical_loss": 3.354281279494766,
      "tokens_seen": 2637103104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020436363636363637,
      "loss": 2.8039,
      "theoretical_loss": 3.3542747012041363,
      "tokens_seen": 2637168640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020434343434343435,
      "loss": 2.672,
      "theoretical_loss": 3.354268123122753,
      "tokens_seen": 2637234176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020432323232323234,
      "loss": 2.3302,
      "theoretical_loss": 3.354261545250605,
      "tokens_seen": 2637299712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002043030303030303,
      "loss": 2.4494,
      "theoretical_loss": 3.3542549675876794,
      "tokens_seen": 2637365248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002042828282828283,
      "loss": 2.3474,
      "theoretical_loss": 3.3542483901339644,
      "tokens_seen": 2637430784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020426262626262626,
      "loss": 2.6203,
      "theoretical_loss": 3.3542418128894487,
      "tokens_seen": 2637496320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020424242424242425,
      "loss": 2.7205,
      "theoretical_loss": 3.3542352358541203,
      "tokens_seen": 2637561856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020422222222222223,
      "loss": 2.3857,
      "theoretical_loss": 3.3542286590279673,
      "tokens_seen": 2637627392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020420202020202019,
      "loss": 2.6463,
      "theoretical_loss": 3.3542220824109776,
      "tokens_seen": 2637692928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041818181818182,
      "loss": 2.3769,
      "theoretical_loss": 3.3542155060031393,
      "tokens_seen": 2637758464
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020416161616161615,
      "loss": 2.4234,
      "theoretical_loss": 3.3542089298044413,
      "tokens_seen": 2637824000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020414141414141416,
      "loss": 2.714,
      "theoretical_loss": 3.354202353814871,
      "tokens_seen": 2637889536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020412121212121212,
      "loss": 2.5929,
      "theoretical_loss": 3.354195778034417,
      "tokens_seen": 2637955072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041010101010101,
      "loss": 2.4963,
      "theoretical_loss": 3.354189202463067,
      "tokens_seen": 2638020608
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1445429,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.842092275619507,
      "objective/train/theoretical_loss": 3.3541826271008097,
      "objective/train/tokens_used": 996945376,
      "theoretical_loss": 3.3541826271008097,
      "tokens_seen": 2638086144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002040808080808081,
      "loss": 2.5018,
      "theoretical_loss": 3.3541826271008097,
      "tokens_seen": 2638086144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020406060606060605,
      "loss": 2.4667,
      "theoretical_loss": 3.3541760519476327,
      "tokens_seen": 2638151680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020404040404040406,
      "loss": 2.4265,
      "theoretical_loss": 3.354169477003525,
      "tokens_seen": 2638217216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020402020202020201,
      "loss": 2.6254,
      "theoretical_loss": 3.354162902268474,
      "tokens_seen": 2638282752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000204,
      "loss": 2.5333,
      "theoretical_loss": 3.3541563277424675,
      "tokens_seen": 2638348288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020397979797979798,
      "loss": 2.5255,
      "theoretical_loss": 3.354149753425495,
      "tokens_seen": 2638413824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020395959595959597,
      "loss": 2.4482,
      "theoretical_loss": 3.354143179317543,
      "tokens_seen": 2638479360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020393939393939395,
      "loss": 2.7569,
      "theoretical_loss": 3.354136605418601,
      "tokens_seen": 2638544896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020391919191919193,
      "loss": 2.4789,
      "theoretical_loss": 3.354130031728657,
      "tokens_seen": 2638610432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002038989898989899,
      "loss": 2.509,
      "theoretical_loss": 3.354123458247699,
      "tokens_seen": 2638675968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020387878787878788,
      "loss": 2.3724,
      "theoretical_loss": 3.3541168849757144,
      "tokens_seen": 2638741504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020385858585858586,
      "loss": 2.5621,
      "theoretical_loss": 3.3541103119126925,
      "tokens_seen": 2638807040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020383838383838384,
      "loss": 2.4991,
      "theoretical_loss": 3.354103739058621,
      "tokens_seen": 2638872576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020381818181818183,
      "loss": 2.5027,
      "theoretical_loss": 3.3540971664134873,
      "tokens_seen": 2638938112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037979797979798,
      "loss": 2.3328,
      "theoretical_loss": 3.354090593977281,
      "tokens_seen": 2639003648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037777777777778,
      "loss": 2.5546,
      "theoretical_loss": 3.3540840217499897,
      "tokens_seen": 2639069184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020375757575757575,
      "loss": 2.4876,
      "theoretical_loss": 3.3540774497316006,
      "tokens_seen": 2639134720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020373737373737374,
      "loss": 2.4025,
      "theoretical_loss": 3.3540708779221036,
      "tokens_seen": 2639200256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020371717171717172,
      "loss": 2.6314,
      "theoretical_loss": 3.3540643063214857,
      "tokens_seen": 2639265792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036969696969697,
      "loss": 2.4343,
      "theoretical_loss": 3.3540577349297354,
      "tokens_seen": 2639331328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036767676767677,
      "loss": 2.4641,
      "theoretical_loss": 3.3540511637468406,
      "tokens_seen": 2639396864
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020365656565656564,
      "loss": 2.6703,
      "theoretical_loss": 3.35404459277279,
      "tokens_seen": 2639462400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020363636363636366,
      "loss": 2.6479,
      "theoretical_loss": 3.354038022007572,
      "tokens_seen": 2639527936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036161616161616,
      "loss": 2.549,
      "theoretical_loss": 3.3540314514511733,
      "tokens_seen": 2639593472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020359595959595962,
      "loss": 2.5515,
      "theoretical_loss": 3.3540248811035838,
      "tokens_seen": 2639659008
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1446170,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.584564685821533,
      "objective/train/theoretical_loss": 3.3540183109647908,
      "objective/train/tokens_used": 998583776,
      "theoretical_loss": 3.3540183109647908,
      "tokens_seen": 2639724544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020357575757575758,
      "loss": 2.6946,
      "theoretical_loss": 3.3540183109647908,
      "tokens_seen": 2639724544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020355555555555554,
      "loss": 2.47,
      "theoretical_loss": 3.3540117410347823,
      "tokens_seen": 2639790080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020353535353535355,
      "loss": 2.4332,
      "theoretical_loss": 3.3540051713135473,
      "tokens_seen": 2639855616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002035151515151515,
      "loss": 2.4539,
      "theoretical_loss": 3.353998601801073,
      "tokens_seen": 2639921152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020349494949494952,
      "loss": 2.4892,
      "theoretical_loss": 3.3539920324973482,
      "tokens_seen": 2639986688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020347474747474747,
      "loss": 2.4403,
      "theoretical_loss": 3.3539854634023607,
      "tokens_seen": 2640052224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020345454545454546,
      "loss": 2.5336,
      "theoretical_loss": 3.3539788945160995,
      "tokens_seen": 2640117760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020343434343434344,
      "loss": 2.4967,
      "theoretical_loss": 3.3539723258385523,
      "tokens_seen": 2640183296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020341414141414142,
      "loss": 2.4621,
      "theoretical_loss": 3.3539657573697066,
      "tokens_seen": 2640248832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002033939393939394,
      "loss": 2.4746,
      "theoretical_loss": 3.3539591891095517,
      "tokens_seen": 2640314368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020337373737373737,
      "loss": 2.7718,
      "theoretical_loss": 3.3539526210580752,
      "tokens_seen": 2640379904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020335353535353535,
      "loss": 2.5249,
      "theoretical_loss": 3.353946053215265,
      "tokens_seen": 2640445440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020333333333333333,
      "loss": 2.5948,
      "theoretical_loss": 3.3539394855811104,
      "tokens_seen": 2640510976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020331313131313132,
      "loss": 2.5381,
      "theoretical_loss": 3.3539329181555986,
      "tokens_seen": 2640576512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002032929292929293,
      "loss": 2.6449,
      "theoretical_loss": 3.3539263509387176,
      "tokens_seen": 2640642048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020327272727272729,
      "loss": 2.3962,
      "theoretical_loss": 3.353919783930457,
      "tokens_seen": 2640707584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020325252525252524,
      "loss": 2.3859,
      "theoretical_loss": 3.3539132171308035,
      "tokens_seen": 2640773120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020323232323232325,
      "loss": 2.5057,
      "theoretical_loss": 3.3539066505397455,
      "tokens_seen": 2640838656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002032121212121212,
      "loss": 2.6876,
      "theoretical_loss": 3.353900084157272,
      "tokens_seen": 2640904192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002031919191919192,
      "loss": 2.666,
      "theoretical_loss": 3.3538935179833707,
      "tokens_seen": 2640969728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020317171717171718,
      "loss": 2.4959,
      "theoretical_loss": 3.35388695201803,
      "tokens_seen": 2641035264
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020315151515151514,
      "loss": 2.6454,
      "theoretical_loss": 3.3538803862612374,
      "tokens_seen": 2641100800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020313131313131315,
      "loss": 2.3571,
      "theoretical_loss": 3.353873820712982,
      "tokens_seen": 2641166336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002031111111111111,
      "loss": 2.341,
      "theoretical_loss": 3.3538672553732516,
      "tokens_seen": 2641231872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020309090909090911,
      "loss": 2.3788,
      "theoretical_loss": 3.3538606902420347,
      "tokens_seen": 2641297408
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1447240,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6527767181396484,
      "objective/train/theoretical_loss": 3.3538541253193195,
      "objective/train/tokens_used": 1000222176,
      "theoretical_loss": 3.3538541253193195,
      "tokens_seen": 2641362944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020307070707070707,
      "loss": 2.5712,
      "theoretical_loss": 3.3538541253193195,
      "tokens_seen": 2641362944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020305050505050503,
      "loss": 2.5421,
      "theoretical_loss": 3.353847560605093,
      "tokens_seen": 2641428480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020303030303030304,
      "loss": 2.2935,
      "theoretical_loss": 3.353840996099345,
      "tokens_seen": 2641494016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000203010101010101,
      "loss": 2.3186,
      "theoretical_loss": 3.3538344318020634,
      "tokens_seen": 2641559552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202989898989899,
      "loss": 2.6663,
      "theoretical_loss": 3.3538278677132354,
      "tokens_seen": 2641625088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020296969696969696,
      "loss": 2.5128,
      "theoretical_loss": 3.3538213038328504,
      "tokens_seen": 2641690624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020294949494949497,
      "loss": 2.4634,
      "theoretical_loss": 3.353814740160896,
      "tokens_seen": 2641756160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020292929292929293,
      "loss": 2.5404,
      "theoretical_loss": 3.3538081766973606,
      "tokens_seen": 2641821696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020290909090909092,
      "loss": 2.4918,
      "theoretical_loss": 3.353801613442232,
      "tokens_seen": 2641887232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002028888888888889,
      "loss": 2.4896,
      "theoretical_loss": 3.353795050395499,
      "tokens_seen": 2641952768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020286868686868686,
      "loss": 2.4449,
      "theoretical_loss": 3.3537884875571495,
      "tokens_seen": 2642018304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020284848484848487,
      "loss": 2.6791,
      "theoretical_loss": 3.353781924927172,
      "tokens_seen": 2642083840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020282828282828282,
      "loss": 2.4065,
      "theoretical_loss": 3.353775362505554,
      "tokens_seen": 2642149376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002028080808080808,
      "loss": 2.4809,
      "theoretical_loss": 3.3537688002922845,
      "tokens_seen": 2642214912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002027878787878788,
      "loss": 2.7207,
      "theoretical_loss": 3.3537622382873513,
      "tokens_seen": 2642280448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020276767676767678,
      "loss": 2.6247,
      "theoretical_loss": 3.353755676490743,
      "tokens_seen": 2642345984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020274747474747476,
      "loss": 2.8295,
      "theoretical_loss": 3.353749114902447,
      "tokens_seen": 2642411520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020272727272727274,
      "loss": 2.3515,
      "theoretical_loss": 3.353742553522453,
      "tokens_seen": 2642477056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002027070707070707,
      "loss": 2.5542,
      "theoretical_loss": 3.3537359923507477,
      "tokens_seen": 2642542592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020268686868686868,
      "loss": 2.363,
      "theoretical_loss": 3.35372943138732,
      "tokens_seen": 2642608128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020266666666666667,
      "loss": 2.4245,
      "theoretical_loss": 3.3537228706321582,
      "tokens_seen": 2642673664
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020264646464646465,
      "loss": 2.4782,
      "theoretical_loss": 3.3537163100852503,
      "tokens_seen": 2642739200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020262626262626264,
      "loss": 2.7439,
      "theoretical_loss": 3.353709749746584,
      "tokens_seen": 2642804736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002026060606060606,
      "loss": 2.4381,
      "theoretical_loss": 3.353703189616149,
      "tokens_seen": 2642870272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002025858585858586,
      "loss": 2.2415,
      "theoretical_loss": 3.3536966296939323,
      "tokens_seen": 2642935808
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1447769,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.581217050552368,
      "objective/train/theoretical_loss": 3.3536900699799226,
      "objective/train/tokens_used": 1001860576,
      "theoretical_loss": 3.3536900699799226,
      "tokens_seen": 2643001344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020256565656565656,
      "loss": 2.4888,
      "theoretical_loss": 3.3536900699799226,
      "tokens_seen": 2643001344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020254545454545457,
      "loss": 2.316,
      "theoretical_loss": 3.353683510474108,
      "tokens_seen": 2643066880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020252525252525253,
      "loss": 2.481,
      "theoretical_loss": 3.353676951176477,
      "tokens_seen": 2643132416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020250505050505049,
      "loss": 2.2993,
      "theoretical_loss": 3.353670392087017,
      "tokens_seen": 2643197952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024848484848485,
      "loss": 2.4659,
      "theoretical_loss": 3.353663833205717,
      "tokens_seen": 2643263488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020246464646464645,
      "loss": 2.41,
      "theoretical_loss": 3.3536572745325652,
      "tokens_seen": 2643329024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020244444444444447,
      "loss": 2.4892,
      "theoretical_loss": 3.3536507160675493,
      "tokens_seen": 2643394560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020242424242424242,
      "loss": 2.5475,
      "theoretical_loss": 3.3536441578106584,
      "tokens_seen": 2643460096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024040404040404,
      "loss": 2.46,
      "theoretical_loss": 3.35363759976188,
      "tokens_seen": 2643525632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023838383838384,
      "loss": 2.4332,
      "theoretical_loss": 3.3536310419212025,
      "tokens_seen": 2643591168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020236363636363637,
      "loss": 2.4998,
      "theoretical_loss": 3.3536244842886145,
      "tokens_seen": 2643656704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020234343434343436,
      "loss": 2.3399,
      "theoretical_loss": 3.3536179268641035,
      "tokens_seen": 2643722240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020232323232323231,
      "loss": 2.5412,
      "theoretical_loss": 3.3536113696476586,
      "tokens_seen": 2643787776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023030303030303,
      "loss": 2.4482,
      "theoretical_loss": 3.3536048126392677,
      "tokens_seen": 2643853312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020228282828282828,
      "loss": 2.4939,
      "theoretical_loss": 3.353598255838919,
      "tokens_seen": 2643918848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020226262626262627,
      "loss": 2.5698,
      "theoretical_loss": 3.3535916992466,
      "tokens_seen": 2643984384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020224242424242425,
      "loss": 2.3163,
      "theoretical_loss": 3.3535851428623005,
      "tokens_seen": 2644049920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020222222222222223,
      "loss": 2.5823,
      "theoretical_loss": 3.3535785866860075,
      "tokens_seen": 2644115456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002022020202020202,
      "loss": 2.3124,
      "theoretical_loss": 3.3535720307177095,
      "tokens_seen": 2644180992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020218181818181818,
      "loss": 2.4176,
      "theoretical_loss": 3.3535654749573953,
      "tokens_seen": 2644246528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020216161616161616,
      "loss": 2.6747,
      "theoretical_loss": 3.3535589194050526,
      "tokens_seen": 2644312064
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020214141414141414,
      "loss": 2.7327,
      "theoretical_loss": 3.3535523640606697,
      "tokens_seen": 2644377600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020212121212121213,
      "loss": 2.4384,
      "theoretical_loss": 3.353545808924235,
      "tokens_seen": 2644443136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002021010101010101,
      "loss": 2.5577,
      "theoretical_loss": 3.353539253995737,
      "tokens_seen": 2644508672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002020808080808081,
      "loss": 2.6404,
      "theoretical_loss": 3.353532699275163,
      "tokens_seen": 2644574208
    },
    {
      "debugging/Self-BLEU-5": 0.3806857716898066,
      "debugging/distinct-1-grams": 0.7114641316153453,
      "debugging/distinct-2-grams": 0.8748620261829383,
      "debugging/entropy-1-grams": 5.322027392293219,
      "debugging/entropy-2-grams": 5.946451670482928,
      "debugging/length": 544.0,
      "debugging/num_segments": 6,
      "epoch": 0.6,
      "objective/train/docs_used": 1449122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.866844892501831,
      "objective/train/theoretical_loss": 3.3535261447625024,
      "objective/train/tokens_used": 1003498976,
      "theoretical_loss": 3.3535261447625024,
      "tokens_seen": 2644639744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020206060606060605,
      "loss": 2.4781,
      "theoretical_loss": 3.3535261447625024,
      "tokens_seen": 2644639744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020204040404040406,
      "loss": 2.5342,
      "theoretical_loss": 3.353519590457743,
      "tokens_seen": 2644705280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020202020202020202,
      "loss": 2.3997,
      "theoretical_loss": 3.353513036360873,
      "tokens_seen": 2644770816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202,
      "loss": 2.5918,
      "theoretical_loss": 3.3535064824718805,
      "tokens_seen": 2644836352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201979797979798,
      "loss": 2.2975,
      "theoretical_loss": 3.353499928790754,
      "tokens_seen": 2644901888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020195959595959594,
      "loss": 2.4527,
      "theoretical_loss": 3.3534933753174814,
      "tokens_seen": 2644967424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020193939393939396,
      "loss": 2.4123,
      "theoretical_loss": 3.3534868220520515,
      "tokens_seen": 2645032960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002019191919191919,
      "loss": 2.5549,
      "theoretical_loss": 3.353480268994452,
      "tokens_seen": 2645098496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020189898989898992,
      "loss": 2.281,
      "theoretical_loss": 3.3534737161446717,
      "tokens_seen": 2645164032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020187878787878788,
      "loss": 2.526,
      "theoretical_loss": 3.3534671635026987,
      "tokens_seen": 2645229568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020185858585858586,
      "loss": 2.4262,
      "theoretical_loss": 3.3534606110685212,
      "tokens_seen": 2645295104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020183838383838385,
      "loss": 2.4313,
      "theoretical_loss": 3.353454058842127,
      "tokens_seen": 2645360640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002018181818181818,
      "loss": 2.4439,
      "theoretical_loss": 3.353447506823505,
      "tokens_seen": 2645426176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020179797979797982,
      "loss": 2.4488,
      "theoretical_loss": 3.3534409550126436,
      "tokens_seen": 2645491712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020177777777777777,
      "loss": 2.5502,
      "theoretical_loss": 3.3534344034095303,
      "tokens_seen": 2645557248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020175757575757576,
      "loss": 2.3812,
      "theoretical_loss": 3.3534278520141543,
      "tokens_seen": 2645622784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020173737373737374,
      "loss": 2.2579,
      "theoretical_loss": 3.353421300826503,
      "tokens_seen": 2645688320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020171717171717173,
      "loss": 2.4524,
      "theoretical_loss": 3.353414749846565,
      "tokens_seen": 2645753856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016969696969697,
      "loss": 2.3762,
      "theoretical_loss": 3.3534081990743285,
      "tokens_seen": 2645819392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016767676767677,
      "loss": 2.4941,
      "theoretical_loss": 3.353401648509782,
      "tokens_seen": 2645884928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020165656565656565,
      "loss": 2.431,
      "theoretical_loss": 3.353395098152914,
      "tokens_seen": 2645950464
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020163636363636363,
      "loss": 2.5094,
      "theoretical_loss": 3.353388548003712,
      "tokens_seen": 2646016000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020161616161616162,
      "loss": 2.4778,
      "theoretical_loss": 3.353381998062165,
      "tokens_seen": 2646081536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002015959595959596,
      "loss": 2.3115,
      "theoretical_loss": 3.3533754483282605,
      "tokens_seen": 2646147072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020157575757575759,
      "loss": 2.5795,
      "theoretical_loss": 3.3533688988019876,
      "tokens_seen": 2646212608
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1449979,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7366762161254883,
      "objective/train/theoretical_loss": 3.3533623494833344,
      "objective/train/tokens_used": 1005137376,
      "theoretical_loss": 3.3533623494833344,
      "tokens_seen": 2646278144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020155555555555554,
      "loss": 2.4919,
      "theoretical_loss": 3.3533623494833344,
      "tokens_seen": 2646278144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020153535353535355,
      "loss": 2.771,
      "theoretical_loss": 3.3533558003722885,
      "tokens_seen": 2646343680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002015151515151515,
      "loss": 2.3588,
      "theoretical_loss": 3.353349251468839,
      "tokens_seen": 2646409216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020149494949494952,
      "loss": 2.5566,
      "theoretical_loss": 3.3533427027729736,
      "tokens_seen": 2646474752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020147474747474748,
      "loss": 2.5956,
      "theoretical_loss": 3.353336154284681,
      "tokens_seen": 2646540288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020145454545454544,
      "loss": 2.6503,
      "theoretical_loss": 3.3533296060039492,
      "tokens_seen": 2646605824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020143434343434345,
      "loss": 2.3714,
      "theoretical_loss": 3.353323057930767,
      "tokens_seen": 2646671360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002014141414141414,
      "loss": 2.564,
      "theoretical_loss": 3.3533165100651217,
      "tokens_seen": 2646736896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020139393939393941,
      "loss": 2.5456,
      "theoretical_loss": 3.3533099624070024,
      "tokens_seen": 2646802432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020137373737373737,
      "loss": 2.5628,
      "theoretical_loss": 3.353303414956397,
      "tokens_seen": 2646867968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020135353535353536,
      "loss": 2.5234,
      "theoretical_loss": 3.3532968677132944,
      "tokens_seen": 2646933504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020133333333333334,
      "loss": 2.5044,
      "theoretical_loss": 3.353290320677682,
      "tokens_seen": 2646999040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002013131313131313,
      "loss": 2.4042,
      "theoretical_loss": 3.3532837738495487,
      "tokens_seen": 2647064576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002012929292929293,
      "loss": 2.3805,
      "theoretical_loss": 3.353277227228882,
      "tokens_seen": 2647130112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020127272727272726,
      "loss": 2.5289,
      "theoretical_loss": 3.3532706808156716,
      "tokens_seen": 2647195648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020125252525252528,
      "loss": 2.4921,
      "theoretical_loss": 3.353264134609905,
      "tokens_seen": 2647261184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020123232323232323,
      "loss": 2.5631,
      "theoretical_loss": 3.3532575886115694,
      "tokens_seen": 2647326720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020121212121212122,
      "loss": 2.3107,
      "theoretical_loss": 3.3532510428206552,
      "tokens_seen": 2647392256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002011919191919192,
      "loss": 2.5545,
      "theoretical_loss": 3.3532444972371493,
      "tokens_seen": 2647457792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020117171717171718,
      "loss": 2.5281,
      "theoretical_loss": 3.3532379518610402,
      "tokens_seen": 2647523328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020115151515151517,
      "loss": 2.5125,
      "theoretical_loss": 3.3532314066923163,
      "tokens_seen": 2647588864
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020113131313131312,
      "loss": 2.5857,
      "theoretical_loss": 3.353224861730966,
      "tokens_seen": 2647654400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002011111111111111,
      "loss": 2.4776,
      "theoretical_loss": 3.3532183169769776,
      "tokens_seen": 2647719936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002010909090909091,
      "loss": 2.5633,
      "theoretical_loss": 3.3532117724303396,
      "tokens_seen": 2647785472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020107070707070708,
      "loss": 2.3298,
      "theoretical_loss": 3.35320522809104,
      "tokens_seen": 2647851008
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1450923,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.731349468231201,
      "objective/train/theoretical_loss": 3.3531986839590666,
      "objective/train/tokens_used": 1006775776,
      "theoretical_loss": 3.3531986839590666,
      "tokens_seen": 2647916544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020105050505050506,
      "loss": 2.7316,
      "theoretical_loss": 3.3531986839590666,
      "tokens_seen": 2647916544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020103030303030304,
      "loss": 2.5103,
      "theoretical_loss": 3.3531921400344085,
      "tokens_seen": 2647982080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201010101010101,
      "loss": 2.463,
      "theoretical_loss": 3.353185596317054,
      "tokens_seen": 2648047616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200989898989899,
      "loss": 2.3869,
      "theoretical_loss": 3.3531790528069907,
      "tokens_seen": 2648113152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020096969696969697,
      "loss": 2.6476,
      "theoretical_loss": 3.3531725095042075,
      "tokens_seen": 2648178688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020094949494949495,
      "loss": 2.43,
      "theoretical_loss": 3.3531659664086924,
      "tokens_seen": 2648244224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020092929292929294,
      "loss": 2.515,
      "theoretical_loss": 3.353159423520434,
      "tokens_seen": 2648309760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002009090909090909,
      "loss": 2.581,
      "theoretical_loss": 3.3531528808394206,
      "tokens_seen": 2648375296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002008888888888889,
      "loss": 2.5799,
      "theoretical_loss": 3.35314633836564,
      "tokens_seen": 2648440832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020086868686868686,
      "loss": 2.3902,
      "theoretical_loss": 3.353139796099081,
      "tokens_seen": 2648506368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020084848484848487,
      "loss": 2.3964,
      "theoretical_loss": 3.353133254039732,
      "tokens_seen": 2648571904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020082828282828283,
      "loss": 2.6709,
      "theoretical_loss": 3.3531267121875805,
      "tokens_seen": 2648637440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020080808080808081,
      "loss": 2.4107,
      "theoretical_loss": 3.353120170542616,
      "tokens_seen": 2648702976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007878787878788,
      "loss": 2.4521,
      "theoretical_loss": 3.353113629104826,
      "tokens_seen": 2648768512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020076767676767675,
      "loss": 2.4593,
      "theoretical_loss": 3.353107087874199,
      "tokens_seen": 2648834048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020074747474747477,
      "loss": 2.4129,
      "theoretical_loss": 3.3531005468507233,
      "tokens_seen": 2648899584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020072727272727272,
      "loss": 2.6882,
      "theoretical_loss": 3.3530940060343872,
      "tokens_seen": 2648965120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007070707070707,
      "loss": 2.489,
      "theoretical_loss": 3.353087465425179,
      "tokens_seen": 2649030656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002006868686868687,
      "loss": 2.5607,
      "theoretical_loss": 3.3530809250230877,
      "tokens_seen": 2649096192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020066666666666667,
      "loss": 2.5956,
      "theoretical_loss": 3.3530743848281004,
      "tokens_seen": 2649161728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020064646464646466,
      "loss": 2.36,
      "theoretical_loss": 3.353067844840206,
      "tokens_seen": 2649227264
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020062626262626264,
      "loss": 2.4052,
      "theoretical_loss": 3.353061305059393,
      "tokens_seen": 2649292800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002006060606060606,
      "loss": 2.5173,
      "theoretical_loss": 3.3530547654856497,
      "tokens_seen": 2649358336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020058585858585858,
      "loss": 2.4487,
      "theoretical_loss": 3.3530482261189642,
      "tokens_seen": 2649423872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020056565656565657,
      "loss": 2.3324,
      "theoretical_loss": 3.353041686959325,
      "tokens_seen": 2649489408
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1451404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.839794635772705,
      "objective/train/theoretical_loss": 3.35303514800672,
      "objective/train/tokens_used": 1008414176,
      "theoretical_loss": 3.35303514800672,
      "tokens_seen": 2649554944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020054545454545455,
      "loss": 2.6139,
      "theoretical_loss": 3.35303514800672,
      "tokens_seen": 2649554944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020052525252525254,
      "loss": 2.7389,
      "theoretical_loss": 3.353028609261138,
      "tokens_seen": 2649620480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002005050505050505,
      "loss": 2.3895,
      "theoretical_loss": 3.3530220707225675,
      "tokens_seen": 2649686016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004848484848485,
      "loss": 2.6618,
      "theoretical_loss": 3.353015532390996,
      "tokens_seen": 2649751552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020046464646464646,
      "loss": 2.3531,
      "theoretical_loss": 3.3530089942664123,
      "tokens_seen": 2649817088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020044444444444444,
      "loss": 2.4143,
      "theoretical_loss": 3.353002456348805,
      "tokens_seen": 2649882624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020042424242424243,
      "loss": 2.5058,
      "theoretical_loss": 3.3529959186381624,
      "tokens_seen": 2649948160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004040404040404,
      "loss": 2.3157,
      "theoretical_loss": 3.3529893811344724,
      "tokens_seen": 2650013696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002003838383838384,
      "loss": 2.4094,
      "theoretical_loss": 3.3529828438377236,
      "tokens_seen": 2650079232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020036363636363635,
      "loss": 2.5233,
      "theoretical_loss": 3.3529763067479044,
      "tokens_seen": 2650144768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020034343434343436,
      "loss": 2.4015,
      "theoretical_loss": 3.3529697698650027,
      "tokens_seen": 2650210304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020032323232323232,
      "loss": 2.6235,
      "theoretical_loss": 3.352963233189007,
      "tokens_seen": 2650275840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020030303030303033,
      "loss": 2.6242,
      "theoretical_loss": 3.352956696719906,
      "tokens_seen": 2650341376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002002828282828283,
      "loss": 2.6108,
      "theoretical_loss": 3.352950160457688,
      "tokens_seen": 2650406912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020026262626262625,
      "loss": 2.3372,
      "theoretical_loss": 3.352943624402341,
      "tokens_seen": 2650472448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020024242424242426,
      "loss": 2.5075,
      "theoretical_loss": 3.352937088553854,
      "tokens_seen": 2650537984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002002222222222222,
      "loss": 2.5586,
      "theoretical_loss": 3.352930552912214,
      "tokens_seen": 2650603520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020020202020202022,
      "loss": 2.6873,
      "theoretical_loss": 3.3529240174774104,
      "tokens_seen": 2650669056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020018181818181818,
      "loss": 2.5061,
      "theoretical_loss": 3.3529174822494316,
      "tokens_seen": 2650734592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020016161616161616,
      "loss": 2.4104,
      "theoretical_loss": 3.352910947228265,
      "tokens_seen": 2650800128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020014141414141415,
      "loss": 2.5027,
      "theoretical_loss": 3.3529044124139005,
      "tokens_seen": 2650865664
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020012121212121213,
      "loss": 2.5356,
      "theoretical_loss": 3.3528978778063254,
      "tokens_seen": 2650931200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020010101010101012,
      "loss": 2.5219,
      "theoretical_loss": 3.3528913434055276,
      "tokens_seen": 2650996736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020008080808080807,
      "loss": 2.6471,
      "theoretical_loss": 3.3528848092114965,
      "tokens_seen": 2651062272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020006060606060606,
      "loss": 2.5698,
      "theoretical_loss": 3.35287827522422,
      "tokens_seen": 2651127808
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1452746,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.292234420776367,
      "objective/train/theoretical_loss": 3.352871741443686,
      "objective/train/tokens_used": 1010052576,
      "theoretical_loss": 3.352871741443686,
      "tokens_seen": 2651193344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020004040404040404,
      "loss": 2.6127,
      "theoretical_loss": 3.352871741443686,
      "tokens_seen": 2651193344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020002020202020203,
      "loss": 2.4828,
      "theoretical_loss": 3.3528652078698835,
      "tokens_seen": 2651258880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 2.548,
      "theoretical_loss": 3.3528586745028006,
      "tokens_seen": 2651324416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000199979797979798,
      "loss": 2.6384,
      "theoretical_loss": 3.3528521413424257,
      "tokens_seen": 2651389952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019995959595959595,
      "loss": 2.478,
      "theoretical_loss": 3.3528456083887472,
      "tokens_seen": 2651455488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019993939393939396,
      "loss": 2.6538,
      "theoretical_loss": 3.352839075641753,
      "tokens_seen": 2651521024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019991919191919192,
      "loss": 2.611,
      "theoretical_loss": 3.3528325431014325,
      "tokens_seen": 2651586560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001998989898989899,
      "loss": 2.6428,
      "theoretical_loss": 3.3528260107677728,
      "tokens_seen": 2651652096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019987878787878789,
      "loss": 2.6609,
      "theoretical_loss": 3.352819478640763,
      "tokens_seen": 2651717632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019985858585858584,
      "loss": 2.4714,
      "theoretical_loss": 3.3528129467203915,
      "tokens_seen": 2651783168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019983838383838385,
      "loss": 2.4131,
      "theoretical_loss": 3.352806415006646,
      "tokens_seen": 2651848704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001998181818181818,
      "loss": 2.4969,
      "theoretical_loss": 3.352799883499516,
      "tokens_seen": 2651914240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019979797979797982,
      "loss": 2.5064,
      "theoretical_loss": 3.3527933521989883,
      "tokens_seen": 2651979776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019977777777777778,
      "loss": 2.5161,
      "theoretical_loss": 3.3527868211050524,
      "tokens_seen": 2652045312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019975757575757574,
      "loss": 2.7671,
      "theoretical_loss": 3.352780290217697,
      "tokens_seen": 2652110848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019973737373737375,
      "loss": 2.4429,
      "theoretical_loss": 3.352773759536909,
      "tokens_seen": 2652176384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001997171717171717,
      "loss": 2.4732,
      "theoretical_loss": 3.352767229062678,
      "tokens_seen": 2652241920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019969696969696971,
      "loss": 2.5058,
      "theoretical_loss": 3.352760698794992,
      "tokens_seen": 2652307456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019967676767676767,
      "loss": 2.4427,
      "theoretical_loss": 3.3527541687338394,
      "tokens_seen": 2652372992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019965656565656566,
      "loss": 2.6361,
      "theoretical_loss": 3.352747638879208,
      "tokens_seen": 2652438528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019963636363636364,
      "loss": 2.528,
      "theoretical_loss": 3.3527411092310873,
      "tokens_seen": 2652504064
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019961616161616162,
      "loss": 2.118,
      "theoretical_loss": 3.3527345797894648,
      "tokens_seen": 2652569600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001995959595959596,
      "loss": 2.6372,
      "theoretical_loss": 3.352728050554329,
      "tokens_seen": 2652635136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019957575757575756,
      "loss": 2.3212,
      "theoretical_loss": 3.3527215215256687,
      "tokens_seen": 2652700672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019955555555555558,
      "loss": 2.6494,
      "theoretical_loss": 3.352714992703471,
      "tokens_seen": 2652766208
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1453835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6858787536621094,
      "objective/train/theoretical_loss": 3.3527084640877263,
      "objective/train/tokens_used": 1011690976,
      "theoretical_loss": 3.3527084640877263,
      "tokens_seen": 2652831744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019953535353535353,
      "loss": 2.5642,
      "theoretical_loss": 3.3527084640877263,
      "tokens_seen": 2652831744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019951515151515152,
      "loss": 2.5744,
      "theoretical_loss": 3.3527019356784216,
      "tokens_seen": 2652897280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001994949494949495,
      "loss": 2.3795,
      "theoretical_loss": 3.352695407475545,
      "tokens_seen": 2652962816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019947474747474748,
      "loss": 2.3516,
      "theoretical_loss": 3.352688879479086,
      "tokens_seen": 2653028352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019945454545454547,
      "loss": 2.5893,
      "theoretical_loss": 3.3526823516890323,
      "tokens_seen": 2653093888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019943434343434345,
      "loss": 2.5813,
      "theoretical_loss": 3.352675824105372,
      "tokens_seen": 2653159424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001994141414141414,
      "loss": 2.463,
      "theoretical_loss": 3.352669296728094,
      "tokens_seen": 2653224960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001993939393939394,
      "loss": 2.5026,
      "theoretical_loss": 3.352662769557187,
      "tokens_seen": 2653290496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019937373737373738,
      "loss": 2.4246,
      "theoretical_loss": 3.3526562425926385,
      "tokens_seen": 2653356032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019935353535353536,
      "loss": 2.5525,
      "theoretical_loss": 3.3526497158344375,
      "tokens_seen": 2653421568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019933333333333334,
      "loss": 2.6686,
      "theoretical_loss": 3.352643189282572,
      "tokens_seen": 2653487104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001993131313131313,
      "loss": 2.5692,
      "theoretical_loss": 3.3526366629370306,
      "tokens_seen": 2653552640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001992929292929293,
      "loss": 2.4908,
      "theoretical_loss": 3.3526301367978015,
      "tokens_seen": 2653618176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019927272727272727,
      "loss": 2.3672,
      "theoretical_loss": 3.3526236108648733,
      "tokens_seen": 2653683712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019925252525252528,
      "loss": 2.5324,
      "theoretical_loss": 3.352617085138234,
      "tokens_seen": 2653749248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019923232323232324,
      "loss": 2.6159,
      "theoretical_loss": 3.352610559617873,
      "tokens_seen": 2653814784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001992121212121212,
      "loss": 2.4106,
      "theoretical_loss": 3.3526040343037775,
      "tokens_seen": 2653880320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001991919191919192,
      "loss": 2.5256,
      "theoretical_loss": 3.3525975091959364,
      "tokens_seen": 2653945856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019917171717171716,
      "loss": 2.4398,
      "theoretical_loss": 3.352590984294338,
      "tokens_seen": 2654011392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019915151515151517,
      "loss": 2.4575,
      "theoretical_loss": 3.3525844595989707,
      "tokens_seen": 2654076928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019913131313131313,
      "loss": 2.4998,
      "theoretical_loss": 3.352577935109823,
      "tokens_seen": 2654142464
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019911111111111111,
      "loss": 2.4917,
      "theoretical_loss": 3.3525714108268834,
      "tokens_seen": 2654208000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001990909090909091,
      "loss": 2.5708,
      "theoretical_loss": 3.35256488675014,
      "tokens_seen": 2654273536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019907070707070708,
      "loss": 2.2758,
      "theoretical_loss": 3.3525583628795808,
      "tokens_seen": 2654339072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019905050505050507,
      "loss": 2.4219,
      "theoretical_loss": 3.352551839215195,
      "tokens_seen": 2654404608
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1454467,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1566362380981445,
      "objective/train/theoretical_loss": 3.352545315756971,
      "objective/train/tokens_used": 1013329376,
      "theoretical_loss": 3.352545315756971,
      "tokens_seen": 2654470144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019903030303030302,
      "loss": 2.7777,
      "theoretical_loss": 3.352545315756971,
      "tokens_seen": 2654470144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000199010101010101,
      "loss": 2.5015,
      "theoretical_loss": 3.352538792504897,
      "tokens_seen": 2654535680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198989898989899,
      "loss": 2.6502,
      "theoretical_loss": 3.3525322694589605,
      "tokens_seen": 2654601216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019896969696969697,
      "loss": 2.7015,
      "theoretical_loss": 3.3525257466191514,
      "tokens_seen": 2654666752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019894949494949496,
      "loss": 2.5124,
      "theoretical_loss": 3.352519223985457,
      "tokens_seen": 2654732288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019892929292929294,
      "loss": 2.4434,
      "theoretical_loss": 3.352512701557866,
      "tokens_seen": 2654797824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001989090909090909,
      "loss": 2.4979,
      "theoretical_loss": 3.352506179336367,
      "tokens_seen": 2654863360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019888888888888888,
      "loss": 2.5917,
      "theoretical_loss": 3.3524996573209482,
      "tokens_seen": 2654928896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019886868686868687,
      "loss": 2.2612,
      "theoretical_loss": 3.3524931355115983,
      "tokens_seen": 2654994432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019884848484848485,
      "loss": 2.5707,
      "theoretical_loss": 3.3524866139083054,
      "tokens_seen": 2655059968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019882828282828284,
      "loss": 2.5653,
      "theoretical_loss": 3.3524800925110574,
      "tokens_seen": 2655125504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001988080808080808,
      "loss": 2.6521,
      "theoretical_loss": 3.352473571319844,
      "tokens_seen": 2655191040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001987878787878788,
      "loss": 2.6293,
      "theoretical_loss": 3.3524670503346528,
      "tokens_seen": 2655256576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019876767676767676,
      "loss": 2.4631,
      "theoretical_loss": 3.352460529555472,
      "tokens_seen": 2655322112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019874747474747477,
      "loss": 2.5296,
      "theoretical_loss": 3.3524540089822903,
      "tokens_seen": 2655387648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019872727272727273,
      "loss": 2.5384,
      "theoretical_loss": 3.3524474886150966,
      "tokens_seen": 2655453184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019870707070707068,
      "loss": 2.2901,
      "theoretical_loss": 3.352440968453878,
      "tokens_seen": 2655518720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001986868686868687,
      "loss": 2.6048,
      "theoretical_loss": 3.3524344484986246,
      "tokens_seen": 2655584256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019866666666666665,
      "loss": 2.5828,
      "theoretical_loss": 3.3524279287493233,
      "tokens_seen": 2655649792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019864646464646466,
      "loss": 2.7395,
      "theoretical_loss": 3.3524214092059634,
      "tokens_seen": 2655715328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019862626262626262,
      "loss": 2.5691,
      "theoretical_loss": 3.352414889868533,
      "tokens_seen": 2655780864
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019860606060606063,
      "loss": 2.4172,
      "theoretical_loss": 3.3524083707370207,
      "tokens_seen": 2655846400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001985858585858586,
      "loss": 2.402,
      "theoretical_loss": 3.3524018518114147,
      "tokens_seen": 2655911936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019856565656565657,
      "loss": 2.534,
      "theoretical_loss": 3.3523953330917036,
      "tokens_seen": 2655977472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019854545454545456,
      "loss": 2.5576,
      "theoretical_loss": 3.3523888145778757,
      "tokens_seen": 2656043008
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1455118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3296263217926025,
      "objective/train/theoretical_loss": 3.3523822962699192,
      "objective/train/tokens_used": 1014967776,
      "theoretical_loss": 3.3523822962699192,
      "tokens_seen": 2656108544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001985252525252525,
      "loss": 2.443,
      "theoretical_loss": 3.3523822962699192,
      "tokens_seen": 2656108544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019850505050505052,
      "loss": 2.3405,
      "theoretical_loss": 3.352375778167823,
      "tokens_seen": 2656174080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019848484848484848,
      "loss": 2.551,
      "theoretical_loss": 3.3523692602715753,
      "tokens_seen": 2656239616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019846464646464647,
      "loss": 2.6063,
      "theoretical_loss": 3.3523627425811644,
      "tokens_seen": 2656305152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019844444444444445,
      "loss": 2.5995,
      "theoretical_loss": 3.3523562250965786,
      "tokens_seen": 2656370688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019842424242424243,
      "loss": 2.3433,
      "theoretical_loss": 3.352349707817807,
      "tokens_seen": 2656436224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019840404040404042,
      "loss": 2.5972,
      "theoretical_loss": 3.352343190744837,
      "tokens_seen": 2656501760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983838383838384,
      "loss": 2.6779,
      "theoretical_loss": 3.3523366738776583,
      "tokens_seen": 2656567296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019836363636363636,
      "loss": 2.5159,
      "theoretical_loss": 3.352330157216258,
      "tokens_seen": 2656632832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019834343434343434,
      "loss": 2.4608,
      "theoretical_loss": 3.3523236407606256,
      "tokens_seen": 2656698368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019832323232323233,
      "loss": 2.4275,
      "theoretical_loss": 3.352317124510749,
      "tokens_seen": 2656763904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019832323232323233,
      "loss": 2.3229,
      "theoretical_loss": 3.3523106084666168,
      "tokens_seen": 2656829440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983030303030303,
      "loss": 2.61,
      "theoretical_loss": 3.352304092628217,
      "tokens_seen": 2656894976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001982828282828283,
      "loss": 2.4089,
      "theoretical_loss": 3.3522975769955385,
      "tokens_seen": 2656960512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019826262626262625,
      "loss": 2.5355,
      "theoretical_loss": 3.3522910615685695,
      "tokens_seen": 2657026048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019824242424242426,
      "loss": 2.5925,
      "theoretical_loss": 3.3522845463472986,
      "tokens_seen": 2657091584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019822222222222222,
      "loss": 2.5389,
      "theoretical_loss": 3.352278031331714,
      "tokens_seen": 2657157120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019820202020202023,
      "loss": 2.5292,
      "theoretical_loss": 3.3522715165218044,
      "tokens_seen": 2657222656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981818181818182,
      "loss": 2.4939,
      "theoretical_loss": 3.352265001917558,
      "tokens_seen": 2657288192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019816161616161614,
      "loss": 2.3584,
      "theoretical_loss": 3.352258487518964,
      "tokens_seen": 2657353728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019814141414141415,
      "loss": 2.3721,
      "theoretical_loss": 3.3522519733260094,
      "tokens_seen": 2657419264
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981212121212121,
      "loss": 2.6653,
      "theoretical_loss": 3.352245459338684,
      "tokens_seen": 2657484800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019810101010101012,
      "loss": 2.5444,
      "theoretical_loss": 3.352238945556975,
      "tokens_seen": 2657550336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019808080808080808,
      "loss": 2.6636,
      "theoretical_loss": 3.3522324319808723,
      "tokens_seen": 2657615872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019806060606060606,
      "loss": 2.4531,
      "theoretical_loss": 3.352225918610363,
      "tokens_seen": 2657681408
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1456223,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8828985691070557,
      "objective/train/theoretical_loss": 3.352219405445436,
      "objective/train/tokens_used": 1016606176,
      "theoretical_loss": 3.352219405445436,
      "tokens_seen": 2657746944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019804040404040405,
      "loss": 2.5353,
      "theoretical_loss": 3.352219405445436,
      "tokens_seen": 2657746944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198020202020202,
      "loss": 2.4716,
      "theoretical_loss": 3.3522128924860803,
      "tokens_seen": 2657812480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019800000000000002,
      "loss": 2.5462,
      "theoretical_loss": 3.3522063797322836,
      "tokens_seen": 2657878016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019797979797979797,
      "loss": 2.4209,
      "theoretical_loss": 3.3521998671840345,
      "tokens_seen": 2657943552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019795959595959596,
      "loss": 2.3932,
      "theoretical_loss": 3.3521933548413223,
      "tokens_seen": 2658009088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019793939393939394,
      "loss": 2.3762,
      "theoretical_loss": 3.3521868427041337,
      "tokens_seen": 2658074624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019791919191919192,
      "loss": 2.4539,
      "theoretical_loss": 3.352180330772459,
      "tokens_seen": 2658140160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001978989898989899,
      "loss": 2.4222,
      "theoretical_loss": 3.3521738190462855,
      "tokens_seen": 2658205696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001978787878787879,
      "loss": 2.3635,
      "theoretical_loss": 3.3521673075256015,
      "tokens_seen": 2658271232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019785858585858585,
      "loss": 2.4932,
      "theoretical_loss": 3.3521607962103963,
      "tokens_seen": 2658336768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019783838383838383,
      "loss": 2.4453,
      "theoretical_loss": 3.3521542851006583,
      "tokens_seen": 2658402304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019781818181818182,
      "loss": 2.4196,
      "theoretical_loss": 3.352147774196375,
      "tokens_seen": 2658467840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977979797979798,
      "loss": 2.4653,
      "theoretical_loss": 3.3521412634975354,
      "tokens_seen": 2658533376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019777777777777778,
      "loss": 2.4768,
      "theoretical_loss": 3.3521347530041288,
      "tokens_seen": 2658598912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019775757575757577,
      "loss": 2.5614,
      "theoretical_loss": 3.3521282427161423,
      "tokens_seen": 2658664448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019773737373737375,
      "loss": 2.6416,
      "theoretical_loss": 3.352121732633565,
      "tokens_seen": 2658729984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977171717171717,
      "loss": 2.4865,
      "theoretical_loss": 3.3521152227563853,
      "tokens_seen": 2658795520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019769696969696972,
      "loss": 2.7369,
      "theoretical_loss": 3.3521087130845912,
      "tokens_seen": 2658861056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019767676767676768,
      "loss": 2.3634,
      "theoretical_loss": 3.352102203618172,
      "tokens_seen": 2658926592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019765656565656566,
      "loss": 2.5013,
      "theoretical_loss": 3.352095694357116,
      "tokens_seen": 2658992128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019763636363636365,
      "loss": 2.5832,
      "theoretical_loss": 3.352089185301411,
      "tokens_seen": 2659057664
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001976161616161616,
      "loss": 2.484,
      "theoretical_loss": 3.352082676451046,
      "tokens_seen": 2659123200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001975959595959596,
      "loss": 2.682,
      "theoretical_loss": 3.3520761678060094,
      "tokens_seen": 2659188736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019757575757575757,
      "loss": 2.6926,
      "theoretical_loss": 3.3520696593662898,
      "tokens_seen": 2659254272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019755555555555558,
      "loss": 2.4182,
      "theoretical_loss": 3.352063151131875,
      "tokens_seen": 2659319808
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1457493,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5922701358795166,
      "objective/train/theoretical_loss": 3.352056643102754,
      "objective/train/tokens_used": 1018244576,
      "theoretical_loss": 3.352056643102754,
      "tokens_seen": 2659385344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019753535353535354,
      "loss": 2.4809,
      "theoretical_loss": 3.352056643102754,
      "tokens_seen": 2659385344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019751515151515152,
      "loss": 2.9616,
      "theoretical_loss": 3.3520501352789154,
      "tokens_seen": 2659450880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001974949494949495,
      "loss": 2.2786,
      "theoretical_loss": 3.3520436276603474,
      "tokens_seen": 2659516416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019747474747474746,
      "loss": 2.3407,
      "theoretical_loss": 3.3520371202470383,
      "tokens_seen": 2659581952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019745454545454547,
      "loss": 2.4585,
      "theoretical_loss": 3.352030613038977,
      "tokens_seen": 2659647488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019743434343434343,
      "loss": 2.7334,
      "theoretical_loss": 3.3520241060361515,
      "tokens_seen": 2659713024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019741414141414141,
      "loss": 2.5413,
      "theoretical_loss": 3.3520175992385512,
      "tokens_seen": 2659778560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001973939393939394,
      "loss": 2.5969,
      "theoretical_loss": 3.3520110926461633,
      "tokens_seen": 2659844096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019737373737373738,
      "loss": 2.4151,
      "theoretical_loss": 3.352004586258977,
      "tokens_seen": 2659909632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019735353535353537,
      "loss": 2.6478,
      "theoretical_loss": 3.3519980800769806,
      "tokens_seen": 2659975168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019733333333333335,
      "loss": 2.2895,
      "theoretical_loss": 3.351991574100163,
      "tokens_seen": 2660040704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001973131313131313,
      "loss": 2.489,
      "theoretical_loss": 3.3519850683285117,
      "tokens_seen": 2660106240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001972929292929293,
      "loss": 2.3768,
      "theoretical_loss": 3.351978562762016,
      "tokens_seen": 2660171776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019727272727272728,
      "loss": 2.634,
      "theoretical_loss": 3.351972057400664,
      "tokens_seen": 2660237312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019725252525252526,
      "loss": 2.5941,
      "theoretical_loss": 3.3519655522444447,
      "tokens_seen": 2660302848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019723232323232324,
      "loss": 2.5072,
      "theoretical_loss": 3.351959047293346,
      "tokens_seen": 2660368384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001972121212121212,
      "loss": 2.6271,
      "theoretical_loss": 3.3519525425473566,
      "tokens_seen": 2660433920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001971919191919192,
      "loss": 2.5462,
      "theoretical_loss": 3.3519460380064654,
      "tokens_seen": 2660499456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019717171717171717,
      "loss": 2.4667,
      "theoretical_loss": 3.35193953367066,
      "tokens_seen": 2660564992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019715151515151515,
      "loss": 2.4661,
      "theoretical_loss": 3.351933029539929,
      "tokens_seen": 2660630528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019713131313131314,
      "loss": 2.4824,
      "theoretical_loss": 3.3519265256142616,
      "tokens_seen": 2660696064
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001971111111111111,
      "loss": 2.3649,
      "theoretical_loss": 3.351920021893646,
      "tokens_seen": 2660761600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001970909090909091,
      "loss": 2.5306,
      "theoretical_loss": 3.35191351837807,
      "tokens_seen": 2660827136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019707070707070706,
      "loss": 2.5332,
      "theoretical_loss": 3.351907015067523,
      "tokens_seen": 2660892672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019705050505050507,
      "loss": 2.6748,
      "theoretical_loss": 3.3519005119619933,
      "tokens_seen": 2660958208
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1458037,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.740746259689331,
      "objective/train/theoretical_loss": 3.3518940090614695,
      "objective/train/tokens_used": 1019882976,
      "theoretical_loss": 3.3518940090614695,
      "tokens_seen": 2661023744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019703030303030303,
      "loss": 2.5832,
      "theoretical_loss": 3.3518940090614695,
      "tokens_seen": 2661023744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000197010101010101,
      "loss": 2.5883,
      "theoretical_loss": 3.3518875063659395,
      "tokens_seen": 2661089280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000196989898989899,
      "loss": 2.4978,
      "theoretical_loss": 3.351881003875392,
      "tokens_seen": 2661154816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019696969696969695,
      "loss": 2.367,
      "theoretical_loss": 3.3518745015898155,
      "tokens_seen": 2661220352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019694949494949496,
      "loss": 2.5075,
      "theoretical_loss": 3.351867999509199,
      "tokens_seen": 2661285888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019692929292929292,
      "loss": 2.5259,
      "theoretical_loss": 3.3518614976335304,
      "tokens_seen": 2661351424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019690909090909093,
      "loss": 2.3953,
      "theoretical_loss": 3.3518549959627983,
      "tokens_seen": 2661416960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001968888888888889,
      "loss": 2.3894,
      "theoretical_loss": 3.3518484944969913,
      "tokens_seen": 2661482496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019686868686868687,
      "loss": 2.6628,
      "theoretical_loss": 3.351841993236098,
      "tokens_seen": 2661548032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019684848484848486,
      "loss": 2.2204,
      "theoretical_loss": 3.351835492180107,
      "tokens_seen": 2661613568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019682828282828284,
      "loss": 2.6109,
      "theoretical_loss": 3.351828991329006,
      "tokens_seen": 2661679104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019680808080808082,
      "loss": 2.5322,
      "theoretical_loss": 3.351822490682784,
      "tokens_seen": 2661744640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019678787878787878,
      "loss": 2.3903,
      "theoretical_loss": 3.3518159902414304,
      "tokens_seen": 2661810176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019676767676767677,
      "loss": 2.4966,
      "theoretical_loss": 3.351809490004932,
      "tokens_seen": 2661875712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019674747474747475,
      "loss": 2.5301,
      "theoretical_loss": 3.3518029899732786,
      "tokens_seen": 2661941248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019672727272727273,
      "loss": 2.4648,
      "theoretical_loss": 3.3517964901464583,
      "tokens_seen": 2662006784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019670707070707072,
      "loss": 2.4157,
      "theoretical_loss": 3.351789990524459,
      "tokens_seen": 2662072320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001966868686868687,
      "loss": 2.5671,
      "theoretical_loss": 3.3517834911072706,
      "tokens_seen": 2662137856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019666666666666666,
      "loss": 2.5077,
      "theoretical_loss": 3.35177699189488,
      "tokens_seen": 2662203392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019664646464646467,
      "loss": 2.489,
      "theoretical_loss": 3.351770492887277,
      "tokens_seen": 2662268928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019662626262626263,
      "loss": 2.545,
      "theoretical_loss": 3.3517639940844495,
      "tokens_seen": 2662334464
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001966060606060606,
      "loss": 2.3762,
      "theoretical_loss": 3.3517574954863862,
      "tokens_seen": 2662400000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001965858585858586,
      "loss": 2.4357,
      "theoretical_loss": 3.3517509970930752,
      "tokens_seen": 2662465536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019656565656565655,
      "loss": 2.6609,
      "theoretical_loss": 3.3517444989045053,
      "tokens_seen": 2662531072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019654545454545456,
      "loss": 2.5247,
      "theoretical_loss": 3.3517380009206654,
      "tokens_seen": 2662596608
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1459076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5953822135925293,
      "objective/train/theoretical_loss": 3.3517315031415436,
      "objective/train/tokens_used": 1021521376,
      "theoretical_loss": 3.3517315031415436,
      "tokens_seen": 2662662144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019652525252525252,
      "loss": 2.7188,
      "theoretical_loss": 3.3517315031415436,
      "tokens_seen": 2662662144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019650505050505053,
      "loss": 2.4411,
      "theoretical_loss": 3.351725005567128,
      "tokens_seen": 2662727680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001964848484848485,
      "loss": 2.5941,
      "theoretical_loss": 3.3517185081974077,
      "tokens_seen": 2662793216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019646464646464644,
      "loss": 2.6263,
      "theoretical_loss": 3.3517120110323715,
      "tokens_seen": 2662858752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019644444444444445,
      "loss": 2.3823,
      "theoretical_loss": 3.351705514072007,
      "tokens_seen": 2662924288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001964242424242424,
      "loss": 2.7029,
      "theoretical_loss": 3.3516990173163035,
      "tokens_seen": 2662989824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019640404040404042,
      "loss": 2.4713,
      "theoretical_loss": 3.351692520765249,
      "tokens_seen": 2663055360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019638383838383838,
      "loss": 2.793,
      "theoretical_loss": 3.3516860244188322,
      "tokens_seen": 2663120896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019636363636363636,
      "loss": 2.6391,
      "theoretical_loss": 3.351679528277042,
      "tokens_seen": 2663186432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019634343434343435,
      "loss": 2.5277,
      "theoretical_loss": 3.3516730323398662,
      "tokens_seen": 2663251968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019632323232323233,
      "loss": 2.4362,
      "theoretical_loss": 3.351666536607294,
      "tokens_seen": 2663317504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019630303030303032,
      "loss": 2.5871,
      "theoretical_loss": 3.351660041079313,
      "tokens_seen": 2663383040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019628282828282827,
      "loss": 2.7775,
      "theoretical_loss": 3.351653545755913,
      "tokens_seen": 2663448576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019626262626262626,
      "loss": 2.5223,
      "theoretical_loss": 3.351647050637082,
      "tokens_seen": 2663514112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019624242424242424,
      "loss": 2.4433,
      "theoretical_loss": 3.351640555722808,
      "tokens_seen": 2663579648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019622222222222222,
      "loss": 2.5425,
      "theoretical_loss": 3.3516340610130797,
      "tokens_seen": 2663645184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962020202020202,
      "loss": 2.5632,
      "theoretical_loss": 3.3516275665078865,
      "tokens_seen": 2663710720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001961818181818182,
      "loss": 2.5539,
      "theoretical_loss": 3.3516210722072155,
      "tokens_seen": 2663776256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019616161616161615,
      "loss": 2.6064,
      "theoretical_loss": 3.3516145781110565,
      "tokens_seen": 2663841792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019614141414141416,
      "loss": 2.4596,
      "theoretical_loss": 3.3516080842193974,
      "tokens_seen": 2663907328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019612121212121212,
      "loss": 2.3964,
      "theoretical_loss": 3.3516015905322267,
      "tokens_seen": 2663972864
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001961010101010101,
      "loss": 2.4594,
      "theoretical_loss": 3.3515950970495334,
      "tokens_seen": 2664038400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019608080808080808,
      "loss": 2.4477,
      "theoretical_loss": 3.3515886037713054,
      "tokens_seen": 2664103936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019606060606060607,
      "loss": 2.3682,
      "theoretical_loss": 3.3515821106975316,
      "tokens_seen": 2664169472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019604040404040405,
      "loss": 2.4924,
      "theoretical_loss": 3.3515756178282006,
      "tokens_seen": 2664235008
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1459609,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.219696283340454,
      "objective/train/theoretical_loss": 3.3515691251633006,
      "objective/train/tokens_used": 1023159776,
      "theoretical_loss": 3.3515691251633006,
      "tokens_seen": 2664300544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000196020202020202,
      "loss": 2.5064,
      "theoretical_loss": 3.3515691251633006,
      "tokens_seen": 2664300544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019600000000000002,
      "loss": 2.3594,
      "theoretical_loss": 3.3515626327028203,
      "tokens_seen": 2664366080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019597979797979798,
      "loss": 2.5238,
      "theoretical_loss": 3.3515561404467484,
      "tokens_seen": 2664431616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000195959595959596,
      "loss": 2.6275,
      "theoretical_loss": 3.3515496483950735,
      "tokens_seen": 2664497152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019593939393939395,
      "loss": 2.5507,
      "theoretical_loss": 3.351543156547784,
      "tokens_seen": 2664562688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001959191919191919,
      "loss": 2.5008,
      "theoretical_loss": 3.351536664904868,
      "tokens_seen": 2664628224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001958989898989899,
      "loss": 2.4596,
      "theoretical_loss": 3.3515301734663145,
      "tokens_seen": 2664693760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019587878787878787,
      "loss": 2.2915,
      "theoretical_loss": 3.351523682232112,
      "tokens_seen": 2664759296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019585858585858588,
      "loss": 2.637,
      "theoretical_loss": 3.3515171912022494,
      "tokens_seen": 2664824832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019583838383838384,
      "loss": 2.477,
      "theoretical_loss": 3.3515107003767146,
      "tokens_seen": 2664890368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019581818181818182,
      "loss": 2.3855,
      "theoretical_loss": 3.3515042097554963,
      "tokens_seen": 2664955904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001957979797979798,
      "loss": 2.6494,
      "theoretical_loss": 3.351497719338583,
      "tokens_seen": 2665021440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001957777777777778,
      "loss": 2.5398,
      "theoretical_loss": 3.3514912291259638,
      "tokens_seen": 2665086976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019575757575757577,
      "loss": 2.4313,
      "theoretical_loss": 3.351484739117627,
      "tokens_seen": 2665152512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019573737373737373,
      "loss": 2.6649,
      "theoretical_loss": 3.35147824931356,
      "tokens_seen": 2665218048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019571717171717171,
      "loss": 2.6761,
      "theoretical_loss": 3.3514717597137533,
      "tokens_seen": 2665283584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001956969696969697,
      "loss": 2.4756,
      "theoretical_loss": 3.351465270318194,
      "tokens_seen": 2665349120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019567676767676768,
      "loss": 2.4651,
      "theoretical_loss": 3.3514587811268712,
      "tokens_seen": 2665414656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019565656565656567,
      "loss": 2.6329,
      "theoretical_loss": 3.351452292139774,
      "tokens_seen": 2665480192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019563636363636365,
      "loss": 2.6946,
      "theoretical_loss": 3.3514458033568895,
      "tokens_seen": 2665545728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001956161616161616,
      "loss": 2.5515,
      "theoretical_loss": 3.351439314778207,
      "tokens_seen": 2665611264
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955959595959596,
      "loss": 2.4736,
      "theoretical_loss": 3.3514328264037156,
      "tokens_seen": 2665676800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019557575757575758,
      "loss": 2.5012,
      "theoretical_loss": 3.3514263382334035,
      "tokens_seen": 2665742336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019555555555555556,
      "loss": 2.4619,
      "theoretical_loss": 3.3514198502672587,
      "tokens_seen": 2665807872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019553535353535354,
      "loss": 2.5445,
      "theoretical_loss": 3.3514133625052707,
      "tokens_seen": 2665873408
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1460723,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5888512134552,
      "objective/train/theoretical_loss": 3.3514068749474273,
      "objective/train/tokens_used": 1024798176,
      "theoretical_loss": 3.3514068749474273,
      "tokens_seen": 2665938944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955151515151515,
      "loss": 2.5897,
      "theoretical_loss": 3.3514068749474273,
      "tokens_seen": 2665938944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001954949494949495,
      "loss": 2.4827,
      "theoretical_loss": 3.351400387593717,
      "tokens_seen": 2666004480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019547474747474747,
      "loss": 2.589,
      "theoretical_loss": 3.3513939004441293,
      "tokens_seen": 2666070016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019545454545454548,
      "loss": 2.6115,
      "theoretical_loss": 3.351387413498652,
      "tokens_seen": 2666135552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019543434343434344,
      "loss": 2.2911,
      "theoretical_loss": 3.3513809267572734,
      "tokens_seen": 2666201088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001954141414141414,
      "loss": 2.6373,
      "theoretical_loss": 3.351374440219983,
      "tokens_seen": 2666266624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001953939393939394,
      "loss": 2.247,
      "theoretical_loss": 3.3513679538867684,
      "tokens_seen": 2666332160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019537373737373736,
      "loss": 2.5155,
      "theoretical_loss": 3.3513614677576187,
      "tokens_seen": 2666397696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019535353535353537,
      "loss": 2.7287,
      "theoretical_loss": 3.3513549818325226,
      "tokens_seen": 2666463232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019533333333333333,
      "loss": 2.4993,
      "theoretical_loss": 3.351348496111468,
      "tokens_seen": 2666528768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001953131313131313,
      "loss": 2.4415,
      "theoretical_loss": 3.3513420105944443,
      "tokens_seen": 2666594304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001952929292929293,
      "loss": 2.3402,
      "theoretical_loss": 3.3513355252814394,
      "tokens_seen": 2666659840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019527272727272728,
      "loss": 2.3422,
      "theoretical_loss": 3.351329040172442,
      "tokens_seen": 2666725376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019525252525252526,
      "loss": 2.5642,
      "theoretical_loss": 3.351322555267441,
      "tokens_seen": 2666790912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019523232323232322,
      "loss": 2.5479,
      "theoretical_loss": 3.3513160705664244,
      "tokens_seen": 2666856448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019521212121212123,
      "loss": 2.7494,
      "theoretical_loss": 3.3513095860693816,
      "tokens_seen": 2666921984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001951919191919192,
      "loss": 2.617,
      "theoretical_loss": 3.3513031017763004,
      "tokens_seen": 2666987520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019517171717171717,
      "loss": 2.5739,
      "theoretical_loss": 3.35129661768717,
      "tokens_seen": 2667053056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019515151515151516,
      "loss": 2.5076,
      "theoretical_loss": 3.3512901338019785,
      "tokens_seen": 2667118592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019513131313131314,
      "loss": 2.3524,
      "theoretical_loss": 3.3512836501207146,
      "tokens_seen": 2667184128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019511111111111113,
      "loss": 2.5395,
      "theoretical_loss": 3.3512771666433667,
      "tokens_seen": 2667249664
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001950909090909091,
      "loss": 2.6239,
      "theoretical_loss": 3.3512706833699237,
      "tokens_seen": 2667315200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019507070707070707,
      "loss": 2.5807,
      "theoretical_loss": 3.351264200300374,
      "tokens_seen": 2667380736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019505050505050505,
      "loss": 2.5741,
      "theoretical_loss": 3.3512577174347067,
      "tokens_seen": 2667446272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019503030303030303,
      "loss": 2.5306,
      "theoretical_loss": 3.3512512347729095,
      "tokens_seen": 2667511808
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1462120,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3585121631622314,
      "objective/train/theoretical_loss": 3.3512447523149715,
      "objective/train/tokens_used": 1026436576,
      "theoretical_loss": 3.3512447523149715,
      "tokens_seen": 2667577344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019501010101010102,
      "loss": 2.4041,
      "theoretical_loss": 3.3512447523149715,
      "tokens_seen": 2667577344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000194989898989899,
      "loss": 2.4452,
      "theoretical_loss": 3.351238270060881,
      "tokens_seen": 2667642880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019496969696969696,
      "loss": 2.3971,
      "theoretical_loss": 3.351231788010627,
      "tokens_seen": 2667708416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019494949494949497,
      "loss": 2.435,
      "theoretical_loss": 3.351225306164198,
      "tokens_seen": 2667773952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019492929292929293,
      "loss": 2.4446,
      "theoretical_loss": 3.351218824521582,
      "tokens_seen": 2667839488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019490909090909094,
      "loss": 2.2768,
      "theoretical_loss": 3.351212343082768,
      "tokens_seen": 2667905024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001948888888888889,
      "loss": 2.5078,
      "theoretical_loss": 3.351205861847745,
      "tokens_seen": 2667970560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019486868686868685,
      "loss": 2.6464,
      "theoretical_loss": 3.351199380816501,
      "tokens_seen": 2668036096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019484848484848486,
      "loss": 2.5922,
      "theoretical_loss": 3.3511928999890244,
      "tokens_seen": 2668101632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019482828282828282,
      "loss": 2.6065,
      "theoretical_loss": 3.3511864193653045,
      "tokens_seen": 2668167168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019480808080808083,
      "loss": 2.6774,
      "theoretical_loss": 3.3511799389453296,
      "tokens_seen": 2668232704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001947878787878788,
      "loss": 2.4449,
      "theoretical_loss": 3.351173458729088,
      "tokens_seen": 2668298240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019476767676767677,
      "loss": 2.4654,
      "theoretical_loss": 3.3511669787165688,
      "tokens_seen": 2668363776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019474747474747476,
      "loss": 2.4668,
      "theoretical_loss": 3.35116049890776,
      "tokens_seen": 2668429312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001947272727272727,
      "loss": 2.522,
      "theoretical_loss": 3.3511540193026508,
      "tokens_seen": 2668494848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019470707070707072,
      "loss": 2.722,
      "theoretical_loss": 3.3511475399012296,
      "tokens_seen": 2668560384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019468686868686868,
      "loss": 2.5914,
      "theoretical_loss": 3.351141060703484,
      "tokens_seen": 2668625920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019466666666666666,
      "loss": 2.5427,
      "theoretical_loss": 3.3511345817094047,
      "tokens_seen": 2668691456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019464646464646465,
      "loss": 2.6406,
      "theoretical_loss": 3.3511281029189783,
      "tokens_seen": 2668756992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019462626262626263,
      "loss": 2.2987,
      "theoretical_loss": 3.351121624332195,
      "tokens_seen": 2668822528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019460606060606062,
      "loss": 2.4408,
      "theoretical_loss": 3.351115145949042,
      "tokens_seen": 2668888064
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001945858585858586,
      "loss": 2.3816,
      "theoretical_loss": 3.351108667769508,
      "tokens_seen": 2668953600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019456565656565656,
      "loss": 2.7126,
      "theoretical_loss": 3.351102189793583,
      "tokens_seen": 2669019136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019454545454545454,
      "loss": 2.4215,
      "theoretical_loss": 3.3510957120212543,
      "tokens_seen": 2669084672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019452525252525252,
      "loss": 2.4052,
      "theoretical_loss": 3.3510892344525107,
      "tokens_seen": 2669150208
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1463069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5914080142974854,
      "objective/train/theoretical_loss": 3.3510827570873416,
      "objective/train/tokens_used": 1028074976,
      "theoretical_loss": 3.3510827570873416,
      "tokens_seen": 2669215744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001945050505050505,
      "loss": 2.5217,
      "theoretical_loss": 3.3510827570873416,
      "tokens_seen": 2669215744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001944848484848485,
      "loss": 2.4921,
      "theoretical_loss": 3.3510762799257345,
      "tokens_seen": 2669281280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019446464646464645,
      "loss": 2.5745,
      "theoretical_loss": 3.351069802967679,
      "tokens_seen": 2669346816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019444444444444446,
      "loss": 2.4156,
      "theoretical_loss": 3.3510633262131626,
      "tokens_seen": 2669412352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019442424242424242,
      "loss": 2.4641,
      "theoretical_loss": 3.3510568496621747,
      "tokens_seen": 2669477888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019440404040404043,
      "loss": 2.5401,
      "theoretical_loss": 3.351050373314704,
      "tokens_seen": 2669543424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019438383838383839,
      "loss": 2.6074,
      "theoretical_loss": 3.3510438971707384,
      "tokens_seen": 2669608960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019436363636363637,
      "loss": 2.6119,
      "theoretical_loss": 3.3510374212302674,
      "tokens_seen": 2669674496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019434343434343435,
      "loss": 2.4556,
      "theoretical_loss": 3.351030945493279,
      "tokens_seen": 2669740032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001943232323232323,
      "loss": 2.3706,
      "theoretical_loss": 3.3510244699597616,
      "tokens_seen": 2669805568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019430303030303032,
      "loss": 2.5111,
      "theoretical_loss": 3.3510179946297045,
      "tokens_seen": 2669871104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019428282828282828,
      "loss": 2.6926,
      "theoretical_loss": 3.3510115195030963,
      "tokens_seen": 2669936640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001942626262626263,
      "loss": 2.6848,
      "theoretical_loss": 3.351005044579925,
      "tokens_seen": 2670002176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019424242424242425,
      "loss": 2.3939,
      "theoretical_loss": 3.350998569860179,
      "tokens_seen": 2670067712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019422222222222223,
      "loss": 2.5511,
      "theoretical_loss": 3.350992095343848,
      "tokens_seen": 2670133248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019420202020202021,
      "loss": 2.7412,
      "theoretical_loss": 3.3509856210309206,
      "tokens_seen": 2670198784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019418181818181817,
      "loss": 2.6661,
      "theoretical_loss": 3.350979146921384,
      "tokens_seen": 2670264320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019416161616161618,
      "loss": 2.4929,
      "theoretical_loss": 3.350972673015228,
      "tokens_seen": 2670329856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019414141414141414,
      "loss": 2.6713,
      "theoretical_loss": 3.350966199312441,
      "tokens_seen": 2670395392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019412121212121212,
      "loss": 2.4785,
      "theoretical_loss": 3.3509597258130115,
      "tokens_seen": 2670460928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001941010101010101,
      "loss": 2.4691,
      "theoretical_loss": 3.350953252516928,
      "tokens_seen": 2670526464
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001940808080808081,
      "loss": 2.5176,
      "theoretical_loss": 3.3509467794241794,
      "tokens_seen": 2670592000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019406060606060607,
      "loss": 2.5314,
      "theoretical_loss": 3.350940306534754,
      "tokens_seen": 2670657536
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019404040404040406,
      "loss": 2.5522,
      "theoretical_loss": 3.3509338338486407,
      "tokens_seen": 2670723072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019402020202020202,
      "loss": 2.7239,
      "theoretical_loss": 3.3509273613658284,
      "tokens_seen": 2670788608
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1463760,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.706230878829956,
      "objective/train/theoretical_loss": 3.350920889086305,
      "objective/train/tokens_used": 1029713376,
      "theoretical_loss": 3.350920889086305,
      "tokens_seen": 2670854144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000194,
      "loss": 2.4463,
      "theoretical_loss": 3.350920889086305,
      "tokens_seen": 2670854144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019397979797979798,
      "loss": 2.4683,
      "theoretical_loss": 3.3509144170100598,
      "tokens_seen": 2670919680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019395959595959597,
      "loss": 2.5659,
      "theoretical_loss": 3.350907945137081,
      "tokens_seen": 2670985216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019393939393939395,
      "loss": 2.3787,
      "theoretical_loss": 3.3509014734673572,
      "tokens_seen": 2671050752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001939191919191919,
      "loss": 2.6089,
      "theoretical_loss": 3.350895002000877,
      "tokens_seen": 2671116288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019389898989898992,
      "loss": 2.6349,
      "theoretical_loss": 3.3508885307376297,
      "tokens_seen": 2671181824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019387878787878788,
      "loss": 2.6378,
      "theoretical_loss": 3.350882059677603,
      "tokens_seen": 2671247360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019385858585858586,
      "loss": 2.818,
      "theoretical_loss": 3.3508755888207866,
      "tokens_seen": 2671312896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019383838383838384,
      "loss": 2.3639,
      "theoretical_loss": 3.350869118167168,
      "tokens_seen": 2671378432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001938181818181818,
      "loss": 2.439,
      "theoretical_loss": 3.350862647716737,
      "tokens_seen": 2671443968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001937979797979798,
      "loss": 2.5524,
      "theoretical_loss": 3.3508561774694807,
      "tokens_seen": 2671509504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019377777777777777,
      "loss": 2.5102,
      "theoretical_loss": 3.350849707425389,
      "tokens_seen": 2671575040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019375757575757578,
      "loss": 2.4817,
      "theoretical_loss": 3.35084323758445,
      "tokens_seen": 2671640576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019373737373737374,
      "loss": 2.4164,
      "theoretical_loss": 3.350836767946653,
      "tokens_seen": 2671706112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019371717171717172,
      "loss": 2.5292,
      "theoretical_loss": 3.350830298511986,
      "tokens_seen": 2671771648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001936969696969697,
      "loss": 2.586,
      "theoretical_loss": 3.3508238292804378,
      "tokens_seen": 2671837184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019367676767676766,
      "loss": 2.5946,
      "theoretical_loss": 3.3508173602519964,
      "tokens_seen": 2671902720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019365656565656567,
      "loss": 2.548,
      "theoretical_loss": 3.3508108914266517,
      "tokens_seen": 2671968256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019363636363636363,
      "loss": 2.5511,
      "theoretical_loss": 3.3508044228043916,
      "tokens_seen": 2672033792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001936161616161616,
      "loss": 2.3673,
      "theoretical_loss": 3.3507979543852047,
      "tokens_seen": 2672099328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001935959595959596,
      "loss": 2.5926,
      "theoretical_loss": 3.35079148616908,
      "tokens_seen": 2672164864
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019357575757575758,
      "loss": 2.5628,
      "theoretical_loss": 3.3507850181560057,
      "tokens_seen": 2672230400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019355555555555557,
      "loss": 2.42,
      "theoretical_loss": 3.3507785503459706,
      "tokens_seen": 2672295936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019353535353535355,
      "loss": 2.5504,
      "theoretical_loss": 3.3507720827389638,
      "tokens_seen": 2672361472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019351515151515153,
      "loss": 2.4483,
      "theoretical_loss": 3.350765615334973,
      "tokens_seen": 2672427008
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1465035,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.667522668838501,
      "objective/train/theoretical_loss": 3.3507591481339882,
      "objective/train/tokens_used": 1031351776,
      "theoretical_loss": 3.3507591481339882,
      "tokens_seen": 2672492544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001934949494949495,
      "loss": 2.664,
      "theoretical_loss": 3.3507591481339882,
      "tokens_seen": 2672492544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019347474747474747,
      "loss": 2.6409,
      "theoretical_loss": 3.350752681135997,
      "tokens_seen": 2672558080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019345454545454546,
      "loss": 2.5572,
      "theoretical_loss": 3.350746214340988,
      "tokens_seen": 2672623616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019343434343434344,
      "loss": 2.588,
      "theoretical_loss": 3.3507397477489507,
      "tokens_seen": 2672689152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019341414141414143,
      "loss": 2.624,
      "theoretical_loss": 3.3507332813598727,
      "tokens_seen": 2672754688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001933939393939394,
      "loss": 2.4797,
      "theoretical_loss": 3.3507268151737435,
      "tokens_seen": 2672820224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019337373737373737,
      "loss": 2.4654,
      "theoretical_loss": 3.350720349190551,
      "tokens_seen": 2672885760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019335353535353538,
      "loss": 2.444,
      "theoretical_loss": 3.350713883410285,
      "tokens_seen": 2672951296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019333333333333333,
      "loss": 2.5188,
      "theoretical_loss": 3.350707417832933,
      "tokens_seen": 2673016832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019331313131313132,
      "loss": 2.4828,
      "theoretical_loss": 3.3507009524584843,
      "tokens_seen": 2673082368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001932929292929293,
      "loss": 2.5014,
      "theoretical_loss": 3.350694487286927,
      "tokens_seen": 2673147904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019327272727272726,
      "loss": 2.496,
      "theoretical_loss": 3.3506880223182507,
      "tokens_seen": 2673213440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019325252525252527,
      "loss": 2.5516,
      "theoretical_loss": 3.350681557552443,
      "tokens_seen": 2673278976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019323232323232323,
      "loss": 2.6604,
      "theoretical_loss": 3.3506750929894933,
      "tokens_seen": 2673344512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019321212121212124,
      "loss": 2.5113,
      "theoretical_loss": 3.35066862862939,
      "tokens_seen": 2673410048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001931919191919192,
      "loss": 2.6697,
      "theoretical_loss": 3.350662164472122,
      "tokens_seen": 2673475584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019317171717171715,
      "loss": 2.6501,
      "theoretical_loss": 3.3506557005176774,
      "tokens_seen": 2673541120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019315151515151516,
      "loss": 2.6429,
      "theoretical_loss": 3.350649236766045,
      "tokens_seen": 2673606656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019313131313131312,
      "loss": 2.4218,
      "theoretical_loss": 3.3506427732172144,
      "tokens_seen": 2673672192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019311111111111113,
      "loss": 2.5571,
      "theoretical_loss": 3.350636309871173,
      "tokens_seen": 2673737728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001930909090909091,
      "loss": 2.4029,
      "theoretical_loss": 3.35062984672791,
      "tokens_seen": 2673803264
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019307070707070707,
      "loss": 2.4958,
      "theoretical_loss": 3.350623383787414,
      "tokens_seen": 2673868800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019305050505050506,
      "loss": 2.4506,
      "theoretical_loss": 3.350616921049674,
      "tokens_seen": 2673934336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019303030303030304,
      "loss": 2.5517,
      "theoretical_loss": 3.3506104585146783,
      "tokens_seen": 2673999872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019301010101010102,
      "loss": 2.4225,
      "theoretical_loss": 3.3506039961824157,
      "tokens_seen": 2674065408
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1465709,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.590970993041992,
      "objective/train/theoretical_loss": 3.3505975340528744,
      "objective/train/tokens_used": 1032990176,
      "theoretical_loss": 3.3505975340528744,
      "tokens_seen": 2674130944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019298989898989898,
      "loss": 2.751,
      "theoretical_loss": 3.3505975340528744,
      "tokens_seen": 2674130944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019296969696969696,
      "loss": 2.5505,
      "theoretical_loss": 3.3505910721260443,
      "tokens_seen": 2674196480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019294949494949495,
      "loss": 2.3657,
      "theoretical_loss": 3.3505846104019126,
      "tokens_seen": 2674262016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019292929292929293,
      "loss": 2.385,
      "theoretical_loss": 3.3505781488804693,
      "tokens_seen": 2674327552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019290909090909092,
      "loss": 2.4174,
      "theoretical_loss": 3.350571687561702,
      "tokens_seen": 2674393088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001928888888888889,
      "loss": 2.5721,
      "theoretical_loss": 3.3505652264456,
      "tokens_seen": 2674458624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019286868686868686,
      "loss": 2.1511,
      "theoretical_loss": 3.350558765532152,
      "tokens_seen": 2674524160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019284848484848487,
      "loss": 2.5657,
      "theoretical_loss": 3.350552304821346,
      "tokens_seen": 2674589696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019282828282828283,
      "loss": 2.4621,
      "theoretical_loss": 3.3505458443131717,
      "tokens_seen": 2674655232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001928080808080808,
      "loss": 2.3927,
      "theoretical_loss": 3.3505393840076167,
      "tokens_seen": 2674720768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001927878787878788,
      "loss": 2.4952,
      "theoretical_loss": 3.3505329239046704,
      "tokens_seen": 2674786304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019276767676767675,
      "loss": 2.6709,
      "theoretical_loss": 3.3505264640043215,
      "tokens_seen": 2674851840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019274747474747476,
      "loss": 2.3387,
      "theoretical_loss": 3.350520004306558,
      "tokens_seen": 2674917376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019272727272727272,
      "loss": 2.5509,
      "theoretical_loss": 3.3505135448113696,
      "tokens_seen": 2674982912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019270707070707073,
      "loss": 2.417,
      "theoretical_loss": 3.3505070855187444,
      "tokens_seen": 2675048448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019268686868686869,
      "loss": 2.5144,
      "theoretical_loss": 3.350500626428671,
      "tokens_seen": 2675113984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001926666666666667,
      "loss": 2.6062,
      "theoretical_loss": 3.350494167541138,
      "tokens_seen": 2675179520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019264646464646465,
      "loss": 2.5596,
      "theoretical_loss": 3.350487708856135,
      "tokens_seen": 2675245056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001926262626262626,
      "loss": 2.5749,
      "theoretical_loss": 3.350481250373649,
      "tokens_seen": 2675310592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019260606060606062,
      "loss": 2.4904,
      "theoretical_loss": 3.3504747920936704,
      "tokens_seen": 2675376128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019258585858585858,
      "loss": 2.5758,
      "theoretical_loss": 3.350468334016187,
      "tokens_seen": 2675441664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001925656565656566,
      "loss": 2.5619,
      "theoretical_loss": 3.3504618761411877,
      "tokens_seen": 2675507200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019254545454545455,
      "loss": 2.349,
      "theoretical_loss": 3.350455418468661,
      "tokens_seen": 2675572736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019252525252525253,
      "loss": 2.7623,
      "theoretical_loss": 3.350448960998596,
      "tokens_seen": 2675638272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019250505050505051,
      "loss": 2.7085,
      "theoretical_loss": 3.350442503730981,
      "tokens_seen": 2675703808
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1466786,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8028337955474854,
      "objective/train/theoretical_loss": 3.3504360466658047,
      "objective/train/tokens_used": 1034628576,
      "theoretical_loss": 3.3504360466658047,
      "tokens_seen": 2675769344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001924848484848485,
      "loss": 2.744,
      "theoretical_loss": 3.3504360466658047,
      "tokens_seen": 2675769344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019246464646464648,
      "loss": 2.5138,
      "theoretical_loss": 3.3504295898030563,
      "tokens_seen": 2675834880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019244444444444444,
      "loss": 2.4419,
      "theoretical_loss": 3.3504231331427237,
      "tokens_seen": 2675900416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019242424242424242,
      "loss": 2.4719,
      "theoretical_loss": 3.3504166766847963,
      "tokens_seen": 2675965952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001924040404040404,
      "loss": 2.5871,
      "theoretical_loss": 3.3504102204292625,
      "tokens_seen": 2676031488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001923838383838384,
      "loss": 2.6861,
      "theoretical_loss": 3.350403764376111,
      "tokens_seen": 2676097024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019236363636363637,
      "loss": 2.6325,
      "theoretical_loss": 3.3503973085253307,
      "tokens_seen": 2676162560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019234343434343436,
      "loss": 2.2442,
      "theoretical_loss": 3.35039085287691,
      "tokens_seen": 2676228096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019232323232323232,
      "loss": 2.5936,
      "theoretical_loss": 3.3503843974308376,
      "tokens_seen": 2676293632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001923030303030303,
      "loss": 2.6557,
      "theoretical_loss": 3.3503779421871025,
      "tokens_seen": 2676359168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019228282828282828,
      "loss": 2.6018,
      "theoretical_loss": 3.350371487145693,
      "tokens_seen": 2676424704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019226262626262627,
      "loss": 2.375,
      "theoretical_loss": 3.3503650323065983,
      "tokens_seen": 2676490240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019224242424242425,
      "loss": 2.5115,
      "theoretical_loss": 3.3503585776698066,
      "tokens_seen": 2676555776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001922222222222222,
      "loss": 2.5157,
      "theoretical_loss": 3.3503521232353073,
      "tokens_seen": 2676621312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019220202020202022,
      "loss": 2.5028,
      "theoretical_loss": 3.3503456690030884,
      "tokens_seen": 2676686848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019218181818181818,
      "loss": 2.5938,
      "theoretical_loss": 3.350339214973139,
      "tokens_seen": 2676752384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001921616161616162,
      "loss": 2.6377,
      "theoretical_loss": 3.350332761145447,
      "tokens_seen": 2676817920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019214141414141414,
      "loss": 2.4994,
      "theoretical_loss": 3.3503263075200023,
      "tokens_seen": 2676883456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001921212121212121,
      "loss": 2.5887,
      "theoretical_loss": 3.350319854096793,
      "tokens_seen": 2676948992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001921010101010101,
      "loss": 2.4202,
      "theoretical_loss": 3.350313400875808,
      "tokens_seen": 2677014528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019208080808080807,
      "loss": 2.5599,
      "theoretical_loss": 3.350306947857036,
      "tokens_seen": 2677080064
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019206060606060608,
      "loss": 2.4609,
      "theoretical_loss": 3.350300495040466,
      "tokens_seen": 2677145600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019204040404040404,
      "loss": 2.5329,
      "theoretical_loss": 3.3502940424260856,
      "tokens_seen": 2677211136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019202020202020202,
      "loss": 2.7932,
      "theoretical_loss": 3.3502875900138847,
      "tokens_seen": 2677276672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000192,
      "loss": 2.4324,
      "theoretical_loss": 3.3502811378038513,
      "tokens_seen": 2677342208
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1468239,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.355358600616455,
      "objective/train/theoretical_loss": 3.3502746857959744,
      "objective/train/tokens_used": 1036266976,
      "theoretical_loss": 3.3502746857959744,
      "tokens_seen": 2677407744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000191979797979798,
      "loss": 2.5288,
      "theoretical_loss": 3.3502746857959744,
      "tokens_seen": 2677407744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019195959595959597,
      "loss": 2.5757,
      "theoretical_loss": 3.350268233990243,
      "tokens_seen": 2677473280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019193939393939393,
      "loss": 2.4711,
      "theoretical_loss": 3.3502617823866454,
      "tokens_seen": 2677538816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019191919191919191,
      "loss": 2.5556,
      "theoretical_loss": 3.3502553309851706,
      "tokens_seen": 2677604352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001918989898989899,
      "loss": 2.4789,
      "theoretical_loss": 3.350248879785807,
      "tokens_seen": 2677669888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001918989898989899,
      "loss": 2.6691,
      "theoretical_loss": 3.3502424287885435,
      "tokens_seen": 2677735424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019187878787878788,
      "loss": 2.4576,
      "theoretical_loss": 3.3502359779933686,
      "tokens_seen": 2677800960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019185858585858587,
      "loss": 2.4932,
      "theoretical_loss": 3.3502295274002716,
      "tokens_seen": 2677866496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019183838383838385,
      "loss": 2.5311,
      "theoretical_loss": 3.350223077009241,
      "tokens_seen": 2677932032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001918181818181818,
      "loss": 2.5367,
      "theoretical_loss": 3.350216626820265,
      "tokens_seen": 2677997568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019179797979797982,
      "loss": 2.5406,
      "theoretical_loss": 3.350210176833333,
      "tokens_seen": 2678063104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019177777777777777,
      "loss": 2.3385,
      "theoretical_loss": 3.3502037270484335,
      "tokens_seen": 2678128640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019175757575757576,
      "loss": 2.3386,
      "theoretical_loss": 3.350197277465555,
      "tokens_seen": 2678194176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019173737373737374,
      "loss": 2.5596,
      "theoretical_loss": 3.350190828084686,
      "tokens_seen": 2678259712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019171717171717173,
      "loss": 2.6733,
      "theoretical_loss": 3.350184378905816,
      "tokens_seen": 2678325248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001916969696969697,
      "loss": 2.5132,
      "theoretical_loss": 3.350177929928934,
      "tokens_seen": 2678390784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019167676767676767,
      "loss": 2.3979,
      "theoretical_loss": 3.3501714811540273,
      "tokens_seen": 2678456320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019165656565656568,
      "loss": 2.5014,
      "theoretical_loss": 3.3501650325810854,
      "tokens_seen": 2678521856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019163636363636363,
      "loss": 2.563,
      "theoretical_loss": 3.3501585842100976,
      "tokens_seen": 2678587392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019161616161616165,
      "loss": 2.423,
      "theoretical_loss": 3.3501521360410518,
      "tokens_seen": 2678652928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001915959595959596,
      "loss": 2.1714,
      "theoretical_loss": 3.350145688073937,
      "tokens_seen": 2678718464
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019157575757575756,
      "loss": 2.6184,
      "theoretical_loss": 3.350139240308742,
      "tokens_seen": 2678784000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019155555555555557,
      "loss": 2.4138,
      "theoretical_loss": 3.3501327927454554,
      "tokens_seen": 2678849536
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019153535353535353,
      "loss": 2.5737,
      "theoretical_loss": 3.350126345384066,
      "tokens_seen": 2678915072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019151515151515154,
      "loss": 2.5624,
      "theoretical_loss": 3.350119898224563,
      "tokens_seen": 2678980608
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1468865,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.819284677505493,
      "objective/train/theoretical_loss": 3.3501134512669344,
      "objective/train/tokens_used": 1037905376,
      "theoretical_loss": 3.3501134512669344,
      "tokens_seen": 2679046144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001914949494949495,
      "loss": 2.4153,
      "theoretical_loss": 3.3501134512669344,
      "tokens_seen": 2679046144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019147474747474748,
      "loss": 2.5791,
      "theoretical_loss": 3.3501070045111696,
      "tokens_seen": 2679111680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019145454545454546,
      "loss": 2.4898,
      "theoretical_loss": 3.3501005579572567,
      "tokens_seen": 2679177216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019143434343434342,
      "loss": 2.4189,
      "theoretical_loss": 3.3500941116051846,
      "tokens_seen": 2679242752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019141414141414143,
      "loss": 2.4052,
      "theoretical_loss": 3.3500876654549425,
      "tokens_seen": 2679308288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001913939393939394,
      "loss": 2.4495,
      "theoretical_loss": 3.350081219506519,
      "tokens_seen": 2679373824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019137373737373737,
      "loss": 2.6772,
      "theoretical_loss": 3.350074773759902,
      "tokens_seen": 2679439360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019135353535353536,
      "loss": 2.5668,
      "theoretical_loss": 3.3500683282150816,
      "tokens_seen": 2679504896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019133333333333334,
      "loss": 2.4562,
      "theoretical_loss": 3.3500618828720454,
      "tokens_seen": 2679570432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019131313131313132,
      "loss": 2.3516,
      "theoretical_loss": 3.350055437730783,
      "tokens_seen": 2679635968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001912929292929293,
      "loss": 2.6521,
      "theoretical_loss": 3.3500489927912827,
      "tokens_seen": 2679701504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019127272727272726,
      "loss": 2.5249,
      "theoretical_loss": 3.350042548053533,
      "tokens_seen": 2679767040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019125252525252525,
      "loss": 2.5869,
      "theoretical_loss": 3.3500361035175232,
      "tokens_seen": 2679832576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019123232323232323,
      "loss": 2.6017,
      "theoretical_loss": 3.350029659183242,
      "tokens_seen": 2679898112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019121212121212122,
      "loss": 2.4304,
      "theoretical_loss": 3.350023215050678,
      "tokens_seen": 2679963648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001911919191919192,
      "loss": 2.3659,
      "theoretical_loss": 3.3500167711198197,
      "tokens_seen": 2680029184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019117171717171716,
      "loss": 2.3833,
      "theoretical_loss": 3.350010327390656,
      "tokens_seen": 2680094720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019115151515151517,
      "loss": 2.641,
      "theoretical_loss": 3.3500038838631756,
      "tokens_seen": 2680160256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019113131313131313,
      "loss": 2.5394,
      "theoretical_loss": 3.3499974405373676,
      "tokens_seen": 2680225792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019111111111111114,
      "loss": 2.5248,
      "theoretical_loss": 3.3499909974132205,
      "tokens_seen": 2680291328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001910909090909091,
      "loss": 2.469,
      "theoretical_loss": 3.3499845544907236,
      "tokens_seen": 2680356864
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019107070707070705,
      "loss": 2.4934,
      "theoretical_loss": 3.3499781117698646,
      "tokens_seen": 2680422400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019105050505050506,
      "loss": 2.4445,
      "theoretical_loss": 3.349971669250633,
      "tokens_seen": 2680487936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019103030303030302,
      "loss": 2.2473,
      "theoretical_loss": 3.3499652269330173,
      "tokens_seen": 2680553472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019101010101010103,
      "loss": 2.2994,
      "theoretical_loss": 3.3499587848170065,
      "tokens_seen": 2680619008
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1469157,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4436075687408447,
      "objective/train/theoretical_loss": 3.3499523429025895,
      "objective/train/tokens_used": 1039543776,
      "theoretical_loss": 3.3499523429025895,
      "tokens_seen": 2680684544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019098989898989899,
      "loss": 2.3605,
      "theoretical_loss": 3.3499523429025895,
      "tokens_seen": 2680684544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019096969696969697,
      "loss": 2.5353,
      "theoretical_loss": 3.3499459011897543,
      "tokens_seen": 2680750080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019094949494949495,
      "loss": 2.5845,
      "theoretical_loss": 3.3499394596784904,
      "tokens_seen": 2680815616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019092929292929294,
      "loss": 2.3071,
      "theoretical_loss": 3.349933018368786,
      "tokens_seen": 2680881152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019090909090909092,
      "loss": 2.2449,
      "theoretical_loss": 3.3499265772606304,
      "tokens_seen": 2680946688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019088888888888888,
      "loss": 2.3663,
      "theoretical_loss": 3.349920136354012,
      "tokens_seen": 2681012224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001908686868686869,
      "loss": 2.5076,
      "theoretical_loss": 3.3499136956489197,
      "tokens_seen": 2681077760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019084848484848485,
      "loss": 2.4093,
      "theoretical_loss": 3.3499072551453426,
      "tokens_seen": 2681143296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019082828282828283,
      "loss": 2.3722,
      "theoretical_loss": 3.349900814843269,
      "tokens_seen": 2681208832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019080808080808081,
      "loss": 2.6375,
      "theoretical_loss": 3.3498943747426875,
      "tokens_seen": 2681274368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001907878787878788,
      "loss": 2.4641,
      "theoretical_loss": 3.3498879348435873,
      "tokens_seen": 2681339904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019076767676767678,
      "loss": 2.4708,
      "theoretical_loss": 3.3498814951459575,
      "tokens_seen": 2681405440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019074747474747474,
      "loss": 2.0143,
      "theoretical_loss": 3.349875055649786,
      "tokens_seen": 2681470976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019072727272727272,
      "loss": 2.558,
      "theoretical_loss": 3.349868616355062,
      "tokens_seen": 2681536512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001907070707070707,
      "loss": 2.3703,
      "theoretical_loss": 3.3498621772617745,
      "tokens_seen": 2681602048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001906868686868687,
      "loss": 2.5092,
      "theoretical_loss": 3.349855738369912,
      "tokens_seen": 2681667584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019066666666666668,
      "loss": 2.6118,
      "theoretical_loss": 3.3498492996794633,
      "tokens_seen": 2681733120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019064646464646466,
      "loss": 2.4143,
      "theoretical_loss": 3.3498428611904174,
      "tokens_seen": 2681798656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019062626262626262,
      "loss": 2.3872,
      "theoretical_loss": 3.3498364229027624,
      "tokens_seen": 2681864192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019060606060606063,
      "loss": 2.4397,
      "theoretical_loss": 3.349829984816488,
      "tokens_seen": 2681929728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019058585858585858,
      "loss": 2.3165,
      "theoretical_loss": 3.3498235469315825,
      "tokens_seen": 2681995264
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019056565656565657,
      "loss": 2.3202,
      "theoretical_loss": 3.3498171092480344,
      "tokens_seen": 2682060800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019054545454545455,
      "loss": 2.4997,
      "theoretical_loss": 3.3498106717658334,
      "tokens_seen": 2682126336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001905252525252525,
      "loss": 2.6068,
      "theoretical_loss": 3.3498042344849672,
      "tokens_seen": 2682191872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019050505050505052,
      "loss": 2.5874,
      "theoretical_loss": 3.3497977974054254,
      "tokens_seen": 2682257408
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1470338,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2492127418518066,
      "objective/train/theoretical_loss": 3.3497913605271963,
      "objective/train/tokens_used": 1041182176,
      "theoretical_loss": 3.3497913605271963,
      "tokens_seen": 2682322944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019048484848484848,
      "loss": 2.3631,
      "theoretical_loss": 3.3497913605271963,
      "tokens_seen": 2682322944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001904646464646465,
      "loss": 2.4778,
      "theoretical_loss": 3.349784923850269,
      "tokens_seen": 2682388480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019044444444444444,
      "loss": 2.3509,
      "theoretical_loss": 3.349778487374632,
      "tokens_seen": 2682454016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019042424242424243,
      "loss": 2.7854,
      "theoretical_loss": 3.349772051100274,
      "tokens_seen": 2682519552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001904040404040404,
      "loss": 2.2866,
      "theoretical_loss": 3.3497656150271844,
      "tokens_seen": 2682585088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019038383838383837,
      "loss": 2.4393,
      "theoretical_loss": 3.3497591791553516,
      "tokens_seen": 2682650624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019036363636363638,
      "loss": 2.537,
      "theoretical_loss": 3.349752743484764,
      "tokens_seen": 2682716160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019034343434343434,
      "loss": 2.4212,
      "theoretical_loss": 3.349746308015411,
      "tokens_seen": 2682781696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019032323232323232,
      "loss": 2.5074,
      "theoretical_loss": 3.349739872747281,
      "tokens_seen": 2682847232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001903030303030303,
      "loss": 2.4737,
      "theoretical_loss": 3.3497334376803636,
      "tokens_seen": 2682912768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001902828282828283,
      "loss": 2.4547,
      "theoretical_loss": 3.3497270028146464,
      "tokens_seen": 2682978304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019026262626262627,
      "loss": 2.3392,
      "theoretical_loss": 3.349720568150119,
      "tokens_seen": 2683043840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019024242424242426,
      "loss": 2.4746,
      "theoretical_loss": 3.3497141336867697,
      "tokens_seen": 2683109376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019022222222222221,
      "loss": 2.3469,
      "theoretical_loss": 3.3497076994245876,
      "tokens_seen": 2683174912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001902020202020202,
      "loss": 2.4776,
      "theoretical_loss": 3.349701265363562,
      "tokens_seen": 2683240448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019018181818181818,
      "loss": 2.4887,
      "theoretical_loss": 3.3496948315036805,
      "tokens_seen": 2683305984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019016161616161617,
      "loss": 2.2377,
      "theoretical_loss": 3.3496883978449326,
      "tokens_seen": 2683371520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019014141414141415,
      "loss": 2.481,
      "theoretical_loss": 3.3496819643873073,
      "tokens_seen": 2683437056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001901212121212121,
      "loss": 2.2663,
      "theoretical_loss": 3.349675531130793,
      "tokens_seen": 2683502592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019010101010101012,
      "loss": 2.3078,
      "theoretical_loss": 3.3496690980753785,
      "tokens_seen": 2683568128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019008080808080807,
      "loss": 2.5163,
      "theoretical_loss": 3.349662665221053,
      "tokens_seen": 2683633664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019006060606060609,
      "loss": 2.581,
      "theoretical_loss": 3.349656232567805,
      "tokens_seen": 2683699200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019004040404040404,
      "loss": 2.6033,
      "theoretical_loss": 3.3496498001156234,
      "tokens_seen": 2683764736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019002020202020203,
      "loss": 2.2805,
      "theoretical_loss": 3.349643367864497,
      "tokens_seen": 2683830272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019,
      "loss": 2.7595,
      "theoretical_loss": 3.3496369358144142,
      "tokens_seen": 2683895808
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1470971,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.071140766143799,
      "objective/train/theoretical_loss": 3.349630503965365,
      "objective/train/tokens_used": 1042820576,
      "theoretical_loss": 3.349630503965365,
      "tokens_seen": 2683961344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018997979797979797,
      "loss": 2.3349,
      "theoretical_loss": 3.349630503965365,
      "tokens_seen": 2683961344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018995959595959598,
      "loss": 2.6993,
      "theoretical_loss": 3.3496240723173365,
      "tokens_seen": 2684026880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018993939393939394,
      "loss": 2.52,
      "theoretical_loss": 3.349617640870319,
      "tokens_seen": 2684092416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018991919191919195,
      "loss": 2.5645,
      "theoretical_loss": 3.3496112096243005,
      "tokens_seen": 2684157952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001898989898989899,
      "loss": 2.5377,
      "theoretical_loss": 3.3496047785792697,
      "tokens_seen": 2684223488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018987878787878786,
      "loss": 2.2663,
      "theoretical_loss": 3.349598347735216,
      "tokens_seen": 2684289024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018985858585858587,
      "loss": 2.5405,
      "theoretical_loss": 3.3495919170921282,
      "tokens_seen": 2684354560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018983838383838383,
      "loss": 2.6888,
      "theoretical_loss": 3.3495854866499943,
      "tokens_seen": 2684420096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018981818181818184,
      "loss": 2.3649,
      "theoretical_loss": 3.349579056408804,
      "tokens_seen": 2684485632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001897979797979798,
      "loss": 2.7438,
      "theoretical_loss": 3.349572626368546,
      "tokens_seen": 2684551168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018977777777777778,
      "loss": 2.6526,
      "theoretical_loss": 3.3495661965292087,
      "tokens_seen": 2684616704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018975757575757576,
      "loss": 2.5452,
      "theoretical_loss": 3.349559766890781,
      "tokens_seen": 2684682240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018973737373737375,
      "loss": 2.6702,
      "theoretical_loss": 3.3495533374532522,
      "tokens_seen": 2684747776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018971717171717173,
      "loss": 2.5469,
      "theoretical_loss": 3.3495469082166105,
      "tokens_seen": 2684813312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001896969696969697,
      "loss": 2.2317,
      "theoretical_loss": 3.3495404791808445,
      "tokens_seen": 2684878848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018967676767676767,
      "loss": 2.495,
      "theoretical_loss": 3.349534050345944,
      "tokens_seen": 2684944384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018965656565656566,
      "loss": 2.2157,
      "theoretical_loss": 3.3495276217118977,
      "tokens_seen": 2685009920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018963636363636364,
      "loss": 2.5796,
      "theoretical_loss": 3.3495211932786937,
      "tokens_seen": 2685075456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018961616161616162,
      "loss": 2.23,
      "theoretical_loss": 3.3495147650463206,
      "tokens_seen": 2685140992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001895959595959596,
      "loss": 2.6779,
      "theoretical_loss": 3.3495083370147682,
      "tokens_seen": 2685206528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018957575757575757,
      "loss": 2.6145,
      "theoretical_loss": 3.3495019091840255,
      "tokens_seen": 2685272064
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018955555555555558,
      "loss": 2.5596,
      "theoretical_loss": 3.3494954815540803,
      "tokens_seen": 2685337600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018953535353535353,
      "loss": 2.4005,
      "theoretical_loss": 3.3494890541249216,
      "tokens_seen": 2685403136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018951515151515152,
      "loss": 2.4784,
      "theoretical_loss": 3.3494826268965388,
      "tokens_seen": 2685468672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001894949494949495,
      "loss": 2.5334,
      "theoretical_loss": 3.34947619986892,
      "tokens_seen": 2685534208
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1472231,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9826325178146362,
      "objective/train/theoretical_loss": 3.349469773042055,
      "objective/train/tokens_used": 1044458976,
      "theoretical_loss": 3.349469773042055,
      "tokens_seen": 2685599744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018947474747474746,
      "loss": 2.487,
      "theoretical_loss": 3.349469773042055,
      "tokens_seen": 2685599744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018945454545454547,
      "loss": 2.3575,
      "theoretical_loss": 3.3494633464159316,
      "tokens_seen": 2685665280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018943434343434343,
      "loss": 2.1996,
      "theoretical_loss": 3.3494569199905397,
      "tokens_seen": 2685730816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018941414141414144,
      "loss": 2.4829,
      "theoretical_loss": 3.3494504937658673,
      "tokens_seen": 2685796352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001893939393939394,
      "loss": 2.3797,
      "theoretical_loss": 3.3494440677419033,
      "tokens_seen": 2685861888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018937373737373738,
      "loss": 2.5123,
      "theoretical_loss": 3.349437641918637,
      "tokens_seen": 2685927424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018935353535353536,
      "loss": 2.5254,
      "theoretical_loss": 3.3494312162960567,
      "tokens_seen": 2685992960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018933333333333332,
      "loss": 2.4118,
      "theoretical_loss": 3.3494247908741515,
      "tokens_seen": 2686058496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018931313131313133,
      "loss": 2.2464,
      "theoretical_loss": 3.34941836565291,
      "tokens_seen": 2686124032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001892929292929293,
      "loss": 2.5086,
      "theoretical_loss": 3.3494119406323217,
      "tokens_seen": 2686189568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018927272727272727,
      "loss": 2.4341,
      "theoretical_loss": 3.349405515812375,
      "tokens_seen": 2686255104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018925252525252525,
      "loss": 2.4747,
      "theoretical_loss": 3.3493990911930585,
      "tokens_seen": 2686320640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018923232323232324,
      "loss": 2.3618,
      "theoretical_loss": 3.3493926667743614,
      "tokens_seen": 2686386176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018921212121212122,
      "loss": 2.2016,
      "theoretical_loss": 3.349386242556273,
      "tokens_seen": 2686451712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001891919191919192,
      "loss": 2.4484,
      "theoretical_loss": 3.3493798185387806,
      "tokens_seen": 2686517248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001891717171717172,
      "loss": 2.5259,
      "theoretical_loss": 3.3493733947218742,
      "tokens_seen": 2686582784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018915151515151515,
      "loss": 2.4334,
      "theoretical_loss": 3.3493669711055425,
      "tokens_seen": 2686648320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018913131313131313,
      "loss": 2.3139,
      "theoretical_loss": 3.3493605476897748,
      "tokens_seen": 2686713856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018911111111111112,
      "loss": 2.4416,
      "theoretical_loss": 3.349354124474559,
      "tokens_seen": 2686779392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001890909090909091,
      "loss": 2.4614,
      "theoretical_loss": 3.3493477014598843,
      "tokens_seen": 2686844928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018907070707070708,
      "loss": 2.4307,
      "theoretical_loss": 3.34934127864574,
      "tokens_seen": 2686910464
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018905050505050507,
      "loss": 2.441,
      "theoretical_loss": 3.3493348560321143,
      "tokens_seen": 2686976000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018903030303030302,
      "loss": 2.6179,
      "theoretical_loss": 3.3493284336189966,
      "tokens_seen": 2687041536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000189010101010101,
      "loss": 2.2023,
      "theoretical_loss": 3.3493220114063753,
      "tokens_seen": 2687107072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000188989898989899,
      "loss": 2.4574,
      "theoretical_loss": 3.3493155893942395,
      "tokens_seen": 2687172608
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1473498,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5242228507995605,
      "objective/train/theoretical_loss": 3.349309167582578,
      "objective/train/tokens_used": 1046097376,
      "theoretical_loss": 3.349309167582578,
      "tokens_seen": 2687238144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018896969696969698,
      "loss": 2.4017,
      "theoretical_loss": 3.349309167582578,
      "tokens_seen": 2687238144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018894949494949496,
      "loss": 2.4346,
      "theoretical_loss": 3.349302745971379,
      "tokens_seen": 2687303680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018892929292929292,
      "loss": 2.5537,
      "theoretical_loss": 3.349296324560633,
      "tokens_seen": 2687369216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018890909090909093,
      "loss": 2.4903,
      "theoretical_loss": 3.349289903350327,
      "tokens_seen": 2687434752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018888888888888888,
      "loss": 2.3062,
      "theoretical_loss": 3.3492834823404514,
      "tokens_seen": 2687500288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001888686868686869,
      "loss": 2.3404,
      "theoretical_loss": 3.349277061530994,
      "tokens_seen": 2687565824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018884848484848485,
      "loss": 2.7609,
      "theoretical_loss": 3.349270640921944,
      "tokens_seen": 2687631360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001888282828282828,
      "loss": 2.4021,
      "theoretical_loss": 3.3492642205132905,
      "tokens_seen": 2687696896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018880808080808082,
      "loss": 2.3577,
      "theoretical_loss": 3.349257800305022,
      "tokens_seen": 2687762432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018878787878787878,
      "loss": 2.4386,
      "theoretical_loss": 3.3492513802971278,
      "tokens_seen": 2687827968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001887676767676768,
      "loss": 2.2587,
      "theoretical_loss": 3.349244960489596,
      "tokens_seen": 2687893504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018874747474747475,
      "loss": 2.2947,
      "theoretical_loss": 3.349238540882416,
      "tokens_seen": 2687959040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018872727272727273,
      "loss": 2.2574,
      "theoretical_loss": 3.349232121475577,
      "tokens_seen": 2688024576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001887070707070707,
      "loss": 2.2962,
      "theoretical_loss": 3.3492257022690666,
      "tokens_seen": 2688090112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001886868686868687,
      "loss": 2.3811,
      "theoretical_loss": 3.349219283262875,
      "tokens_seen": 2688155648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018866666666666668,
      "loss": 2.4024,
      "theoretical_loss": 3.349212864456991,
      "tokens_seen": 2688221184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018864646464646464,
      "loss": 2.4508,
      "theoretical_loss": 3.3492064458514026,
      "tokens_seen": 2688286720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018862626262626262,
      "loss": 2.6113,
      "theoretical_loss": 3.349200027446099,
      "tokens_seen": 2688352256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001886060606060606,
      "loss": 2.6725,
      "theoretical_loss": 3.3491936092410692,
      "tokens_seen": 2688417792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001885858585858586,
      "loss": 2.5437,
      "theoretical_loss": 3.349187191236302,
      "tokens_seen": 2688483328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018856565656565657,
      "loss": 2.4646,
      "theoretical_loss": 3.3491807734317867,
      "tokens_seen": 2688548864
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018854545454545456,
      "loss": 2.3727,
      "theoretical_loss": 3.3491743558275116,
      "tokens_seen": 2688614400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018852525252525251,
      "loss": 2.3472,
      "theoretical_loss": 3.3491679384234656,
      "tokens_seen": 2688679936
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018850505050505053,
      "loss": 2.1493,
      "theoretical_loss": 3.349161521219638,
      "tokens_seen": 2688745472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018848484848484848,
      "loss": 2.3411,
      "theoretical_loss": 3.3491551042160173,
      "tokens_seen": 2688811008
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1474112,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3311548233032227,
      "objective/train/theoretical_loss": 3.349148687412592,
      "objective/train/tokens_used": 1047735776,
      "theoretical_loss": 3.349148687412592,
      "tokens_seen": 2688876544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018846464646464647,
      "loss": 2.2077,
      "theoretical_loss": 3.349148687412592,
      "tokens_seen": 2688876544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018844444444444445,
      "loss": 2.4615,
      "theoretical_loss": 3.3491422708093523,
      "tokens_seen": 2688942080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001884242424242424,
      "loss": 2.4166,
      "theoretical_loss": 3.3491358544062857,
      "tokens_seen": 2689007616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018840404040404042,
      "loss": 2.7887,
      "theoretical_loss": 3.3491294382033816,
      "tokens_seen": 2689073152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018838383838383838,
      "loss": 2.355,
      "theoretical_loss": 3.3491230222006294,
      "tokens_seen": 2689138688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018836363636363639,
      "loss": 2.5055,
      "theoretical_loss": 3.349116606398017,
      "tokens_seen": 2689204224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018834343434343434,
      "loss": 2.3707,
      "theoretical_loss": 3.3491101907955336,
      "tokens_seen": 2689269760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018832323232323235,
      "loss": 2.389,
      "theoretical_loss": 3.3491037753931687,
      "tokens_seen": 2689335296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001883030303030303,
      "loss": 2.4807,
      "theoretical_loss": 3.3490973601909104,
      "tokens_seen": 2689400832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018828282828282827,
      "loss": 2.3522,
      "theoretical_loss": 3.349090945188748,
      "tokens_seen": 2689466368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018826262626262628,
      "loss": 2.5167,
      "theoretical_loss": 3.34908453038667,
      "tokens_seen": 2689531904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018824242424242424,
      "loss": 2.6034,
      "theoretical_loss": 3.349078115784666,
      "tokens_seen": 2689597440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018822222222222225,
      "loss": 2.2696,
      "theoretical_loss": 3.3490717013827243,
      "tokens_seen": 2689662976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001882020202020202,
      "loss": 2.5018,
      "theoretical_loss": 3.3490652871808337,
      "tokens_seen": 2689728512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001881818181818182,
      "loss": 2.406,
      "theoretical_loss": 3.3490588731789837,
      "tokens_seen": 2689794048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018816161616161617,
      "loss": 2.464,
      "theoretical_loss": 3.3490524593771624,
      "tokens_seen": 2689859584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018814141414141413,
      "loss": 2.5187,
      "theoretical_loss": 3.349046045775359,
      "tokens_seen": 2689925120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018812121212121214,
      "loss": 2.5216,
      "theoretical_loss": 3.3490396323735627,
      "tokens_seen": 2689990656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001881010101010101,
      "loss": 2.3615,
      "theoretical_loss": 3.3490332191717624,
      "tokens_seen": 2690056192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018808080808080808,
      "loss": 2.584,
      "theoretical_loss": 3.3490268061699466,
      "tokens_seen": 2690121728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018806060606060606,
      "loss": 2.2031,
      "theoretical_loss": 3.349020393368104,
      "tokens_seen": 2690187264
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018804040404040405,
      "loss": 2.4321,
      "theoretical_loss": 3.349013980766224,
      "tokens_seen": 2690252800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018802020202020203,
      "loss": 2.3794,
      "theoretical_loss": 3.349007568364296,
      "tokens_seen": 2690318336
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018800000000000002,
      "loss": 2.3306,
      "theoretical_loss": 3.3490011561623074,
      "tokens_seen": 2690383872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018797979797979797,
      "loss": 2.6744,
      "theoretical_loss": 3.3489947441602483,
      "tokens_seen": 2690449408
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1475320,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5901687145233154,
      "objective/train/theoretical_loss": 3.348988332358107,
      "objective/train/tokens_used": 1049374176,
      "theoretical_loss": 3.348988332358107,
      "tokens_seen": 2690514944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018795959595959596,
      "loss": 2.7233,
      "theoretical_loss": 3.348988332358107,
      "tokens_seen": 2690514944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018793939393939394,
      "loss": 2.6352,
      "theoretical_loss": 3.348981920755873,
      "tokens_seen": 2690580480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018791919191919192,
      "loss": 2.5087,
      "theoretical_loss": 3.3489755093535347,
      "tokens_seen": 2690646016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001878989898989899,
      "loss": 2.523,
      "theoretical_loss": 3.348969098151081,
      "tokens_seen": 2690711552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018787878787878787,
      "loss": 2.3825,
      "theoretical_loss": 3.348962687148501,
      "tokens_seen": 2690777088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018785858585858588,
      "loss": 2.539,
      "theoretical_loss": 3.3489562763457834,
      "tokens_seen": 2690842624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018783838383838383,
      "loss": 2.151,
      "theoretical_loss": 3.348949865742917,
      "tokens_seen": 2690908160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018781818181818184,
      "loss": 2.4829,
      "theoretical_loss": 3.3489434553398914,
      "tokens_seen": 2690973696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001877979797979798,
      "loss": 2.412,
      "theoretical_loss": 3.3489370451366947,
      "tokens_seen": 2691039232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018777777777777776,
      "loss": 2.6671,
      "theoretical_loss": 3.3489306351333163,
      "tokens_seen": 2691104768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018775757575757577,
      "loss": 2.5635,
      "theoretical_loss": 3.3489242253297453,
      "tokens_seen": 2691170304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018773737373737373,
      "loss": 2.5179,
      "theoretical_loss": 3.3489178157259696,
      "tokens_seen": 2691235840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018771717171717174,
      "loss": 2.4773,
      "theoretical_loss": 3.348911406321979,
      "tokens_seen": 2691301376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001876969696969697,
      "loss": 2.5822,
      "theoretical_loss": 3.3489049971177622,
      "tokens_seen": 2691366912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018767676767676768,
      "loss": 2.3443,
      "theoretical_loss": 3.348898588113308,
      "tokens_seen": 2691432448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018765656565656566,
      "loss": 2.486,
      "theoretical_loss": 3.3488921793086055,
      "tokens_seen": 2691497984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018763636363636365,
      "loss": 2.7644,
      "theoretical_loss": 3.3488857707036432,
      "tokens_seen": 2691563520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018761616161616163,
      "loss": 2.3513,
      "theoretical_loss": 3.348879362298411,
      "tokens_seen": 2691629056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001875959595959596,
      "loss": 2.57,
      "theoretical_loss": 3.3488729540928963,
      "tokens_seen": 2691694592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018757575757575757,
      "loss": 2.4854,
      "theoretical_loss": 3.3488665460870894,
      "tokens_seen": 2691760128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018755555555555555,
      "loss": 2.5739,
      "theoretical_loss": 3.3488601382809784,
      "tokens_seen": 2691825664
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018753535353535354,
      "loss": 2.4218,
      "theoretical_loss": 3.348853730674552,
      "tokens_seen": 2691891200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018751515151515152,
      "loss": 2.2809,
      "theoretical_loss": 3.3488473232678,
      "tokens_seen": 2691956736
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001874949494949495,
      "loss": 2.6068,
      "theoretical_loss": 3.3488409160607113,
      "tokens_seen": 2692022272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001874747474747475,
      "loss": 2.512,
      "theoretical_loss": 3.348834509053274,
      "tokens_seen": 2692087808
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1476480,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.195870876312256,
      "objective/train/theoretical_loss": 3.3488281022454776,
      "objective/train/tokens_used": 1051012576,
      "theoretical_loss": 3.3488281022454776,
      "tokens_seen": 2692153344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018745454545454545,
      "loss": 2.2735,
      "theoretical_loss": 3.3488281022454776,
      "tokens_seen": 2692153344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018743434343434343,
      "loss": 2.5262,
      "theoretical_loss": 3.3488216956373105,
      "tokens_seen": 2692218880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018741414141414142,
      "loss": 2.4886,
      "theoretical_loss": 3.3488152892287624,
      "tokens_seen": 2692284416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001873939393939394,
      "loss": 2.4564,
      "theoretical_loss": 3.3488088830198213,
      "tokens_seen": 2692349952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018737373737373738,
      "loss": 2.3529,
      "theoretical_loss": 3.348802477010477,
      "tokens_seen": 2692415488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018735353535353537,
      "loss": 2.3423,
      "theoretical_loss": 3.348796071200718,
      "tokens_seen": 2692481024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018733333333333332,
      "loss": 2.48,
      "theoretical_loss": 3.348789665590533,
      "tokens_seen": 2692546560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018731313131313134,
      "loss": 2.4425,
      "theoretical_loss": 3.3487832601799115,
      "tokens_seen": 2692612096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001872929292929293,
      "loss": 2.3937,
      "theoretical_loss": 3.3487768549688415,
      "tokens_seen": 2692677632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018727272727272728,
      "loss": 2.6744,
      "theoretical_loss": 3.348770449957313,
      "tokens_seen": 2692743168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018725252525252526,
      "loss": 2.4658,
      "theoretical_loss": 3.3487640451453147,
      "tokens_seen": 2692808704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018723232323232322,
      "loss": 2.5764,
      "theoretical_loss": 3.3487576405328348,
      "tokens_seen": 2692874240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018721212121212123,
      "loss": 2.52,
      "theoretical_loss": 3.3487512361198633,
      "tokens_seen": 2692939776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018719191919191918,
      "loss": 2.5945,
      "theoretical_loss": 3.348744831906388,
      "tokens_seen": 2693005312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001871717171717172,
      "loss": 2.6165,
      "theoretical_loss": 3.348738427892399,
      "tokens_seen": 2693070848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018715151515151515,
      "loss": 2.4594,
      "theoretical_loss": 3.3487320240778837,
      "tokens_seen": 2693136384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018713131313131314,
      "loss": 2.3541,
      "theoretical_loss": 3.3487256204628326,
      "tokens_seen": 2693201920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018711111111111112,
      "loss": 2.4568,
      "theoretical_loss": 3.3487192170472335,
      "tokens_seen": 2693267456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018709090909090908,
      "loss": 2.5012,
      "theoretical_loss": 3.348712813831076,
      "tokens_seen": 2693332992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001870707070707071,
      "loss": 2.4288,
      "theoretical_loss": 3.3487064108143496,
      "tokens_seen": 2693398528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018705050505050505,
      "loss": 2.7131,
      "theoretical_loss": 3.3487000079970417,
      "tokens_seen": 2693464064
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018703030303030303,
      "loss": 2.1204,
      "theoretical_loss": 3.3486936053791423,
      "tokens_seen": 2693529600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000187010101010101,
      "loss": 2.4358,
      "theoretical_loss": 3.34868720296064,
      "tokens_seen": 2693595136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000186989898989899,
      "loss": 2.1905,
      "theoretical_loss": 3.3486808007415236,
      "tokens_seen": 2693660672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018696969696969698,
      "loss": 2.3522,
      "theoretical_loss": 3.348674398721782,
      "tokens_seen": 2693726208
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1476942,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.778862237930298,
      "objective/train/theoretical_loss": 3.348667996901405,
      "objective/train/tokens_used": 1052650976,
      "theoretical_loss": 3.348667996901405,
      "tokens_seen": 2693791744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018694949494949497,
      "loss": 2.6326,
      "theoretical_loss": 3.348667996901405,
      "tokens_seen": 2693791744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018692929292929292,
      "loss": 2.2662,
      "theoretical_loss": 3.348661595280381,
      "tokens_seen": 2693857280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001869090909090909,
      "loss": 2.5923,
      "theoretical_loss": 3.3486551938586984,
      "tokens_seen": 2693922816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001868888888888889,
      "loss": 2.2519,
      "theoretical_loss": 3.3486487926363466,
      "tokens_seen": 2693988352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018686868686868687,
      "loss": 2.4467,
      "theoretical_loss": 3.348642391613315,
      "tokens_seen": 2694053888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018684848484848486,
      "loss": 2.4839,
      "theoretical_loss": 3.3486359907895915,
      "tokens_seen": 2694119424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018682828282828281,
      "loss": 2.2833,
      "theoretical_loss": 3.348629590165166,
      "tokens_seen": 2694184960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018680808080808083,
      "loss": 2.4663,
      "theoretical_loss": 3.348623189740027,
      "tokens_seen": 2694250496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018678787878787878,
      "loss": 2.5509,
      "theoretical_loss": 3.3486167895141636,
      "tokens_seen": 2694316032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001867676767676768,
      "loss": 2.5238,
      "theoretical_loss": 3.3486103894875647,
      "tokens_seen": 2694381568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018674747474747475,
      "loss": 2.4002,
      "theoretical_loss": 3.3486039896602193,
      "tokens_seen": 2694447104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001867272727272727,
      "loss": 2.2114,
      "theoretical_loss": 3.3485975900321163,
      "tokens_seen": 2694512640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018670707070707072,
      "loss": 2.559,
      "theoretical_loss": 3.348591190603244,
      "tokens_seen": 2694578176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018668686868686868,
      "loss": 2.4803,
      "theoretical_loss": 3.348584791373593,
      "tokens_seen": 2694643712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001866666666666667,
      "loss": 2.5313,
      "theoretical_loss": 3.3485783923431507,
      "tokens_seen": 2694709248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018664646464646464,
      "loss": 2.5669,
      "theoretical_loss": 3.3485719935119067,
      "tokens_seen": 2694774784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018662626262626265,
      "loss": 2.7007,
      "theoretical_loss": 3.3485655948798496,
      "tokens_seen": 2694840320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001866060606060606,
      "loss": 2.5186,
      "theoretical_loss": 3.3485591964469688,
      "tokens_seen": 2694905856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018658585858585857,
      "loss": 2.4547,
      "theoretical_loss": 3.348552798213253,
      "tokens_seen": 2694971392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018656565656565658,
      "loss": 2.587,
      "theoretical_loss": 3.3485464001786918,
      "tokens_seen": 2695036928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018654545454545454,
      "loss": 2.353,
      "theoretical_loss": 3.348540002343273,
      "tokens_seen": 2695102464
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018652525252525255,
      "loss": 2.1833,
      "theoretical_loss": 3.348533604706986,
      "tokens_seen": 2695168000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001865050505050505,
      "loss": 2.434,
      "theoretical_loss": 3.3485272072698202,
      "tokens_seen": 2695233536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001864848484848485,
      "loss": 2.5977,
      "theoretical_loss": 3.348520810031764,
      "tokens_seen": 2695299072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018646464646464647,
      "loss": 2.2726,
      "theoretical_loss": 3.348514412992807,
      "tokens_seen": 2695364608
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1478003,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5612661838531494,
      "objective/train/theoretical_loss": 3.3485080161529375,
      "objective/train/tokens_used": 1054289376,
      "theoretical_loss": 3.3485080161529375,
      "tokens_seen": 2695430144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018644444444444446,
      "loss": 2.4453,
      "theoretical_loss": 3.3485080161529375,
      "tokens_seen": 2695430144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018642424242424244,
      "loss": 2.5392,
      "theoretical_loss": 3.348501619512145,
      "tokens_seen": 2695495680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001864040404040404,
      "loss": 2.5594,
      "theoretical_loss": 3.348495223070418,
      "tokens_seen": 2695561216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018638383838383838,
      "loss": 2.3624,
      "theoretical_loss": 3.3484888268277455,
      "tokens_seen": 2695626752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018636363636363636,
      "loss": 2.6489,
      "theoretical_loss": 3.348482430784117,
      "tokens_seen": 2695692288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018634343434343435,
      "loss": 2.4926,
      "theoretical_loss": 3.3484760349395213,
      "tokens_seen": 2695757824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018632323232323233,
      "loss": 2.5222,
      "theoretical_loss": 3.3484696392939464,
      "tokens_seen": 2695823360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018630303030303032,
      "loss": 2.3682,
      "theoretical_loss": 3.348463243847383,
      "tokens_seen": 2695888896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018628282828282827,
      "loss": 2.5248,
      "theoretical_loss": 3.3484568485998185,
      "tokens_seen": 2695954432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018626262626262628,
      "loss": 2.5558,
      "theoretical_loss": 3.3484504535512425,
      "tokens_seen": 2696019968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018624242424242424,
      "loss": 2.4428,
      "theoretical_loss": 3.348444058701644,
      "tokens_seen": 2696085504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018622222222222223,
      "loss": 2.3469,
      "theoretical_loss": 3.348437664051012,
      "tokens_seen": 2696151040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001862020202020202,
      "loss": 2.4062,
      "theoretical_loss": 3.3484312695993355,
      "tokens_seen": 2696216576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018618181818181817,
      "loss": 2.6158,
      "theoretical_loss": 3.3484248753466033,
      "tokens_seen": 2696282112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018616161616161618,
      "loss": 2.444,
      "theoretical_loss": 3.3484184812928044,
      "tokens_seen": 2696347648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018614141414141413,
      "loss": 2.5906,
      "theoretical_loss": 3.348412087437928,
      "tokens_seen": 2696413184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018612121212121214,
      "loss": 2.6042,
      "theoretical_loss": 3.348405693781963,
      "tokens_seen": 2696478720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001861010101010101,
      "loss": 2.6892,
      "theoretical_loss": 3.348399300324898,
      "tokens_seen": 2696544256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018608080808080809,
      "loss": 2.6522,
      "theoretical_loss": 3.348392907066722,
      "tokens_seen": 2696609792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018606060606060607,
      "loss": 2.6185,
      "theoretical_loss": 3.3483865140074247,
      "tokens_seen": 2696675328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018604040404040403,
      "loss": 2.5195,
      "theoretical_loss": 3.348380121146995,
      "tokens_seen": 2696740864
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018602020202020204,
      "loss": 2.336,
      "theoretical_loss": 3.348373728485421,
      "tokens_seen": 2696806400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000186,
      "loss": 2.5492,
      "theoretical_loss": 3.348367336022692,
      "tokens_seen": 2696871936
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018597979797979798,
      "loss": 2.3328,
      "theoretical_loss": 3.3483609437587973,
      "tokens_seen": 2696937472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018595959595959596,
      "loss": 2.4845,
      "theoretical_loss": 3.348354551693726,
      "tokens_seen": 2697003008
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1478515,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5991339683532715,
      "objective/train/theoretical_loss": 3.348348159827467,
      "objective/train/tokens_used": 1055927776,
      "theoretical_loss": 3.348348159827467,
      "tokens_seen": 2697068544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018593939393939395,
      "loss": 2.5127,
      "theoretical_loss": 3.348348159827467,
      "tokens_seen": 2697068544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018591919191919193,
      "loss": 2.3531,
      "theoretical_loss": 3.3483417681600085,
      "tokens_seen": 2697134080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018589898989898991,
      "loss": 2.4944,
      "theoretical_loss": 3.348335376691341,
      "tokens_seen": 2697199616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018587878787878787,
      "loss": 2.5394,
      "theoretical_loss": 3.3483289854214515,
      "tokens_seen": 2697265152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018585858585858586,
      "loss": 2.6762,
      "theoretical_loss": 3.3483225943503307,
      "tokens_seen": 2697330688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018583838383838384,
      "loss": 2.3825,
      "theoretical_loss": 3.348316203477967,
      "tokens_seen": 2697396224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018581818181818182,
      "loss": 2.4218,
      "theoretical_loss": 3.3483098128043496,
      "tokens_seen": 2697461760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001857979797979798,
      "loss": 2.176,
      "theoretical_loss": 3.3483034223294665,
      "tokens_seen": 2697527296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018577777777777776,
      "loss": 2.2991,
      "theoretical_loss": 3.348297032053308,
      "tokens_seen": 2697592832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018575757575757577,
      "loss": 2.4322,
      "theoretical_loss": 3.3482906419758627,
      "tokens_seen": 2697658368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018573737373737373,
      "loss": 2.5228,
      "theoretical_loss": 3.3482842520971188,
      "tokens_seen": 2697723904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018571717171717172,
      "loss": 2.5293,
      "theoretical_loss": 3.3482778624170666,
      "tokens_seen": 2697789440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001856969696969697,
      "loss": 2.4668,
      "theoretical_loss": 3.348271472935694,
      "tokens_seen": 2697854976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018567676767676768,
      "loss": 2.6747,
      "theoretical_loss": 3.348265083652991,
      "tokens_seen": 2697920512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018565656565656567,
      "loss": 2.6877,
      "theoretical_loss": 3.3482586945689454,
      "tokens_seen": 2697986048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018563636363636362,
      "loss": 2.2513,
      "theoretical_loss": 3.348252305683547,
      "tokens_seen": 2698051584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018561616161616164,
      "loss": 2.3112,
      "theoretical_loss": 3.3482459169967846,
      "tokens_seen": 2698117120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001855959595959596,
      "loss": 2.4948,
      "theoretical_loss": 3.348239528508647,
      "tokens_seen": 2698182656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001855757575757576,
      "loss": 2.4935,
      "theoretical_loss": 3.348233140219124,
      "tokens_seen": 2698248192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018555555555555556,
      "loss": 2.6531,
      "theoretical_loss": 3.3482267521282036,
      "tokens_seen": 2698313728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018553535353535352,
      "loss": 2.4196,
      "theoretical_loss": 3.348220364235875,
      "tokens_seen": 2698379264
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018551515151515153,
      "loss": 2.419,
      "theoretical_loss": 3.348213976542128,
      "tokens_seen": 2698444800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018549494949494949,
      "loss": 2.4405,
      "theoretical_loss": 3.3482075890469507,
      "tokens_seen": 2698510336
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001854747474747475,
      "loss": 2.3262,
      "theoretical_loss": 3.3482012017503324,
      "tokens_seen": 2698575872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018545454545454545,
      "loss": 2.4757,
      "theoretical_loss": 3.3481948146522624,
      "tokens_seen": 2698641408
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1479744,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.583965301513672,
      "objective/train/theoretical_loss": 3.348188427752729,
      "objective/train/tokens_used": 1057566176,
      "theoretical_loss": 3.348188427752729,
      "tokens_seen": 2698706944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018543434343434344,
      "loss": 2.4659,
      "theoretical_loss": 3.348188427752729,
      "tokens_seen": 2698706944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018541414141414142,
      "loss": 2.4803,
      "theoretical_loss": 3.348182041051722,
      "tokens_seen": 2698772480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001853939393939394,
      "loss": 2.6625,
      "theoretical_loss": 3.34817565454923,
      "tokens_seen": 2698838016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001853737373737374,
      "loss": 2.5764,
      "theoretical_loss": 3.348169268245242,
      "tokens_seen": 2698903552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018535353535353535,
      "loss": 2.2431,
      "theoretical_loss": 3.3481628821397473,
      "tokens_seen": 2698969088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018533333333333333,
      "loss": 2.3912,
      "theoretical_loss": 3.3481564962327344,
      "tokens_seen": 2699034624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018531313131313131,
      "loss": 2.3538,
      "theoretical_loss": 3.348150110524193,
      "tokens_seen": 2699100160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001852929292929293,
      "loss": 2.4493,
      "theoretical_loss": 3.3481437250141113,
      "tokens_seen": 2699165696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018527272727272728,
      "loss": 2.4769,
      "theoretical_loss": 3.348137339702479,
      "tokens_seen": 2699231232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018525252525252527,
      "loss": 2.416,
      "theoretical_loss": 3.3481309545892843,
      "tokens_seen": 2699296768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018523232323232322,
      "loss": 2.3809,
      "theoretical_loss": 3.348124569674517,
      "tokens_seen": 2699362304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018521212121212123,
      "loss": 2.3255,
      "theoretical_loss": 3.3481181849581665,
      "tokens_seen": 2699427840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001851919191919192,
      "loss": 2.4066,
      "theoretical_loss": 3.3481118004402206,
      "tokens_seen": 2699493376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018517171717171717,
      "loss": 2.6292,
      "theoretical_loss": 3.348105416120669,
      "tokens_seen": 2699558912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018515151515151516,
      "loss": 2.4581,
      "theoretical_loss": 3.3480990319995008,
      "tokens_seen": 2699624448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018513131313131312,
      "loss": 2.5486,
      "theoretical_loss": 3.3480926480767046,
      "tokens_seen": 2699689984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018511111111111113,
      "loss": 2.3949,
      "theoretical_loss": 3.34808626435227,
      "tokens_seen": 2699755520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018509090909090908,
      "loss": 2.4495,
      "theoretical_loss": 3.3480798808261856,
      "tokens_seen": 2699821056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001850707070707071,
      "loss": 2.582,
      "theoretical_loss": 3.34807349749844,
      "tokens_seen": 2699886592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018505050505050505,
      "loss": 2.3803,
      "theoretical_loss": 3.3480671143690235,
      "tokens_seen": 2699952128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000185030303030303,
      "loss": 2.5114,
      "theoretical_loss": 3.348060731437924,
      "tokens_seen": 2700017664
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018501010101010102,
      "loss": 2.2333,
      "theoretical_loss": 3.348054348705131,
      "tokens_seen": 2700083200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018498989898989898,
      "loss": 2.4672,
      "theoretical_loss": 3.3480479661706335,
      "tokens_seen": 2700148736
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000184969696969697,
      "loss": 2.2601,
      "theoretical_loss": 3.34804158383442,
      "tokens_seen": 2700214272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018494949494949494,
      "loss": 2.4793,
      "theoretical_loss": 3.3480352016964803,
      "tokens_seen": 2700279808
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1480347,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4358396530151367,
      "objective/train/theoretical_loss": 3.348028819756803,
      "objective/train/tokens_used": 1059204576,
      "theoretical_loss": 3.348028819756803,
      "tokens_seen": 2700345344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018492929292929293,
      "loss": 2.3853,
      "theoretical_loss": 3.348028819756803,
      "tokens_seen": 2700345344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001849090909090909,
      "loss": 2.4037,
      "theoretical_loss": 3.348022438015377,
      "tokens_seen": 2700410880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001848888888888889,
      "loss": 2.5219,
      "theoretical_loss": 3.348016056472192,
      "tokens_seen": 2700476416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018486868686868688,
      "loss": 2.5485,
      "theoretical_loss": 3.3480096751272366,
      "tokens_seen": 2700541952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018484848484848484,
      "loss": 2.4248,
      "theoretical_loss": 3.3480032939805,
      "tokens_seen": 2700607488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018482828282828285,
      "loss": 2.5916,
      "theoretical_loss": 3.3479969130319702,
      "tokens_seen": 2700673024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001848080808080808,
      "loss": 2.4734,
      "theoretical_loss": 3.3479905322816377,
      "tokens_seen": 2700738560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001847878787878788,
      "loss": 2.3459,
      "theoretical_loss": 3.3479841517294906,
      "tokens_seen": 2700804096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018476767676767677,
      "loss": 2.7078,
      "theoretical_loss": 3.3479777713755188,
      "tokens_seen": 2700869632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018474747474747476,
      "loss": 2.6226,
      "theoretical_loss": 3.34797139121971,
      "tokens_seen": 2700935168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018472727272727274,
      "loss": 2.4492,
      "theoretical_loss": 3.3479650112620547,
      "tokens_seen": 2701000704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018470707070707072,
      "loss": 2.4536,
      "theoretical_loss": 3.347958631502541,
      "tokens_seen": 2701066240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018468686868686868,
      "loss": 2.3387,
      "theoretical_loss": 3.3479522519411584,
      "tokens_seen": 2701131776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018466666666666666,
      "loss": 2.4472,
      "theoretical_loss": 3.3479458725778954,
      "tokens_seen": 2701197312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018464646464646465,
      "loss": 2.355,
      "theoretical_loss": 3.3479394934127416,
      "tokens_seen": 2701262848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018462626262626263,
      "loss": 2.372,
      "theoretical_loss": 3.347933114445686,
      "tokens_seen": 2701328384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018460606060606062,
      "loss": 2.4098,
      "theoretical_loss": 3.347926735676717,
      "tokens_seen": 2701393920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018458585858585857,
      "loss": 2.5509,
      "theoretical_loss": 3.3479203571058247,
      "tokens_seen": 2701459456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018456565656565658,
      "loss": 2.3439,
      "theoretical_loss": 3.347913978732997,
      "tokens_seen": 2701524992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018454545454545454,
      "loss": 2.5472,
      "theoretical_loss": 3.347907600558224,
      "tokens_seen": 2701590528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018452525252525255,
      "loss": 2.7265,
      "theoretical_loss": 3.347901222581494,
      "tokens_seen": 2701656064
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001845050505050505,
      "loss": 2.4103,
      "theoretical_loss": 3.3478948448027963,
      "tokens_seen": 2701721600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018448484848484847,
      "loss": 2.5985,
      "theoretical_loss": 3.34788846722212,
      "tokens_seen": 2701787136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018446464646464648,
      "loss": 2.4802,
      "theoretical_loss": 3.3478820898394543,
      "tokens_seen": 2701852672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018444444444444443,
      "loss": 2.5821,
      "theoretical_loss": 3.347875712654788,
      "tokens_seen": 2701918208
    },
    {
      "epoch": 0.63,
      "objective/train/docs_used": 1481517,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.663572311401367,
      "objective/train/theoretical_loss": 3.3478693356681095,
      "objective/train/tokens_used": 1060842976,
      "theoretical_loss": 3.3478693356681095,
      "tokens_seen": 2701983744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018442424242424245,
      "loss": 2.449,
      "theoretical_loss": 3.3478693356681095,
      "tokens_seen": 2701983744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001844040404040404,
      "loss": 2.44,
      "theoretical_loss": 3.3478629588794093,
      "tokens_seen": 2702049280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018438383838383839,
      "loss": 2.698,
      "theoretical_loss": 3.347856582288675,
      "tokens_seen": 2702114816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018436363636363637,
      "loss": 2.6436,
      "theoretical_loss": 3.347850205895897,
      "tokens_seen": 2702180352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018434343434343435,
      "loss": 2.6549,
      "theoretical_loss": 3.3478438297010635,
      "tokens_seen": 2702245888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018432323232323234,
      "loss": 2.4352,
      "theoretical_loss": 3.347837453704164,
      "tokens_seen": 2702311424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001843030303030303,
      "loss": 2.4105,
      "theoretical_loss": 3.347831077905187,
      "tokens_seen": 2702376960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018428282828282828,
      "loss": 2.585,
      "theoretical_loss": 3.3478247023041217,
      "tokens_seen": 2702442496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018426262626262626,
      "loss": 2.2564,
      "theoretical_loss": 3.347818326900958,
      "tokens_seen": 2702508032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018424242424242425,
      "loss": 2.5492,
      "theoretical_loss": 3.3478119516956837,
      "tokens_seen": 2702573568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018422222222222223,
      "loss": 2.6858,
      "theoretical_loss": 3.3478055766882884,
      "tokens_seen": 2702639104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018420202020202021,
      "loss": 2.6041,
      "theoretical_loss": 3.3477992018787615,
      "tokens_seen": 2702704640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018418181818181817,
      "loss": 2.6253,
      "theoretical_loss": 3.347792827267092,
      "tokens_seen": 2702770176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018416161616161616,
      "loss": 2.306,
      "theoretical_loss": 3.3477864528532684,
      "tokens_seen": 2702835712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018414141414141414,
      "loss": 2.5949,
      "theoretical_loss": 3.34778007863728,
      "tokens_seen": 2702901248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018412121212121212,
      "loss": 2.4144,
      "theoretical_loss": 3.347773704619116,
      "tokens_seen": 2702966784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001841010101010101,
      "loss": 2.384,
      "theoretical_loss": 3.3477673307987654,
      "tokens_seen": 2703032320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018408080808080806,
      "loss": 2.5751,
      "theoretical_loss": 3.3477609571762175,
      "tokens_seen": 2703097856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018406060606060608,
      "loss": 2.4568,
      "theoretical_loss": 3.347754583751461,
      "tokens_seen": 2703163392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018404040404040403,
      "loss": 2.5424,
      "theoretical_loss": 3.347748210524485,
      "tokens_seen": 2703228928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018402020202020204,
      "loss": 2.4473,
      "theoretical_loss": 3.347741837495279,
      "tokens_seen": 2703294464
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000184,
      "loss": 2.427,
      "theoretical_loss": 3.3477354646638315,
      "tokens_seen": 2703360000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018397979797979798,
      "loss": 2.49,
      "theoretical_loss": 3.347729092030132,
      "tokens_seen": 2703425536
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018395959595959597,
      "loss": 2.4909,
      "theoretical_loss": 3.3477227195941692,
      "tokens_seen": 2703491072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018393939393939392,
      "loss": 2.5905,
      "theoretical_loss": 3.3477163473559326,
      "tokens_seen": 2703556608
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1482062,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9572315216064453,
      "objective/train/theoretical_loss": 3.3477099753154107,
      "objective/train/tokens_used": 1062481376,
      "theoretical_loss": 3.3477099753154107,
      "tokens_seen": 2703622144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018391919191919194,
      "loss": 2.3128,
      "theoretical_loss": 3.3477099753154107,
      "tokens_seen": 2703622144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001838989898989899,
      "loss": 2.4784,
      "theoretical_loss": 3.347703603472593,
      "tokens_seen": 2703687680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001838787878787879,
      "loss": 2.3243,
      "theoretical_loss": 3.3476972318274685,
      "tokens_seen": 2703753216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018385858585858586,
      "loss": 2.6366,
      "theoretical_loss": 3.3476908603800264,
      "tokens_seen": 2703818752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018383838383838384,
      "loss": 2.473,
      "theoretical_loss": 3.3476844891302555,
      "tokens_seen": 2703884288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018381818181818183,
      "loss": 2.7443,
      "theoretical_loss": 3.347678118078145,
      "tokens_seen": 2703949824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018379797979797979,
      "loss": 2.5302,
      "theoretical_loss": 3.3476717472236843,
      "tokens_seen": 2704015360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001837777777777778,
      "loss": 2.465,
      "theoretical_loss": 3.347665376566862,
      "tokens_seen": 2704080896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018375757575757575,
      "loss": 2.4601,
      "theoretical_loss": 3.3476590061076674,
      "tokens_seen": 2704146432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018373737373737374,
      "loss": 2.5635,
      "theoretical_loss": 3.347652635846089,
      "tokens_seen": 2704211968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018371717171717172,
      "loss": 2.5015,
      "theoretical_loss": 3.3476462657821173,
      "tokens_seen": 2704277504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001836969696969697,
      "loss": 2.5868,
      "theoretical_loss": 3.34763989591574,
      "tokens_seen": 2704343040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001836767676767677,
      "loss": 2.515,
      "theoretical_loss": 3.3476335262469465,
      "tokens_seen": 2704408576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018365656565656567,
      "loss": 2.533,
      "theoretical_loss": 3.3476271567757263,
      "tokens_seen": 2704474112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018363636363636363,
      "loss": 2.4978,
      "theoretical_loss": 3.347620787502068,
      "tokens_seen": 2704539648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018361616161616161,
      "loss": 2.5581,
      "theoretical_loss": 3.347614418425961,
      "tokens_seen": 2704605184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001835959595959596,
      "loss": 2.4623,
      "theoretical_loss": 3.3476080495473943,
      "tokens_seen": 2704670720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018357575757575758,
      "loss": 2.7358,
      "theoretical_loss": 3.347601680866357,
      "tokens_seen": 2704736256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018355555555555557,
      "loss": 2.4622,
      "theoretical_loss": 3.347595312382839,
      "tokens_seen": 2704801792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018353535353535352,
      "loss": 2.2301,
      "theoretical_loss": 3.3475889440968274,
      "tokens_seen": 2704867328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018351515151515153,
      "loss": 2.3475,
      "theoretical_loss": 3.347582576008313,
      "tokens_seen": 2704932864
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001834949494949495,
      "loss": 2.8454,
      "theoretical_loss": 3.3475762081172844,
      "tokens_seen": 2704998400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001834747474747475,
      "loss": 2.4604,
      "theoretical_loss": 3.34756984042373,
      "tokens_seen": 2705063936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018345454545454546,
      "loss": 2.5308,
      "theoretical_loss": 3.3475634729276402,
      "tokens_seen": 2705129472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018343434343434342,
      "loss": 2.5021,
      "theoretical_loss": 3.3475571056290034,
      "tokens_seen": 2705195008
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1483223,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.250884771347046,
      "objective/train/theoretical_loss": 3.3475507385278087,
      "objective/train/tokens_used": 1064119776,
      "theoretical_loss": 3.3475507385278087,
      "tokens_seen": 2705260544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018341414141414143,
      "loss": 2.4728,
      "theoretical_loss": 3.3475507385278087,
      "tokens_seen": 2705260544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018339393939393938,
      "loss": 2.4835,
      "theoretical_loss": 3.347544371624045,
      "tokens_seen": 2705326080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001833737373737374,
      "loss": 2.3555,
      "theoretical_loss": 3.347538004917702,
      "tokens_seen": 2705391616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018335353535353535,
      "loss": 2.4698,
      "theoretical_loss": 3.3475316384087677,
      "tokens_seen": 2705457152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018333333333333334,
      "loss": 2.5391,
      "theoretical_loss": 3.3475252720972324,
      "tokens_seen": 2705522688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018331313131313132,
      "loss": 2.5753,
      "theoretical_loss": 3.3475189059830845,
      "tokens_seen": 2705588224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018329292929292928,
      "loss": 2.4569,
      "theoretical_loss": 3.3475125400663135,
      "tokens_seen": 2705653760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001832727272727273,
      "loss": 2.7072,
      "theoretical_loss": 3.3475061743469086,
      "tokens_seen": 2705719296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018325252525252524,
      "loss": 2.3648,
      "theoretical_loss": 3.3474998088248578,
      "tokens_seen": 2705784832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018323232323232323,
      "loss": 2.618,
      "theoretical_loss": 3.3474934435001513,
      "tokens_seen": 2705850368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001832121212121212,
      "loss": 2.5072,
      "theoretical_loss": 3.347487078372778,
      "tokens_seen": 2705915904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001831919191919192,
      "loss": 2.5421,
      "theoretical_loss": 3.347480713442727,
      "tokens_seen": 2705981440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018317171717171718,
      "loss": 2.4983,
      "theoretical_loss": 3.3474743487099876,
      "tokens_seen": 2706046976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018315151515151516,
      "loss": 2.1668,
      "theoretical_loss": 3.347467984174548,
      "tokens_seen": 2706112512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018313131313131315,
      "loss": 2.5562,
      "theoretical_loss": 3.3474616198363982,
      "tokens_seen": 2706178048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001831111111111111,
      "loss": 2.4365,
      "theoretical_loss": 3.347455255695527,
      "tokens_seen": 2706243584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001830909090909091,
      "loss": 2.493,
      "theoretical_loss": 3.347448891751924,
      "tokens_seen": 2706309120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018307070707070707,
      "loss": 2.4579,
      "theoretical_loss": 3.347442528005577,
      "tokens_seen": 2706374656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018305050505050506,
      "loss": 2.5042,
      "theoretical_loss": 3.3474361644564765,
      "tokens_seen": 2706440192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018303030303030304,
      "loss": 2.6841,
      "theoretical_loss": 3.347429801104611,
      "tokens_seen": 2706505728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018301010101010102,
      "loss": 2.4324,
      "theoretical_loss": 3.3474234379499697,
      "tokens_seen": 2706571264
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018298989898989898,
      "loss": 2.2687,
      "theoretical_loss": 3.3474170749925416,
      "tokens_seen": 2706636800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000182969696969697,
      "loss": 2.6401,
      "theoretical_loss": 3.3474107122323162,
      "tokens_seen": 2706702336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018294949494949495,
      "loss": 2.4576,
      "theoretical_loss": 3.347404349669282,
      "tokens_seen": 2706767872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018292929292929293,
      "loss": 2.4525,
      "theoretical_loss": 3.3473979873034283,
      "tokens_seen": 2706833408
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1483803,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.663747787475586,
      "objective/train/theoretical_loss": 3.347391625134745,
      "objective/train/tokens_used": 1065758176,
      "theoretical_loss": 3.347391625134745,
      "tokens_seen": 2706898944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018290909090909092,
      "loss": 2.5869,
      "theoretical_loss": 3.347391625134745,
      "tokens_seen": 2706898944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018288888888888887,
      "loss": 2.1871,
      "theoretical_loss": 3.3473852631632197,
      "tokens_seen": 2706964480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018286868686868689,
      "loss": 2.4987,
      "theoretical_loss": 3.3473789013888426,
      "tokens_seen": 2707030016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018284848484848484,
      "loss": 2.4858,
      "theoretical_loss": 3.347372539811603,
      "tokens_seen": 2707095552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018282828282828285,
      "loss": 2.565,
      "theoretical_loss": 3.3473661784314896,
      "tokens_seen": 2707161088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001828080808080808,
      "loss": 2.574,
      "theoretical_loss": 3.347359817248491,
      "tokens_seen": 2707226624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001827878787878788,
      "loss": 2.551,
      "theoretical_loss": 3.3473534562625975,
      "tokens_seen": 2707292160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018276767676767678,
      "loss": 2.68,
      "theoretical_loss": 3.347347095473797,
      "tokens_seen": 2707357696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018274747474747473,
      "loss": 2.4394,
      "theoretical_loss": 3.3473407348820796,
      "tokens_seen": 2707423232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018272727272727275,
      "loss": 2.3885,
      "theoretical_loss": 3.347334374487434,
      "tokens_seen": 2707488768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001827070707070707,
      "loss": 2.2589,
      "theoretical_loss": 3.3473280142898494,
      "tokens_seen": 2707554304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001826868686868687,
      "loss": 2.605,
      "theoretical_loss": 3.3473216542893147,
      "tokens_seen": 2707619840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018266666666666667,
      "loss": 2.5402,
      "theoretical_loss": 3.347315294485819,
      "tokens_seen": 2707685376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018264646464646465,
      "loss": 2.6106,
      "theoretical_loss": 3.347308934879352,
      "tokens_seen": 2707750912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018262626262626264,
      "loss": 2.4888,
      "theoretical_loss": 3.347302575469902,
      "tokens_seen": 2707816448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018260606060606062,
      "loss": 2.6001,
      "theoretical_loss": 3.347296216257459,
      "tokens_seen": 2707881984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018258585858585858,
      "loss": 2.3081,
      "theoretical_loss": 3.3472898572420116,
      "tokens_seen": 2707947520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018256565656565656,
      "loss": 2.629,
      "theoretical_loss": 3.347283498423549,
      "tokens_seen": 2708013056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018254545454545455,
      "loss": 2.5801,
      "theoretical_loss": 3.3472771398020607,
      "tokens_seen": 2708078592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018252525252525253,
      "loss": 2.4218,
      "theoretical_loss": 3.347270781377535,
      "tokens_seen": 2708144128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018250505050505052,
      "loss": 2.4671,
      "theoretical_loss": 3.347264423149962,
      "tokens_seen": 2708209664
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018248484848484847,
      "loss": 2.2512,
      "theoretical_loss": 3.34725806511933,
      "tokens_seen": 2708275200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018246464646464648,
      "loss": 2.5189,
      "theoretical_loss": 3.3472517072856287,
      "tokens_seen": 2708340736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018244444444444444,
      "loss": 2.6719,
      "theoretical_loss": 3.347245349648847,
      "tokens_seen": 2708406272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018242424242424242,
      "loss": 2.7243,
      "theoretical_loss": 3.347238992208974,
      "tokens_seen": 2708471808
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1485001,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9092819690704346,
      "objective/train/theoretical_loss": 3.347232634965999,
      "objective/train/tokens_used": 1067396576,
      "theoretical_loss": 3.347232634965999,
      "tokens_seen": 2708537344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001824040404040404,
      "loss": 2.3197,
      "theoretical_loss": 3.347232634965999,
      "tokens_seen": 2708537344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018238383838383836,
      "loss": 2.4197,
      "theoretical_loss": 3.347226277919911,
      "tokens_seen": 2708602880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018236363636363638,
      "loss": 2.365,
      "theoretical_loss": 3.347219921070699,
      "tokens_seen": 2708668416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018234343434343433,
      "loss": 2.4228,
      "theoretical_loss": 3.347213564418353,
      "tokens_seen": 2708733952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018232323232323234,
      "loss": 2.4779,
      "theoretical_loss": 3.347207207962861,
      "tokens_seen": 2708799488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001823030303030303,
      "loss": 2.4212,
      "theoretical_loss": 3.3472008517042124,
      "tokens_seen": 2708865024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001822828282828283,
      "loss": 2.345,
      "theoretical_loss": 3.3471944956423973,
      "tokens_seen": 2708930560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018226262626262627,
      "loss": 2.3913,
      "theoretical_loss": 3.3471881397774035,
      "tokens_seen": 2708996096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018224242424242423,
      "loss": 2.4625,
      "theoretical_loss": 3.3471817841092206,
      "tokens_seen": 2709061632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018222222222222224,
      "loss": 2.5894,
      "theoretical_loss": 3.347175428637838,
      "tokens_seen": 2709127168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001822020202020202,
      "loss": 2.294,
      "theoretical_loss": 3.3471690733632453,
      "tokens_seen": 2709192704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001821818181818182,
      "loss": 2.4062,
      "theoretical_loss": 3.34716271828543,
      "tokens_seen": 2709258240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018216161616161616,
      "loss": 2.3877,
      "theoretical_loss": 3.3471563634043835,
      "tokens_seen": 2709323776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018214141414141415,
      "loss": 2.5089,
      "theoretical_loss": 3.347150008720093,
      "tokens_seen": 2709389312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018212121212121213,
      "loss": 2.3448,
      "theoretical_loss": 3.3471436542325486,
      "tokens_seen": 2709454848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001821010101010101,
      "loss": 2.5059,
      "theoretical_loss": 3.3471372999417395,
      "tokens_seen": 2709520384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001820808080808081,
      "loss": 2.5152,
      "theoretical_loss": 3.3471309458476544,
      "tokens_seen": 2709585920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018206060606060605,
      "loss": 2.2663,
      "theoretical_loss": 3.347124591950283,
      "tokens_seen": 2709651456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018204040404040404,
      "loss": 2.564,
      "theoretical_loss": 3.347118238249614,
      "tokens_seen": 2709716992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018202020202020202,
      "loss": 2.7046,
      "theoretical_loss": 3.3471118847456363,
      "tokens_seen": 2709782528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000182,
      "loss": 2.7923,
      "theoretical_loss": 3.3471055314383396,
      "tokens_seen": 2709848064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000181979797979798,
      "loss": 2.4299,
      "theoretical_loss": 3.347099178327713,
      "tokens_seen": 2709913600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018195959595959597,
      "loss": 2.4373,
      "theoretical_loss": 3.3470928254137453,
      "tokens_seen": 2709979136
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018193939393939393,
      "loss": 2.4307,
      "theoretical_loss": 3.3470864726964265,
      "tokens_seen": 2710044672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018191919191919194,
      "loss": 2.5111,
      "theoretical_loss": 3.3470801201757445,
      "tokens_seen": 2710110208
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1485587,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5603830814361572,
      "objective/train/theoretical_loss": 3.3470737678516898,
      "objective/train/tokens_used": 1069034976,
      "theoretical_loss": 3.3470737678516898,
      "tokens_seen": 2710175744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001818989898989899,
      "loss": 2.4378,
      "theoretical_loss": 3.3470737678516898,
      "tokens_seen": 2710175744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018187878787878788,
      "loss": 2.63,
      "theoretical_loss": 3.34706741572425,
      "tokens_seen": 2710241280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018185858585858587,
      "loss": 2.476,
      "theoretical_loss": 3.347061063793416,
      "tokens_seen": 2710306816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018183838383838382,
      "loss": 2.4691,
      "theoretical_loss": 3.3470547120591756,
      "tokens_seen": 2710372352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018181818181818183,
      "loss": 2.4109,
      "theoretical_loss": 3.3470483605215184,
      "tokens_seen": 2710437888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001817979797979798,
      "loss": 2.3391,
      "theoretical_loss": 3.3470420091804343,
      "tokens_seen": 2710503424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001817777777777778,
      "loss": 2.4988,
      "theoretical_loss": 3.347035658035911,
      "tokens_seen": 2710568960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018175757575757576,
      "loss": 2.249,
      "theoretical_loss": 3.3470293070879387,
      "tokens_seen": 2710634496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018173737373737372,
      "loss": 2.5243,
      "theoretical_loss": 3.3470229563365064,
      "tokens_seen": 2710700032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018171717171717173,
      "loss": 2.6344,
      "theoretical_loss": 3.347016605781603,
      "tokens_seen": 2710765568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018169696969696968,
      "loss": 2.609,
      "theoretical_loss": 3.347010255423218,
      "tokens_seen": 2710831104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001816767676767677,
      "loss": 2.1508,
      "theoretical_loss": 3.3470039052613405,
      "tokens_seen": 2710896640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018165656565656565,
      "loss": 2.4535,
      "theoretical_loss": 3.3469975552959594,
      "tokens_seen": 2710962176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018163636363636364,
      "loss": 2.5204,
      "theoretical_loss": 3.3469912055270647,
      "tokens_seen": 2711027712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018161616161616162,
      "loss": 2.4313,
      "theoretical_loss": 3.346984855954644,
      "tokens_seen": 2711093248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001815959595959596,
      "loss": 2.4751,
      "theoretical_loss": 3.3469785065786883,
      "tokens_seen": 2711158784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001815757575757576,
      "loss": 2.3699,
      "theoretical_loss": 3.3469721573991853,
      "tokens_seen": 2711224320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018155555555555554,
      "loss": 2.552,
      "theoretical_loss": 3.3469658084161247,
      "tokens_seen": 2711289856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018153535353535353,
      "loss": 2.4496,
      "theoretical_loss": 3.346959459629496,
      "tokens_seen": 2711355392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001815151515151515,
      "loss": 2.353,
      "theoretical_loss": 3.346953111039288,
      "tokens_seen": 2711420928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001814949494949495,
      "loss": 2.4166,
      "theoretical_loss": 3.34694676264549,
      "tokens_seen": 2711486464
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018147474747474748,
      "loss": 2.4535,
      "theoretical_loss": 3.346940414448091,
      "tokens_seen": 2711552000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018145454545454546,
      "loss": 2.381,
      "theoretical_loss": 3.3469340664470804,
      "tokens_seen": 2711617536
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018143434343434345,
      "loss": 2.3516,
      "theoretical_loss": 3.3469277186424473,
      "tokens_seen": 2711683072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018141414141414143,
      "loss": 2.6999,
      "theoretical_loss": 3.3469213710341807,
      "tokens_seen": 2711748608
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1487028,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.833167552947998,
      "objective/train/theoretical_loss": 3.3469150236222704,
      "objective/train/tokens_used": 1070673376,
      "theoretical_loss": 3.3469150236222704,
      "tokens_seen": 2711814144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001813939393939394,
      "loss": 2.5415,
      "theoretical_loss": 3.3469150236222704,
      "tokens_seen": 2711814144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018137373737373737,
      "loss": 2.4226,
      "theoretical_loss": 3.346908676406705,
      "tokens_seen": 2711879680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018135353535353536,
      "loss": 2.5692,
      "theoretical_loss": 3.3469023293874733,
      "tokens_seen": 2711945216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018133333333333334,
      "loss": 2.3508,
      "theoretical_loss": 3.3468959825645657,
      "tokens_seen": 2712010752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018131313131313132,
      "loss": 2.4882,
      "theoretical_loss": 3.3468896359379703,
      "tokens_seen": 2712076288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018129292929292928,
      "loss": 2.4915,
      "theoretical_loss": 3.3468832895076766,
      "tokens_seen": 2712141824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001812727272727273,
      "loss": 2.7021,
      "theoretical_loss": 3.3468769432736742,
      "tokens_seen": 2712207360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018125252525252525,
      "loss": 2.2858,
      "theoretical_loss": 3.346870597235952,
      "tokens_seen": 2712272896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018123232323232326,
      "loss": 2.5339,
      "theoretical_loss": 3.346864251394499,
      "tokens_seen": 2712338432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018121212121212122,
      "loss": 2.5822,
      "theoretical_loss": 3.3468579057493044,
      "tokens_seen": 2712403968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018119191919191917,
      "loss": 2.5105,
      "theoretical_loss": 3.3468515603003572,
      "tokens_seen": 2712469504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018117171717171719,
      "loss": 2.4078,
      "theoretical_loss": 3.3468452150476478,
      "tokens_seen": 2712535040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018115151515151514,
      "loss": 2.2959,
      "theoretical_loss": 3.3468388699911635,
      "tokens_seen": 2712600576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018113131313131315,
      "loss": 2.3479,
      "theoretical_loss": 3.3468325251308952,
      "tokens_seen": 2712666112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001811111111111111,
      "loss": 2.4701,
      "theoretical_loss": 3.3468261804668313,
      "tokens_seen": 2712731648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001810909090909091,
      "loss": 2.3747,
      "theoretical_loss": 3.3468198359989607,
      "tokens_seen": 2712797184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018107070707070708,
      "loss": 2.5224,
      "theoretical_loss": 3.346813491727273,
      "tokens_seen": 2712862720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018105050505050506,
      "loss": 2.525,
      "theoretical_loss": 3.3468071476517576,
      "tokens_seen": 2712928256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018103030303030305,
      "loss": 2.2141,
      "theoretical_loss": 3.3468008037724033,
      "tokens_seen": 2712993792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000181010101010101,
      "loss": 2.4438,
      "theoretical_loss": 3.3467944600891997,
      "tokens_seen": 2713059328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000180989898989899,
      "loss": 2.4823,
      "theoretical_loss": 3.3467881166021356,
      "tokens_seen": 2713124864
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018096969696969697,
      "loss": 2.6494,
      "theoretical_loss": 3.3467817733112,
      "tokens_seen": 2713190400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018094949494949495,
      "loss": 2.4183,
      "theoretical_loss": 3.346775430216383,
      "tokens_seen": 2713255936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018092929292929294,
      "loss": 2.6091,
      "theoretical_loss": 3.346769087317673,
      "tokens_seen": 2713321472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018090909090909092,
      "loss": 2.4667,
      "theoretical_loss": 3.3467627446150594,
      "tokens_seen": 2713387008
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1487700,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3146984577178955,
      "objective/train/theoretical_loss": 3.3467564021085314,
      "objective/train/tokens_used": 1072311776,
      "theoretical_loss": 3.3467564021085314,
      "tokens_seen": 2713452544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018088888888888888,
      "loss": 2.2269,
      "theoretical_loss": 3.3467564021085314,
      "tokens_seen": 2713452544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018086868686868686,
      "loss": 2.5734,
      "theoretical_loss": 3.3467500597980786,
      "tokens_seen": 2713518080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018084848484848485,
      "loss": 2.3583,
      "theoretical_loss": 3.3467437176836894,
      "tokens_seen": 2713583616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018082828282828283,
      "loss": 2.3448,
      "theoretical_loss": 3.3467373757653536,
      "tokens_seen": 2713649152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018080808080808082,
      "loss": 2.7295,
      "theoretical_loss": 3.34673103404306,
      "tokens_seen": 2713714688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018078787878787877,
      "loss": 2.4642,
      "theoretical_loss": 3.3467246925167986,
      "tokens_seen": 2713780224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018076767676767678,
      "loss": 2.3671,
      "theoretical_loss": 3.3467183511865577,
      "tokens_seen": 2713845760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018074747474747474,
      "loss": 2.5793,
      "theoretical_loss": 3.346712010052327,
      "tokens_seen": 2713911296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018072727272727275,
      "loss": 2.3826,
      "theoretical_loss": 3.3467056691140957,
      "tokens_seen": 2713976832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001807070707070707,
      "loss": 2.5504,
      "theoretical_loss": 3.3466993283718525,
      "tokens_seen": 2714042368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018068686868686867,
      "loss": 2.4045,
      "theoretical_loss": 3.3466929878255876,
      "tokens_seen": 2714107904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018066666666666668,
      "loss": 2.4039,
      "theoretical_loss": 3.346686647475289,
      "tokens_seen": 2714173440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018064646464646463,
      "loss": 2.124,
      "theoretical_loss": 3.3466803073209466,
      "tokens_seen": 2714238976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018062626262626264,
      "loss": 2.5203,
      "theoretical_loss": 3.34667396736255,
      "tokens_seen": 2714304512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001806060606060606,
      "loss": 2.3105,
      "theoretical_loss": 3.3466676276000875,
      "tokens_seen": 2714370048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001805858585858586,
      "loss": 2.46,
      "theoretical_loss": 3.346661288033549,
      "tokens_seen": 2714435584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018056565656565657,
      "loss": 2.4667,
      "theoretical_loss": 3.3466549486629233,
      "tokens_seen": 2714501120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018054545454545455,
      "loss": 2.2639,
      "theoretical_loss": 3.3466486094882,
      "tokens_seen": 2714566656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018052525252525254,
      "loss": 2.6771,
      "theoretical_loss": 3.346642270509368,
      "tokens_seen": 2714632192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001805050505050505,
      "loss": 2.5091,
      "theoretical_loss": 3.346635931726417,
      "tokens_seen": 2714697728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001804848484848485,
      "loss": 2.476,
      "theoretical_loss": 3.346629593139335,
      "tokens_seen": 2714763264
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018046464646464646,
      "loss": 2.6063,
      "theoretical_loss": 3.346623254748113,
      "tokens_seen": 2714828800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018044444444444445,
      "loss": 2.4468,
      "theoretical_loss": 3.3466169165527386,
      "tokens_seen": 2714894336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018042424242424243,
      "loss": 2.3416,
      "theoretical_loss": 3.3466105785532023,
      "tokens_seen": 2714959872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001804040404040404,
      "loss": 2.6436,
      "theoretical_loss": 3.3466042407494925,
      "tokens_seen": 2715025408
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1488584,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8602285385131836,
      "objective/train/theoretical_loss": 3.3465979031415984,
      "objective/train/tokens_used": 1073950176,
      "theoretical_loss": 3.3465979031415984,
      "tokens_seen": 2715090944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001803838383838384,
      "loss": 2.4297,
      "theoretical_loss": 3.3465979031415984,
      "tokens_seen": 2715090944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018036363636363638,
      "loss": 2.3063,
      "theoretical_loss": 3.3465915657295096,
      "tokens_seen": 2715156480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018034343434343434,
      "loss": 2.7623,
      "theoretical_loss": 3.346585228513215,
      "tokens_seen": 2715222016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018032323232323232,
      "loss": 2.4433,
      "theoretical_loss": 3.3465788914927046,
      "tokens_seen": 2715287552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001803030303030303,
      "loss": 2.5431,
      "theoretical_loss": 3.3465725546679668,
      "tokens_seen": 2715353088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001802828282828283,
      "loss": 2.5037,
      "theoretical_loss": 3.346566218038991,
      "tokens_seen": 2715418624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018026262626262627,
      "loss": 2.7025,
      "theoretical_loss": 3.3465598816057667,
      "tokens_seen": 2715484160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018024242424242423,
      "loss": 2.7195,
      "theoretical_loss": 3.3465535453682826,
      "tokens_seen": 2715549696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018022222222222224,
      "loss": 2.5807,
      "theoretical_loss": 3.3465472093265287,
      "tokens_seen": 2715615232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001802020202020202,
      "loss": 2.432,
      "theoretical_loss": 3.3465408734804933,
      "tokens_seen": 2715680768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001801818181818182,
      "loss": 2.3513,
      "theoretical_loss": 3.3465345378301663,
      "tokens_seen": 2715746304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018016161616161617,
      "loss": 2.5359,
      "theoretical_loss": 3.346528202375537,
      "tokens_seen": 2715811840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018014141414141412,
      "loss": 2.6096,
      "theoretical_loss": 3.346521867116594,
      "tokens_seen": 2715877376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018012121212121213,
      "loss": 2.604,
      "theoretical_loss": 3.3465155320533273,
      "tokens_seen": 2715942912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001801010101010101,
      "loss": 2.6083,
      "theoretical_loss": 3.3465091971857257,
      "tokens_seen": 2716008448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001800808080808081,
      "loss": 2.1912,
      "theoretical_loss": 3.3465028625137787,
      "tokens_seen": 2716073984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018006060606060606,
      "loss": 2.4818,
      "theoretical_loss": 3.3464965280374748,
      "tokens_seen": 2716139520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018004040404040404,
      "loss": 2.2881,
      "theoretical_loss": 3.346490193756804,
      "tokens_seen": 2716205056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018002020202020203,
      "loss": 2.465,
      "theoretical_loss": 3.3464838596717557,
      "tokens_seen": 2716270592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017999999999999998,
      "loss": 2.6133,
      "theoretical_loss": 3.3464775257823183,
      "tokens_seen": 2716336128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000179979797979798,
      "loss": 2.3731,
      "theoretical_loss": 3.3464711920884813,
      "tokens_seen": 2716401664
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017995959595959595,
      "loss": 2.4465,
      "theoretical_loss": 3.3464648585902346,
      "tokens_seen": 2716467200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017993939393939394,
      "loss": 2.565,
      "theoretical_loss": 3.346458525287567,
      "tokens_seen": 2716532736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017991919191919192,
      "loss": 2.552,
      "theoretical_loss": 3.3464521921804673,
      "tokens_seen": 2716598272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001798989898989899,
      "loss": 2.5929,
      "theoretical_loss": 3.3464458592689255,
      "tokens_seen": 2716663808
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1489419,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.378549337387085,
      "objective/train/theoretical_loss": 3.3464395265529303,
      "objective/train/tokens_used": 1075588576,
      "theoretical_loss": 3.3464395265529303,
      "tokens_seen": 2716729344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001798787878787879,
      "loss": 2.3665,
      "theoretical_loss": 3.3464395265529303,
      "tokens_seen": 2716729344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017985858585858587,
      "loss": 2.5537,
      "theoretical_loss": 3.346433194032471,
      "tokens_seen": 2716794880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017983838383838383,
      "loss": 2.4699,
      "theoretical_loss": 3.3464268617075374,
      "tokens_seen": 2716860416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001798181818181818,
      "loss": 2.5484,
      "theoretical_loss": 3.3464205295781184,
      "tokens_seen": 2716925952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001797979797979798,
      "loss": 2.5882,
      "theoretical_loss": 3.346414197644203,
      "tokens_seen": 2716991488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017977777777777778,
      "loss": 2.4017,
      "theoretical_loss": 3.3464078659057805,
      "tokens_seen": 2717057024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017975757575757576,
      "loss": 2.5802,
      "theoretical_loss": 3.3464015343628404,
      "tokens_seen": 2717122560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017973737373737375,
      "loss": 2.4962,
      "theoretical_loss": 3.346395203015372,
      "tokens_seen": 2717188096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017971717171717173,
      "loss": 2.3226,
      "theoretical_loss": 3.346388871863364,
      "tokens_seen": 2717253632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001796969696969697,
      "loss": 2.4841,
      "theoretical_loss": 3.346382540906806,
      "tokens_seen": 2717319168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001796767676767677,
      "loss": 2.3176,
      "theoretical_loss": 3.3463762101456878,
      "tokens_seen": 2717384704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017965656565656566,
      "loss": 2.6631,
      "theoretical_loss": 3.346369879579998,
      "tokens_seen": 2717450240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017963636363636364,
      "loss": 2.4856,
      "theoretical_loss": 3.346363549209726,
      "tokens_seen": 2717515776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017961616161616163,
      "loss": 2.5622,
      "theoretical_loss": 3.3463572190348607,
      "tokens_seen": 2717581312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017959595959595958,
      "loss": 2.4627,
      "theoretical_loss": 3.346350889055392,
      "tokens_seen": 2717646848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001795757575757576,
      "loss": 2.2519,
      "theoretical_loss": 3.3463445592713086,
      "tokens_seen": 2717712384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017955555555555555,
      "loss": 2.6045,
      "theoretical_loss": 3.3463382296826003,
      "tokens_seen": 2717777920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017953535353535356,
      "loss": 2.4538,
      "theoretical_loss": 3.3463319002892558,
      "tokens_seen": 2717843456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017951515151515152,
      "loss": 2.4552,
      "theoretical_loss": 3.346325571091265,
      "tokens_seen": 2717908992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001794949494949495,
      "loss": 2.4055,
      "theoretical_loss": 3.346319242088617,
      "tokens_seen": 2717974528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017947474747474749,
      "loss": 2.5917,
      "theoretical_loss": 3.3463129132813005,
      "tokens_seen": 2718040064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017945454545454544,
      "loss": 2.6042,
      "theoretical_loss": 3.3463065846693048,
      "tokens_seen": 2718105600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017943434343434345,
      "loss": 2.4016,
      "theoretical_loss": 3.34630025625262,
      "tokens_seen": 2718171136
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001794141414141414,
      "loss": 2.4873,
      "theoretical_loss": 3.3462939280312347,
      "tokens_seen": 2718236672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001793939393939394,
      "loss": 2.6103,
      "theoretical_loss": 3.3462876000051383,
      "tokens_seen": 2718302208
    },
    {
      "epoch": 0.64,
      "objective/train/docs_used": 1489769,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7963850498199463,
      "objective/train/theoretical_loss": 3.34628127217432,
      "objective/train/tokens_used": 1077226976,
      "theoretical_loss": 3.34628127217432,
      "tokens_seen": 2718367744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017937373737373738,
      "loss": 2.6305,
      "theoretical_loss": 3.34628127217432,
      "tokens_seen": 2718367744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017935353535353536,
      "loss": 2.2062,
      "theoretical_loss": 3.346274944538769,
      "tokens_seen": 2718433280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017933333333333335,
      "loss": 2.4493,
      "theoretical_loss": 3.3462686170984752,
      "tokens_seen": 2718498816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001793131313131313,
      "loss": 2.7743,
      "theoretical_loss": 3.346262289853427,
      "tokens_seen": 2718564352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001792929292929293,
      "loss": 2.5875,
      "theoretical_loss": 3.3462559628036144,
      "tokens_seen": 2718629888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017927272727272727,
      "loss": 2.4126,
      "theoretical_loss": 3.346249635949026,
      "tokens_seen": 2718695424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017925252525252526,
      "loss": 2.6688,
      "theoretical_loss": 3.3462433092896515,
      "tokens_seen": 2718760960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017923232323232324,
      "loss": 2.4169,
      "theoretical_loss": 3.34623698282548,
      "tokens_seen": 2718826496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017921212121212122,
      "loss": 2.4834,
      "theoretical_loss": 3.346230656556501,
      "tokens_seen": 2718892032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017919191919191918,
      "loss": 2.5111,
      "theoretical_loss": 3.3462243304827033,
      "tokens_seen": 2718957568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001791717171717172,
      "loss": 2.4619,
      "theoretical_loss": 3.3462180046040766,
      "tokens_seen": 2719023104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017915151515151515,
      "loss": 2.564,
      "theoretical_loss": 3.34621167892061,
      "tokens_seen": 2719088640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017913131313131313,
      "loss": 2.3703,
      "theoretical_loss": 3.346205353432293,
      "tokens_seen": 2719154176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017911111111111112,
      "loss": 2.5059,
      "theoretical_loss": 3.346199028139115,
      "tokens_seen": 2719219712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017909090909090907,
      "loss": 2.624,
      "theoretical_loss": 3.3461927030410648,
      "tokens_seen": 2719285248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017907070707070708,
      "loss": 2.6148,
      "theoretical_loss": 3.346186378138132,
      "tokens_seen": 2719350784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017905050505050504,
      "loss": 2.3291,
      "theoretical_loss": 3.346180053430305,
      "tokens_seen": 2719416320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017903030303030305,
      "loss": 2.4597,
      "theoretical_loss": 3.3461737289175746,
      "tokens_seen": 2719481856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000179010101010101,
      "loss": 2.3228,
      "theoretical_loss": 3.3461674045999286,
      "tokens_seen": 2719547392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000178989898989899,
      "loss": 2.1143,
      "theoretical_loss": 3.3461610804773576,
      "tokens_seen": 2719612928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017896969696969698,
      "loss": 2.4988,
      "theoretical_loss": 3.34615475654985,
      "tokens_seen": 2719678464
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017894949494949493,
      "loss": 2.6052,
      "theoretical_loss": 3.3461484328173956,
      "tokens_seen": 2719744000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017892929292929294,
      "loss": 2.3632,
      "theoretical_loss": 3.346142109279983,
      "tokens_seen": 2719809536
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001789090909090909,
      "loss": 2.3987,
      "theoretical_loss": 3.3461357859376024,
      "tokens_seen": 2719875072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017888888888888889,
      "loss": 2.2826,
      "theoretical_loss": 3.346129462790242,
      "tokens_seen": 2719940608
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1491149,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4027974605560303,
      "objective/train/theoretical_loss": 3.346123139837892,
      "objective/train/tokens_used": 1078865376,
      "theoretical_loss": 3.346123139837892,
      "tokens_seen": 2720006144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017886868686868687,
      "loss": 2.3313,
      "theoretical_loss": 3.346123139837892,
      "tokens_seen": 2720006144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017884848484848485,
      "loss": 2.2946,
      "theoretical_loss": 3.3461168170805418,
      "tokens_seen": 2720071680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017882828282828284,
      "loss": 2.6664,
      "theoretical_loss": 3.3461104945181797,
      "tokens_seen": 2720137216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017880808080808082,
      "loss": 2.554,
      "theoretical_loss": 3.346104172150796,
      "tokens_seen": 2720202752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001787878787878788,
      "loss": 2.5731,
      "theoretical_loss": 3.346097849978379,
      "tokens_seen": 2720268288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017876767676767676,
      "loss": 2.5387,
      "theoretical_loss": 3.346091528000919,
      "tokens_seen": 2720333824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017874747474747475,
      "loss": 2.622,
      "theoretical_loss": 3.3460852062184046,
      "tokens_seen": 2720399360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017872727272727273,
      "loss": 2.3865,
      "theoretical_loss": 3.3460788846308254,
      "tokens_seen": 2720464896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017870707070707071,
      "loss": 2.6337,
      "theoretical_loss": 3.3460725632381703,
      "tokens_seen": 2720530432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001786868686868687,
      "loss": 2.4348,
      "theoretical_loss": 3.3460662420404295,
      "tokens_seen": 2720595968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017866666666666668,
      "loss": 2.2586,
      "theoretical_loss": 3.3460599210375914,
      "tokens_seen": 2720661504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017864646464646464,
      "loss": 2.5056,
      "theoretical_loss": 3.3460536002296455,
      "tokens_seen": 2720727040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017862626262626265,
      "loss": 2.3681,
      "theoretical_loss": 3.3460472796165814,
      "tokens_seen": 2720792576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001786060606060606,
      "loss": 2.4087,
      "theoretical_loss": 3.346040959198388,
      "tokens_seen": 2720858112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001785858585858586,
      "loss": 2.7711,
      "theoretical_loss": 3.346034638975055,
      "tokens_seen": 2720923648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017856565656565657,
      "loss": 2.673,
      "theoretical_loss": 3.3460283189465714,
      "tokens_seen": 2720989184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017854545454545453,
      "loss": 2.4926,
      "theoretical_loss": 3.3460219991129265,
      "tokens_seen": 2721054720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017852525252525254,
      "loss": 2.7564,
      "theoretical_loss": 3.34601567947411,
      "tokens_seen": 2721120256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001785050505050505,
      "loss": 2.3423,
      "theoretical_loss": 3.34600936003011,
      "tokens_seen": 2721185792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001784848484848485,
      "loss": 2.4657,
      "theoretical_loss": 3.346003040780918,
      "tokens_seen": 2721251328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017846464646464647,
      "loss": 2.4228,
      "theoretical_loss": 3.345996721726521,
      "tokens_seen": 2721316864
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017844444444444442,
      "loss": 2.4939,
      "theoretical_loss": 3.3459904028669096,
      "tokens_seen": 2721382400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017842424242424244,
      "loss": 2.5651,
      "theoretical_loss": 3.345984084202073,
      "tokens_seen": 2721447936
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001784040404040404,
      "loss": 2.4951,
      "theoretical_loss": 3.345977765732,
      "tokens_seen": 2721513472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001783838383838384,
      "loss": 2.448,
      "theoretical_loss": 3.3459714474566806,
      "tokens_seen": 2721579008
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1491715,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6255178451538086,
      "objective/train/theoretical_loss": 3.3459651293761032,
      "objective/train/tokens_used": 1080503776,
      "theoretical_loss": 3.3459651293761032,
      "tokens_seen": 2721644544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017836363636363636,
      "loss": 2.6509,
      "theoretical_loss": 3.3459651293761032,
      "tokens_seen": 2721644544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017834343434343434,
      "loss": 2.277,
      "theoretical_loss": 3.345958811490258,
      "tokens_seen": 2721710080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017832323232323233,
      "loss": 2.5304,
      "theoretical_loss": 3.345952493799134,
      "tokens_seen": 2721775616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001783030303030303,
      "loss": 2.4756,
      "theoretical_loss": 3.3459461763027205,
      "tokens_seen": 2721841152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001782828282828283,
      "loss": 2.5428,
      "theoretical_loss": 3.3459398590010068,
      "tokens_seen": 2721906688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017826262626262625,
      "loss": 2.3946,
      "theoretical_loss": 3.3459335418939817,
      "tokens_seen": 2721972224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017824242424242424,
      "loss": 2.4894,
      "theoretical_loss": 3.3459272249816356,
      "tokens_seen": 2722037760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017822222222222222,
      "loss": 2.4735,
      "theoretical_loss": 3.345920908263957,
      "tokens_seen": 2722103296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001782020202020202,
      "loss": 2.3455,
      "theoretical_loss": 3.345914591740935,
      "tokens_seen": 2722168832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001781818181818182,
      "loss": 2.5208,
      "theoretical_loss": 3.34590827541256,
      "tokens_seen": 2722234368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017816161616161617,
      "loss": 2.2801,
      "theoretical_loss": 3.3459019592788204,
      "tokens_seen": 2722299904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017814141414141413,
      "loss": 2.522,
      "theoretical_loss": 3.3458956433397056,
      "tokens_seen": 2722365440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017812121212121214,
      "loss": 2.457,
      "theoretical_loss": 3.3458893275952053,
      "tokens_seen": 2722430976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001781010101010101,
      "loss": 2.3007,
      "theoretical_loss": 3.3458830120453085,
      "tokens_seen": 2722496512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017808080808080808,
      "loss": 2.5132,
      "theoretical_loss": 3.3458766966900044,
      "tokens_seen": 2722562048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017806060606060607,
      "loss": 2.6731,
      "theoretical_loss": 3.345870381529283,
      "tokens_seen": 2722627584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017804040404040402,
      "loss": 2.4723,
      "theoretical_loss": 3.3458640665631325,
      "tokens_seen": 2722693120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017802020202020203,
      "loss": 2.6628,
      "theoretical_loss": 3.3458577517915438,
      "tokens_seen": 2722758656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000178,
      "loss": 2.4993,
      "theoretical_loss": 3.3458514372145047,
      "tokens_seen": 2722824192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000177979797979798,
      "loss": 2.3658,
      "theoretical_loss": 3.345845122832005,
      "tokens_seen": 2722889728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017795959595959596,
      "loss": 2.5685,
      "theoretical_loss": 3.3458388086440345,
      "tokens_seen": 2722955264
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017793939393939397,
      "loss": 2.3321,
      "theoretical_loss": 3.3458324946505824,
      "tokens_seen": 2723020800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017791919191919193,
      "loss": 2.6086,
      "theoretical_loss": 3.3458261808516374,
      "tokens_seen": 2723086336
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017789898989898988,
      "loss": 2.3901,
      "theoretical_loss": 3.3458198672471893,
      "tokens_seen": 2723151872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001778787878787879,
      "loss": 2.4532,
      "theoretical_loss": 3.3458135538372273,
      "tokens_seen": 2723217408
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1492888,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6339731216430664,
      "objective/train/theoretical_loss": 3.345807240621741,
      "objective/train/tokens_used": 1082142176,
      "theoretical_loss": 3.345807240621741,
      "tokens_seen": 2723282944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017785858585858585,
      "loss": 2.4715,
      "theoretical_loss": 3.345807240621741,
      "tokens_seen": 2723282944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017783838383838386,
      "loss": 2.3342,
      "theoretical_loss": 3.3458009276007195,
      "tokens_seen": 2723348480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017781818181818182,
      "loss": 2.4575,
      "theoretical_loss": 3.345794614774152,
      "tokens_seen": 2723414016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001777979797979798,
      "loss": 2.5119,
      "theoretical_loss": 3.3457883021420285,
      "tokens_seen": 2723479552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017777777777777779,
      "loss": 2.4197,
      "theoretical_loss": 3.3457819897043373,
      "tokens_seen": 2723545088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017775757575757577,
      "loss": 2.4893,
      "theoretical_loss": 3.3457756774610683,
      "tokens_seen": 2723610624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017773737373737375,
      "loss": 2.543,
      "theoretical_loss": 3.345769365412211,
      "tokens_seen": 2723676160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001777171717171717,
      "loss": 2.3916,
      "theoretical_loss": 3.3457630535577545,
      "tokens_seen": 2723741696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001776969696969697,
      "loss": 2.3492,
      "theoretical_loss": 3.345756741897688,
      "tokens_seen": 2723807232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017767676767676768,
      "loss": 2.5064,
      "theoretical_loss": 3.345750430432001,
      "tokens_seen": 2723872768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017765656565656566,
      "loss": 2.447,
      "theoretical_loss": 3.345744119160683,
      "tokens_seen": 2723938304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017763636363636365,
      "loss": 2.498,
      "theoretical_loss": 3.345737808083723,
      "tokens_seen": 2724003840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017761616161616163,
      "loss": 2.0675,
      "theoretical_loss": 3.3457314972011103,
      "tokens_seen": 2724069376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001775959595959596,
      "loss": 2.2933,
      "theoretical_loss": 3.345725186512835,
      "tokens_seen": 2724134912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017757575757575757,
      "loss": 2.4245,
      "theoretical_loss": 3.3457188760188856,
      "tokens_seen": 2724200448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017755555555555556,
      "loss": 2.4136,
      "theoretical_loss": 3.3457125657192512,
      "tokens_seen": 2724265984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017753535353535354,
      "loss": 2.3768,
      "theoretical_loss": 3.3457062556139223,
      "tokens_seen": 2724331520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017751515151515152,
      "loss": 2.3923,
      "theoretical_loss": 3.3456999457028878,
      "tokens_seen": 2724397056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017749494949494948,
      "loss": 2.5208,
      "theoretical_loss": 3.3456936359861364,
      "tokens_seen": 2724462592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001774747474747475,
      "loss": 2.4495,
      "theoretical_loss": 3.345687326463658,
      "tokens_seen": 2724528128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017745454545454545,
      "loss": 2.4585,
      "theoretical_loss": 3.345681017135442,
      "tokens_seen": 2724593664
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017743434343434346,
      "loss": 2.5762,
      "theoretical_loss": 3.3456747080014773,
      "tokens_seen": 2724659200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017741414141414142,
      "loss": 2.4656,
      "theoretical_loss": 3.3456683990617537,
      "tokens_seen": 2724724736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017739393939393937,
      "loss": 2.2935,
      "theoretical_loss": 3.3456620903162606,
      "tokens_seen": 2724790272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017737373737373738,
      "loss": 2.4084,
      "theoretical_loss": 3.345655781764987,
      "tokens_seen": 2724855808
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1493539,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.459871768951416,
      "objective/train/theoretical_loss": 3.3456494734079225,
      "objective/train/tokens_used": 1083780576,
      "theoretical_loss": 3.3456494734079225,
      "tokens_seen": 2724921344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017735353535353534,
      "loss": 2.5583,
      "theoretical_loss": 3.3456494734079225,
      "tokens_seen": 2724921344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017733333333333335,
      "loss": 2.4765,
      "theoretical_loss": 3.345643165245056,
      "tokens_seen": 2724986880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001773131313131313,
      "loss": 2.3926,
      "theoretical_loss": 3.3456368572763773,
      "tokens_seen": 2725052416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001772929292929293,
      "loss": 2.2519,
      "theoretical_loss": 3.345630549501876,
      "tokens_seen": 2725117952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017727272727272728,
      "loss": 2.6669,
      "theoretical_loss": 3.3456242419215405,
      "tokens_seen": 2725183488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017725252525252526,
      "loss": 2.6278,
      "theoretical_loss": 3.3456179345353614,
      "tokens_seen": 2725249024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017723232323232324,
      "loss": 2.4478,
      "theoretical_loss": 3.345611627343327,
      "tokens_seen": 2725314560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001772121212121212,
      "loss": 2.5039,
      "theoretical_loss": 3.345605320345427,
      "tokens_seen": 2725380096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017719191919191919,
      "loss": 2.6057,
      "theoretical_loss": 3.345599013541651,
      "tokens_seen": 2725445632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017717171717171717,
      "loss": 2.3337,
      "theoretical_loss": 3.345592706931988,
      "tokens_seen": 2725511168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017715151515151515,
      "loss": 2.4373,
      "theoretical_loss": 3.345586400516428,
      "tokens_seen": 2725576704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017713131313131314,
      "loss": 2.3604,
      "theoretical_loss": 3.3455800942949594,
      "tokens_seen": 2725642240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017711111111111112,
      "loss": 2.4601,
      "theoretical_loss": 3.345573788267572,
      "tokens_seen": 2725707776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001770909090909091,
      "loss": 2.578,
      "theoretical_loss": 3.3455674824342556,
      "tokens_seen": 2725773312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001770707070707071,
      "loss": 2.4867,
      "theoretical_loss": 3.345561176794999,
      "tokens_seen": 2725838848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017705050505050505,
      "loss": 2.3889,
      "theoretical_loss": 3.3455548713497913,
      "tokens_seen": 2725904384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017703030303030303,
      "loss": 2.4814,
      "theoretical_loss": 3.345548566098623,
      "tokens_seen": 2725969920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017701010101010101,
      "loss": 2.3653,
      "theoretical_loss": 3.3455422610414827,
      "tokens_seen": 2726035456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000176989898989899,
      "loss": 2.3246,
      "theoretical_loss": 3.3455359561783595,
      "tokens_seen": 2726100992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017696969696969698,
      "loss": 2.3211,
      "theoretical_loss": 3.345529651509243,
      "tokens_seen": 2726166528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017694949494949494,
      "loss": 2.3399,
      "theoretical_loss": 3.3455233470341232,
      "tokens_seen": 2726232064
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017692929292929295,
      "loss": 2.3903,
      "theoretical_loss": 3.3455170427529883,
      "tokens_seen": 2726297600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001769090909090909,
      "loss": 2.454,
      "theoretical_loss": 3.345510738665829,
      "tokens_seen": 2726363136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017688888888888892,
      "loss": 2.7146,
      "theoretical_loss": 3.3455044347726335,
      "tokens_seen": 2726428672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017686868686868687,
      "loss": 2.4177,
      "theoretical_loss": 3.345498131073392,
      "tokens_seen": 2726494208
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1494687,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.248422622680664,
      "objective/train/theoretical_loss": 3.3454918275680927,
      "objective/train/tokens_used": 1085418976,
      "theoretical_loss": 3.3454918275680927,
      "tokens_seen": 2726559744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017684848484848483,
      "loss": 2.6784,
      "theoretical_loss": 3.3454918275680927,
      "tokens_seen": 2726559744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017682828282828284,
      "loss": 2.5723,
      "theoretical_loss": 3.3454855242567265,
      "tokens_seen": 2726625280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001768080808080808,
      "loss": 2.6373,
      "theoretical_loss": 3.3454792211392816,
      "tokens_seen": 2726690816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001767878787878788,
      "loss": 2.255,
      "theoretical_loss": 3.3454729182157483,
      "tokens_seen": 2726756352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017676767676767677,
      "loss": 2.6882,
      "theoretical_loss": 3.3454666154861155,
      "tokens_seen": 2726821888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017674747474747475,
      "loss": 2.5089,
      "theoretical_loss": 3.345460312950372,
      "tokens_seen": 2726887424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017672727272727274,
      "loss": 2.4428,
      "theoretical_loss": 3.345454010608508,
      "tokens_seen": 2726952960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001767070707070707,
      "loss": 2.1321,
      "theoretical_loss": 3.345447708460513,
      "tokens_seen": 2727018496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001766868686868687,
      "loss": 2.2741,
      "theoretical_loss": 3.3454414065063753,
      "tokens_seen": 2727084032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017666666666666666,
      "loss": 2.5295,
      "theoretical_loss": 3.3454351047460857,
      "tokens_seen": 2727149568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017664646464646464,
      "loss": 2.3459,
      "theoretical_loss": 3.3454288031796326,
      "tokens_seen": 2727215104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017662626262626263,
      "loss": 2.5008,
      "theoretical_loss": 3.3454225018070054,
      "tokens_seen": 2727280640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001766060606060606,
      "loss": 2.5305,
      "theoretical_loss": 3.3454162006281942,
      "tokens_seen": 2727346176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001765858585858586,
      "loss": 2.5268,
      "theoretical_loss": 3.3454098996431876,
      "tokens_seen": 2727411712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017656565656565658,
      "loss": 2.5351,
      "theoretical_loss": 3.3454035988519752,
      "tokens_seen": 2727477248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017654545454545454,
      "loss": 2.5015,
      "theoretical_loss": 3.3453972982545466,
      "tokens_seen": 2727542784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017652525252525252,
      "loss": 2.3463,
      "theoretical_loss": 3.345390997850891,
      "tokens_seen": 2727608320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001765050505050505,
      "loss": 2.4823,
      "theoretical_loss": 3.345384697640998,
      "tokens_seen": 2727673856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001764848484848485,
      "loss": 2.5309,
      "theoretical_loss": 3.3453783976248563,
      "tokens_seen": 2727739392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017646464646464647,
      "loss": 2.7278,
      "theoretical_loss": 3.345372097802456,
      "tokens_seen": 2727804928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017644444444444443,
      "loss": 2.3179,
      "theoretical_loss": 3.3453657981737868,
      "tokens_seen": 2727870464
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017642424242424244,
      "loss": 2.6319,
      "theoretical_loss": 3.3453594987388375,
      "tokens_seen": 2727936000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001764040404040404,
      "loss": 2.4107,
      "theoretical_loss": 3.3453531994975974,
      "tokens_seen": 2728001536
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001763838383838384,
      "loss": 2.4977,
      "theoretical_loss": 3.345346900450056,
      "tokens_seen": 2728067072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017636363636363637,
      "loss": 2.484,
      "theoretical_loss": 3.345340601596203,
      "tokens_seen": 2728132608
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1495196,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6293399333953857,
      "objective/train/theoretical_loss": 3.3453343029360267,
      "objective/train/tokens_used": 1087057376,
      "theoretical_loss": 3.3453343029360267,
      "tokens_seen": 2728198144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017634343434343432,
      "loss": 2.6006,
      "theoretical_loss": 3.3453343029360267,
      "tokens_seen": 2728198144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017632323232323233,
      "loss": 2.2541,
      "theoretical_loss": 3.3453280044695184,
      "tokens_seen": 2728263680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001763030303030303,
      "loss": 2.6532,
      "theoretical_loss": 3.345321706196666,
      "tokens_seen": 2728329216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001762828282828283,
      "loss": 2.6359,
      "theoretical_loss": 3.345315408117459,
      "tokens_seen": 2728394752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017626262626262626,
      "loss": 2.3939,
      "theoretical_loss": 3.345309110231888,
      "tokens_seen": 2728460288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017624242424242427,
      "loss": 2.6302,
      "theoretical_loss": 3.3453028125399404,
      "tokens_seen": 2728525824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017622222222222223,
      "loss": 2.3937,
      "theoretical_loss": 3.345296515041608,
      "tokens_seen": 2728591360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001762020202020202,
      "loss": 2.4791,
      "theoretical_loss": 3.3452902177368777,
      "tokens_seen": 2728656896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001761818181818182,
      "loss": 2.292,
      "theoretical_loss": 3.3452839206257408,
      "tokens_seen": 2728722432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017616161616161615,
      "loss": 2.368,
      "theoretical_loss": 3.345277623708186,
      "tokens_seen": 2728787968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017614141414141416,
      "loss": 2.3361,
      "theoretical_loss": 3.3452713269842023,
      "tokens_seen": 2728853504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017612121212121212,
      "loss": 2.5059,
      "theoretical_loss": 3.34526503045378,
      "tokens_seen": 2728919040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001761010101010101,
      "loss": 2.3078,
      "theoretical_loss": 3.3452587341169076,
      "tokens_seen": 2728984576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001760808080808081,
      "loss": 2.3487,
      "theoretical_loss": 3.345252437973575,
      "tokens_seen": 2729050112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017606060606060607,
      "loss": 2.3791,
      "theoretical_loss": 3.345246142023772,
      "tokens_seen": 2729115648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017604040404040405,
      "loss": 2.5723,
      "theoretical_loss": 3.3452398462674866,
      "tokens_seen": 2729181184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000176020202020202,
      "loss": 2.5314,
      "theoretical_loss": 3.3452335507047097,
      "tokens_seen": 2729246720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000176,
      "loss": 2.5865,
      "theoretical_loss": 3.34522725533543,
      "tokens_seen": 2729312256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017597979797979798,
      "loss": 2.513,
      "theoretical_loss": 3.3452209601596374,
      "tokens_seen": 2729377792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017595959595959596,
      "loss": 2.5898,
      "theoretical_loss": 3.3452146651773207,
      "tokens_seen": 2729443328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017593939393939395,
      "loss": 2.2181,
      "theoretical_loss": 3.3452083703884696,
      "tokens_seen": 2729508864
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017591919191919193,
      "loss": 2.5111,
      "theoretical_loss": 3.345202075793073,
      "tokens_seen": 2729574400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001758989898989899,
      "loss": 2.294,
      "theoretical_loss": 3.3451957813911215,
      "tokens_seen": 2729639936
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001758787878787879,
      "loss": 2.513,
      "theoretical_loss": 3.3451894871826036,
      "tokens_seen": 2729705472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017585858585858586,
      "loss": 2.5602,
      "theoretical_loss": 3.3451831931675087,
      "tokens_seen": 2729771008
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1496443,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.345172882080078,
      "objective/train/theoretical_loss": 3.3451768993458266,
      "objective/train/tokens_used": 1088695776,
      "theoretical_loss": 3.3451768993458266,
      "tokens_seen": 2729836544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017583838383838384,
      "loss": 2.5411,
      "theoretical_loss": 3.3451768993458266,
      "tokens_seen": 2729836544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017581818181818182,
      "loss": 2.6169,
      "theoretical_loss": 3.3451706057175463,
      "tokens_seen": 2729902080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017579797979797978,
      "loss": 2.4187,
      "theoretical_loss": 3.3451643122826575,
      "tokens_seen": 2729967616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001757777777777778,
      "loss": 2.5986,
      "theoretical_loss": 3.3451580190411496,
      "tokens_seen": 2730033152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017575757575757575,
      "loss": 2.6854,
      "theoretical_loss": 3.345151725993012,
      "tokens_seen": 2730098688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017573737373737376,
      "loss": 2.6153,
      "theoretical_loss": 3.345145433138234,
      "tokens_seen": 2730164224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017571717171717172,
      "loss": 2.5143,
      "theoretical_loss": 3.345139140476805,
      "tokens_seen": 2730229760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001756969696969697,
      "loss": 2.7027,
      "theoretical_loss": 3.345132848008715,
      "tokens_seen": 2730295296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017567676767676768,
      "loss": 2.2668,
      "theoretical_loss": 3.3451265557339527,
      "tokens_seen": 2730360832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017565656565656564,
      "loss": 2.501,
      "theoretical_loss": 3.3451202636525075,
      "tokens_seen": 2730426368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017563636363636365,
      "loss": 2.5209,
      "theoretical_loss": 3.3451139717643694,
      "tokens_seen": 2730491904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001756161616161616,
      "loss": 2.5928,
      "theoretical_loss": 3.3451076800695274,
      "tokens_seen": 2730557440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001755959595959596,
      "loss": 2.276,
      "theoretical_loss": 3.3451013885679712,
      "tokens_seen": 2730622976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017557575757575758,
      "loss": 2.2878,
      "theoretical_loss": 3.34509509725969,
      "tokens_seen": 2730688512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017555555555555556,
      "loss": 2.6027,
      "theoretical_loss": 3.345088806144673,
      "tokens_seen": 2730754048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017553535353535355,
      "loss": 2.7153,
      "theoretical_loss": 3.34508251522291,
      "tokens_seen": 2730819584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017551515151515153,
      "loss": 2.4384,
      "theoretical_loss": 3.3450762244943903,
      "tokens_seen": 2730885120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017549494949494949,
      "loss": 2.5186,
      "theoretical_loss": 3.3450699339591035,
      "tokens_seen": 2730950656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017547474747474747,
      "loss": 2.3682,
      "theoretical_loss": 3.3450636436170385,
      "tokens_seen": 2731016192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017545454545454545,
      "loss": 2.5585,
      "theoretical_loss": 3.3450573534681856,
      "tokens_seen": 2731081728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017543434343434344,
      "loss": 2.3031,
      "theoretical_loss": 3.3450510635125337,
      "tokens_seen": 2731147264
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017541414141414142,
      "loss": 2.48,
      "theoretical_loss": 3.3450447737500717,
      "tokens_seen": 2731212800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001753939393939394,
      "loss": 2.3915,
      "theoretical_loss": 3.3450384841807903,
      "tokens_seen": 2731278336
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001753737373737374,
      "loss": 2.6075,
      "theoretical_loss": 3.345032194804678,
      "tokens_seen": 2731343872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017535353535353535,
      "loss": 2.3865,
      "theoretical_loss": 3.3450259056217244,
      "tokens_seen": 2731409408
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1497834,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.771178960800171,
      "objective/train/theoretical_loss": 3.345019616631919,
      "objective/train/tokens_used": 1090334176,
      "theoretical_loss": 3.345019616631919,
      "tokens_seen": 2731474944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017533333333333336,
      "loss": 2.5906,
      "theoretical_loss": 3.345019616631919,
      "tokens_seen": 2731474944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017531313131313131,
      "loss": 2.5288,
      "theoretical_loss": 3.345013327835251,
      "tokens_seen": 2731540480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001752929292929293,
      "loss": 2.4116,
      "theoretical_loss": 3.3450070392317106,
      "tokens_seen": 2731606016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017527272727272728,
      "loss": 2.5544,
      "theoretical_loss": 3.3450007508212862,
      "tokens_seen": 2731671552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017525252525252524,
      "loss": 2.3793,
      "theoretical_loss": 3.3449944626039683,
      "tokens_seen": 2731737088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017523232323232325,
      "loss": 2.6695,
      "theoretical_loss": 3.344988174579745,
      "tokens_seen": 2731802624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001752121212121212,
      "loss": 2.6443,
      "theoretical_loss": 3.344981886748607,
      "tokens_seen": 2731868160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017519191919191922,
      "loss": 2.3688,
      "theoretical_loss": 3.344975599110543,
      "tokens_seen": 2731933696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017517171717171718,
      "loss": 2.4677,
      "theoretical_loss": 3.344969311665543,
      "tokens_seen": 2731999232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017515151515151513,
      "loss": 2.7443,
      "theoretical_loss": 3.3449630244135964,
      "tokens_seen": 2732064768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017513131313131314,
      "loss": 2.4406,
      "theoretical_loss": 3.344956737354692,
      "tokens_seen": 2732130304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001751111111111111,
      "loss": 2.6151,
      "theoretical_loss": 3.3449504504888194,
      "tokens_seen": 2732195840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001750909090909091,
      "loss": 2.334,
      "theoretical_loss": 3.3449441638159687,
      "tokens_seen": 2732261376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017507070707070707,
      "loss": 2.4341,
      "theoretical_loss": 3.344937877336129,
      "tokens_seen": 2732326912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017505050505050505,
      "loss": 2.556,
      "theoretical_loss": 3.344931591049289,
      "tokens_seen": 2732392448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017503030303030304,
      "loss": 2.5427,
      "theoretical_loss": 3.344925304955439,
      "tokens_seen": 2732457984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017501010101010102,
      "loss": 2.5749,
      "theoretical_loss": 3.3449190190545686,
      "tokens_seen": 2732523520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000174989898989899,
      "loss": 2.4344,
      "theoretical_loss": 3.344912733346667,
      "tokens_seen": 2732589056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017496969696969696,
      "loss": 2.5658,
      "theoretical_loss": 3.3449064478317228,
      "tokens_seen": 2732654592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017494949494949494,
      "loss": 2.4196,
      "theoretical_loss": 3.3449001625097265,
      "tokens_seen": 2732720128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017492929292929293,
      "loss": 2.3603,
      "theoretical_loss": 3.3448938773806676,
      "tokens_seen": 2732785664
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001749090909090909,
      "loss": 2.5436,
      "theoretical_loss": 3.344887592444535,
      "tokens_seen": 2732851200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001748888888888889,
      "loss": 2.449,
      "theoretical_loss": 3.344881307701318,
      "tokens_seen": 2732916736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017486868686868688,
      "loss": 2.6443,
      "theoretical_loss": 3.344875023151007,
      "tokens_seen": 2732982272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017484848484848484,
      "loss": 2.5382,
      "theoretical_loss": 3.3448687387935903,
      "tokens_seen": 2733047808
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1498342,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.313222885131836,
      "objective/train/theoretical_loss": 3.344862454629058,
      "objective/train/tokens_used": 1091972576,
      "theoretical_loss": 3.344862454629058,
      "tokens_seen": 2733113344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017482828282828285,
      "loss": 2.5151,
      "theoretical_loss": 3.344862454629058,
      "tokens_seen": 2733113344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001748080808080808,
      "loss": 2.582,
      "theoretical_loss": 3.3448561706574,
      "tokens_seen": 2733178880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001747878787878788,
      "loss": 2.5593,
      "theoretical_loss": 3.3448498868786047,
      "tokens_seen": 2733244416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017476767676767677,
      "loss": 2.5863,
      "theoretical_loss": 3.344843603292662,
      "tokens_seen": 2733309952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017474747474747473,
      "loss": 2.4391,
      "theoretical_loss": 3.3448373198995616,
      "tokens_seen": 2733375488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017472727272727274,
      "loss": 2.5547,
      "theoretical_loss": 3.3448310366992926,
      "tokens_seen": 2733441024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001747070707070707,
      "loss": 2.3641,
      "theoretical_loss": 3.344824753691845,
      "tokens_seen": 2733506560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001746868686868687,
      "loss": 2.5537,
      "theoretical_loss": 3.3448184708772075,
      "tokens_seen": 2733572096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017466666666666667,
      "loss": 2.4433,
      "theoretical_loss": 3.3448121882553705,
      "tokens_seen": 2733637632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017464646464646465,
      "loss": 2.3546,
      "theoretical_loss": 3.3448059058263224,
      "tokens_seen": 2733703168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017462626262626263,
      "loss": 2.3936,
      "theoretical_loss": 3.3447996235900535,
      "tokens_seen": 2733768704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001746060606060606,
      "loss": 2.5517,
      "theoretical_loss": 3.344793341546553,
      "tokens_seen": 2733834240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001745858585858586,
      "loss": 2.4965,
      "theoretical_loss": 3.3447870596958102,
      "tokens_seen": 2733899776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017456565656565656,
      "loss": 2.7289,
      "theoretical_loss": 3.3447807780378147,
      "tokens_seen": 2733965312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017454545454545457,
      "loss": 2.343,
      "theoretical_loss": 3.3447744965725557,
      "tokens_seen": 2734030848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017452525252525253,
      "loss": 2.4557,
      "theoretical_loss": 3.3447682153000233,
      "tokens_seen": 2734096384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001745050505050505,
      "loss": 2.2667,
      "theoretical_loss": 3.344761934220206,
      "tokens_seen": 2734161920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001744848484848485,
      "loss": 2.4983,
      "theoretical_loss": 3.3447556533330944,
      "tokens_seen": 2734227456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017446464646464648,
      "loss": 2.581,
      "theoretical_loss": 3.3447493726386774,
      "tokens_seen": 2734292992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017444444444444446,
      "loss": 2.6461,
      "theoretical_loss": 3.3447430921369445,
      "tokens_seen": 2734358528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017442424242424242,
      "loss": 2.7583,
      "theoretical_loss": 3.344736811827885,
      "tokens_seen": 2734424064
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001744040404040404,
      "loss": 2.5833,
      "theoretical_loss": 3.3447305317114884,
      "tokens_seen": 2734489600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001743838383838384,
      "loss": 2.463,
      "theoretical_loss": 3.3447242517877447,
      "tokens_seen": 2734555136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017436363636363637,
      "loss": 2.6402,
      "theoretical_loss": 3.3447179720566425,
      "tokens_seen": 2734620672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017434343434343435,
      "loss": 2.646,
      "theoretical_loss": 3.344711692518172,
      "tokens_seen": 2734686208
    },
    {
      "epoch": 0.65,
      "objective/train/docs_used": 1499499,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.72967791557312,
      "objective/train/theoretical_loss": 3.3447054131723224,
      "objective/train/tokens_used": 1093610976,
      "theoretical_loss": 3.3447054131723224,
      "tokens_seen": 2734751744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017432323232323234,
      "loss": 2.5635,
      "theoretical_loss": 3.3447054131723224,
      "tokens_seen": 2734751744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001743030303030303,
      "loss": 2.3556,
      "theoretical_loss": 3.3446991340190833,
      "tokens_seen": 2734817280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017428282828282828,
      "loss": 2.4384,
      "theoretical_loss": 3.344692855058444,
      "tokens_seen": 2734882816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017426262626262626,
      "loss": 2.5683,
      "theoretical_loss": 3.344686576290394,
      "tokens_seen": 2734948352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017424242424242425,
      "loss": 2.5014,
      "theoretical_loss": 3.344680297714923,
      "tokens_seen": 2735013888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017422222222222223,
      "loss": 2.3371,
      "theoretical_loss": 3.34467401933202,
      "tokens_seen": 2735079424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001742020202020202,
      "loss": 2.5409,
      "theoretical_loss": 3.344667741141675,
      "tokens_seen": 2735144960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001741818181818182,
      "loss": 2.3913,
      "theoretical_loss": 3.3446614631438774,
      "tokens_seen": 2735210496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017416161616161616,
      "loss": 2.5409,
      "theoretical_loss": 3.3446551853386164,
      "tokens_seen": 2735276032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017414141414141417,
      "loss": 2.3571,
      "theoretical_loss": 3.3446489077258814,
      "tokens_seen": 2735341568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017412121212121212,
      "loss": 2.3962,
      "theoretical_loss": 3.3446426303056622,
      "tokens_seen": 2735407104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017410101010101008,
      "loss": 2.4922,
      "theoretical_loss": 3.3446363530779486,
      "tokens_seen": 2735472640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001740808080808081,
      "loss": 2.5545,
      "theoretical_loss": 3.3446300760427294,
      "tokens_seen": 2735538176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017406060606060605,
      "loss": 2.3725,
      "theoretical_loss": 3.3446237991999945,
      "tokens_seen": 2735603712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017404040404040406,
      "loss": 2.4952,
      "theoretical_loss": 3.344617522549733,
      "tokens_seen": 2735669248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017402020202020202,
      "loss": 2.3943,
      "theoretical_loss": 3.344611246091935,
      "tokens_seen": 2735734784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000174,
      "loss": 2.4569,
      "theoretical_loss": 3.3446049698265896,
      "tokens_seen": 2735800320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017397979797979798,
      "loss": 2.66,
      "theoretical_loss": 3.344598693753686,
      "tokens_seen": 2735865856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017395959595959597,
      "loss": 2.4489,
      "theoretical_loss": 3.3445924178732143,
      "tokens_seen": 2735931392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017393939393939395,
      "loss": 2.5073,
      "theoretical_loss": 3.3445861421851637,
      "tokens_seen": 2735996928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001739191919191919,
      "loss": 2.286,
      "theoretical_loss": 3.3445798666895237,
      "tokens_seen": 2736062464
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001738989898989899,
      "loss": 2.637,
      "theoretical_loss": 3.3445735913862835,
      "tokens_seen": 2736128000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017387878787878788,
      "loss": 2.4915,
      "theoretical_loss": 3.344567316275433,
      "tokens_seen": 2736193536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017385858585858586,
      "loss": 2.3633,
      "theoretical_loss": 3.344561041356962,
      "tokens_seen": 2736259072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017383838383838385,
      "loss": 2.3622,
      "theoretical_loss": 3.344554766630859,
      "tokens_seen": 2736324608
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1499930,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.625838041305542,
      "objective/train/theoretical_loss": 3.3445484920971142,
      "objective/train/tokens_used": 1095249376,
      "theoretical_loss": 3.3445484920971142,
      "tokens_seen": 2736390144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017381818181818183,
      "loss": 2.6424,
      "theoretical_loss": 3.3445484920971142,
      "tokens_seen": 2736390144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017379797979797979,
      "loss": 2.6782,
      "theoretical_loss": 3.344542217755717,
      "tokens_seen": 2736455680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001737777777777778,
      "loss": 2.5903,
      "theoretical_loss": 3.344535943606657,
      "tokens_seen": 2736521216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017375757575757575,
      "loss": 2.3787,
      "theoretical_loss": 3.3445296696499236,
      "tokens_seen": 2736586752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017373737373737374,
      "loss": 2.501,
      "theoretical_loss": 3.344523395885506,
      "tokens_seen": 2736652288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017371717171717172,
      "loss": 2.377,
      "theoretical_loss": 3.344517122313394,
      "tokens_seen": 2736717824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001736969696969697,
      "loss": 2.3833,
      "theoretical_loss": 3.3445108489335773,
      "tokens_seen": 2736783360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001736767676767677,
      "loss": 2.6227,
      "theoretical_loss": 3.3445045757460448,
      "tokens_seen": 2736848896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017365656565656565,
      "loss": 2.5532,
      "theoretical_loss": 3.3444983027507864,
      "tokens_seen": 2736914432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017363636363636366,
      "loss": 2.3314,
      "theoretical_loss": 3.3444920299477916,
      "tokens_seen": 2736979968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017361616161616161,
      "loss": 2.7198,
      "theoretical_loss": 3.34448575733705,
      "tokens_seen": 2737045504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017359595959595963,
      "loss": 2.587,
      "theoretical_loss": 3.3444794849185513,
      "tokens_seen": 2737111040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017357575757575758,
      "loss": 2.4204,
      "theoretical_loss": 3.344473212692284,
      "tokens_seen": 2737176576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017355555555555554,
      "loss": 2.3673,
      "theoretical_loss": 3.344466940658239,
      "tokens_seen": 2737242112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017353535353535355,
      "loss": 2.6659,
      "theoretical_loss": 3.3444606688164047,
      "tokens_seen": 2737307648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001735151515151515,
      "loss": 2.4983,
      "theoretical_loss": 3.344454397166771,
      "tokens_seen": 2737373184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017349494949494952,
      "loss": 2.3814,
      "theoretical_loss": 3.3444481257093273,
      "tokens_seen": 2737438720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017347474747474748,
      "loss": 2.6721,
      "theoretical_loss": 3.3444418544440633,
      "tokens_seen": 2737504256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017345454545454546,
      "loss": 2.4244,
      "theoretical_loss": 3.3444355833709682,
      "tokens_seen": 2737569792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017343434343434344,
      "loss": 2.5724,
      "theoretical_loss": 3.344429312490032,
      "tokens_seen": 2737635328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001734141414141414,
      "loss": 2.68,
      "theoretical_loss": 3.344423041801244,
      "tokens_seen": 2737700864
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001733939393939394,
      "loss": 2.5312,
      "theoretical_loss": 3.3444167713045934,
      "tokens_seen": 2737766400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017337373737373737,
      "loss": 2.45,
      "theoretical_loss": 3.34441050100007,
      "tokens_seen": 2737831936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017335353535353535,
      "loss": 2.4462,
      "theoretical_loss": 3.3444042308876636,
      "tokens_seen": 2737897472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017333333333333334,
      "loss": 2.4034,
      "theoretical_loss": 3.3443979609673633,
      "tokens_seen": 2737963008
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1501408,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4893600940704346,
      "objective/train/theoretical_loss": 3.3443916912391587,
      "objective/train/tokens_used": 1096887776,
      "theoretical_loss": 3.3443916912391587,
      "tokens_seen": 2738028544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017331313131313132,
      "loss": 2.4473,
      "theoretical_loss": 3.3443916912391587,
      "tokens_seen": 2738028544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001732929292929293,
      "loss": 2.4524,
      "theoretical_loss": 3.3443854217030395,
      "tokens_seen": 2738094080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001732727272727273,
      "loss": 2.5657,
      "theoretical_loss": 3.3443791523589947,
      "tokens_seen": 2738159616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017325252525252524,
      "loss": 2.4131,
      "theoretical_loss": 3.3443728832070145,
      "tokens_seen": 2738225152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017323232323232323,
      "loss": 2.3343,
      "theoretical_loss": 3.344366614247088,
      "tokens_seen": 2738290688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001732121212121212,
      "loss": 2.3796,
      "theoretical_loss": 3.344360345479205,
      "tokens_seen": 2738356224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001731919191919192,
      "loss": 2.4422,
      "theoretical_loss": 3.3443540769033544,
      "tokens_seen": 2738421760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017317171717171718,
      "loss": 2.5798,
      "theoretical_loss": 3.344347808519526,
      "tokens_seen": 2738487296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017315151515151514,
      "loss": 2.3749,
      "theoretical_loss": 3.34434154032771,
      "tokens_seen": 2738552832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017313131313131315,
      "loss": 2.1612,
      "theoretical_loss": 3.3443352723278954,
      "tokens_seen": 2738618368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001731111111111111,
      "loss": 2.4693,
      "theoretical_loss": 3.3443290045200715,
      "tokens_seen": 2738683904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017309090909090912,
      "loss": 2.4712,
      "theoretical_loss": 3.344322736904228,
      "tokens_seen": 2738749440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017307070707070707,
      "loss": 2.4421,
      "theoretical_loss": 3.3443164694803547,
      "tokens_seen": 2738814976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017305050505050503,
      "loss": 2.4397,
      "theoretical_loss": 3.3443102022484408,
      "tokens_seen": 2738880512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017303030303030304,
      "loss": 2.688,
      "theoretical_loss": 3.344303935208476,
      "tokens_seen": 2738946048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000173010101010101,
      "loss": 2.6083,
      "theoretical_loss": 3.3442976683604493,
      "tokens_seen": 2739011584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000172989898989899,
      "loss": 2.6289,
      "theoretical_loss": 3.3442914017043512,
      "tokens_seen": 2739077120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017296969696969697,
      "loss": 2.5174,
      "theoretical_loss": 3.344285135240171,
      "tokens_seen": 2739142656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017294949494949495,
      "loss": 2.6495,
      "theoretical_loss": 3.3442788689678973,
      "tokens_seen": 2739208192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017292929292929293,
      "loss": 2.3474,
      "theoretical_loss": 3.3442726028875205,
      "tokens_seen": 2739273728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017290909090909092,
      "loss": 2.4873,
      "theoretical_loss": 3.3442663369990298,
      "tokens_seen": 2739339264
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001728888888888889,
      "loss": 2.3001,
      "theoretical_loss": 3.344260071302415,
      "tokens_seen": 2739404800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017286868686868686,
      "loss": 2.6948,
      "theoretical_loss": 3.3442538057976656,
      "tokens_seen": 2739470336
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017284848484848484,
      "loss": 2.3724,
      "theoretical_loss": 3.344247540484771,
      "tokens_seen": 2739535872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017282828282828283,
      "loss": 2.3583,
      "theoretical_loss": 3.3442412753637205,
      "tokens_seen": 2739601408
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1502052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.483609199523926,
      "objective/train/theoretical_loss": 3.344235010434504,
      "objective/train/tokens_used": 1098526176,
      "theoretical_loss": 3.344235010434504,
      "tokens_seen": 2739666944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001728080808080808,
      "loss": 2.5029,
      "theoretical_loss": 3.344235010434504,
      "tokens_seen": 2739666944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001727878787878788,
      "loss": 2.5748,
      "theoretical_loss": 3.3442287456971114,
      "tokens_seen": 2739732480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017276767676767678,
      "loss": 2.5983,
      "theoretical_loss": 3.344222481151531,
      "tokens_seen": 2739798016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017274747474747476,
      "loss": 2.5335,
      "theoretical_loss": 3.3442162167977534,
      "tokens_seen": 2739863552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017272727272727272,
      "loss": 2.4976,
      "theoretical_loss": 3.344209952635768,
      "tokens_seen": 2739929088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001727070707070707,
      "loss": 2.6703,
      "theoretical_loss": 3.344203688665564,
      "tokens_seen": 2739994624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001726868686868687,
      "loss": 2.4644,
      "theoretical_loss": 3.344197424887131,
      "tokens_seen": 2740060160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017266666666666667,
      "loss": 2.4428,
      "theoretical_loss": 3.344191161300459,
      "tokens_seen": 2740125696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017264646464646466,
      "loss": 2.5132,
      "theoretical_loss": 3.3441848979055373,
      "tokens_seen": 2740191232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017262626262626264,
      "loss": 2.4459,
      "theoretical_loss": 3.344178634702355,
      "tokens_seen": 2740256768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001726060606060606,
      "loss": 2.3425,
      "theoretical_loss": 3.3441723716909024,
      "tokens_seen": 2740322304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001725858585858586,
      "loss": 2.5537,
      "theoretical_loss": 3.3441661088711685,
      "tokens_seen": 2740387840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017256565656565656,
      "loss": 2.3677,
      "theoretical_loss": 3.3441598462431426,
      "tokens_seen": 2740453376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017254545454545455,
      "loss": 2.4816,
      "theoretical_loss": 3.344153583806815,
      "tokens_seen": 2740518912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017252525252525253,
      "loss": 2.6452,
      "theoretical_loss": 3.3441473215621746,
      "tokens_seen": 2740584448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001725050505050505,
      "loss": 2.4854,
      "theoretical_loss": 3.344141059509212,
      "tokens_seen": 2740649984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001724848484848485,
      "loss": 2.5062,
      "theoretical_loss": 3.344134797647915,
      "tokens_seen": 2740715520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017246464646464646,
      "loss": 2.3549,
      "theoretical_loss": 3.344128535978275,
      "tokens_seen": 2740781056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017244444444444447,
      "loss": 2.2427,
      "theoretical_loss": 3.3441222745002803,
      "tokens_seen": 2740846592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017242424242424242,
      "loss": 2.5639,
      "theoretical_loss": 3.3441160132139207,
      "tokens_seen": 2740912128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001724040404040404,
      "loss": 2.1027,
      "theoretical_loss": 3.344109752119186,
      "tokens_seen": 2740977664
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001723838383838384,
      "loss": 2.4572,
      "theoretical_loss": 3.3441034912160656,
      "tokens_seen": 2741043200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017236363636363635,
      "loss": 2.3577,
      "theoretical_loss": 3.3440972305045493,
      "tokens_seen": 2741108736
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017234343434343436,
      "loss": 2.4761,
      "theoretical_loss": 3.3440909699846264,
      "tokens_seen": 2741174272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017232323232323232,
      "loss": 2.251,
      "theoretical_loss": 3.3440847096562867,
      "tokens_seen": 2741239808
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1503417,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7360618114471436,
      "objective/train/theoretical_loss": 3.344078449519519,
      "objective/train/tokens_used": 1100164576,
      "theoretical_loss": 3.344078449519519,
      "tokens_seen": 2741305344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001723030303030303,
      "loss": 2.5625,
      "theoretical_loss": 3.344078449519519,
      "tokens_seen": 2741305344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017228282828282829,
      "loss": 2.4135,
      "theoretical_loss": 3.344072189574314,
      "tokens_seen": 2741370880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017226262626262627,
      "loss": 2.5926,
      "theoretical_loss": 3.3440659298206605,
      "tokens_seen": 2741436416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017224242424242425,
      "loss": 2.6506,
      "theoretical_loss": 3.3440596702585483,
      "tokens_seen": 2741501952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017222222222222224,
      "loss": 2.3075,
      "theoretical_loss": 3.344053410887967,
      "tokens_seen": 2741567488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001722020202020202,
      "loss": 2.4401,
      "theoretical_loss": 3.3440471517089057,
      "tokens_seen": 2741633024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017218181818181818,
      "loss": 2.5614,
      "theoretical_loss": 3.3440408927213547,
      "tokens_seen": 2741698560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017216161616161616,
      "loss": 2.707,
      "theoretical_loss": 3.344034633925303,
      "tokens_seen": 2741764096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017214141414141415,
      "loss": 2.4705,
      "theoretical_loss": 3.3440283753207405,
      "tokens_seen": 2741829632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017212121212121213,
      "loss": 2.3959,
      "theoretical_loss": 3.3440221169076563,
      "tokens_seen": 2741895168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001721010101010101,
      "loss": 2.4262,
      "theoretical_loss": 3.3440158586860407,
      "tokens_seen": 2741960704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001720808080808081,
      "loss": 2.6377,
      "theoretical_loss": 3.3440096006558826,
      "tokens_seen": 2742026240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017206060606060605,
      "loss": 2.6157,
      "theoretical_loss": 3.344003342817172,
      "tokens_seen": 2742091776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017204040404040407,
      "loss": 2.3218,
      "theoretical_loss": 3.343997085169898,
      "tokens_seen": 2742157312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017202020202020202,
      "loss": 2.3688,
      "theoretical_loss": 3.343990827714051,
      "tokens_seen": 2742222848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017199999999999998,
      "loss": 2.7244,
      "theoretical_loss": 3.3439845704496194,
      "tokens_seen": 2742288384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000171979797979798,
      "loss": 2.4522,
      "theoretical_loss": 3.343978313376594,
      "tokens_seen": 2742353920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017195959595959595,
      "loss": 2.3829,
      "theoretical_loss": 3.3439720564949633,
      "tokens_seen": 2742419456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017193939393939396,
      "loss": 2.5424,
      "theoretical_loss": 3.3439657998047174,
      "tokens_seen": 2742484992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017191919191919192,
      "loss": 2.4937,
      "theoretical_loss": 3.343959543305846,
      "tokens_seen": 2742550528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017189898989898993,
      "loss": 2.4347,
      "theoretical_loss": 3.3439532869983384,
      "tokens_seen": 2742616064
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017187878787878788,
      "loss": 2.6582,
      "theoretical_loss": 3.3439470308821844,
      "tokens_seen": 2742681600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017185858585858584,
      "loss": 2.4595,
      "theoretical_loss": 3.3439407749573733,
      "tokens_seen": 2742747136
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017183838383838385,
      "loss": 2.5899,
      "theoretical_loss": 3.343934519223895,
      "tokens_seen": 2742812672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001718181818181818,
      "loss": 2.5765,
      "theoretical_loss": 3.3439282636817387,
      "tokens_seen": 2742878208
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1504564,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4868407249450684,
      "objective/train/theoretical_loss": 3.3439220083308943,
      "objective/train/tokens_used": 1101802976,
      "theoretical_loss": 3.3439220083308943,
      "tokens_seen": 2742943744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017179797979797982,
      "loss": 2.295,
      "theoretical_loss": 3.3439220083308943,
      "tokens_seen": 2742943744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017177777777777778,
      "loss": 2.4951,
      "theoretical_loss": 3.343915753171351,
      "tokens_seen": 2743009280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017175757575757576,
      "loss": 2.4665,
      "theoretical_loss": 3.343909498203099,
      "tokens_seen": 2743074816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017173737373737374,
      "loss": 2.5766,
      "theoretical_loss": 3.3439032434261273,
      "tokens_seen": 2743140352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017171717171717173,
      "loss": 2.5329,
      "theoretical_loss": 3.3438969888404255,
      "tokens_seen": 2743205888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001716969696969697,
      "loss": 2.4222,
      "theoretical_loss": 3.3438907344459836,
      "tokens_seen": 2743271424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017167676767676767,
      "loss": 2.3349,
      "theoretical_loss": 3.343884480242791,
      "tokens_seen": 2743336960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017165656565656565,
      "loss": 2.3954,
      "theoretical_loss": 3.343878226230837,
      "tokens_seen": 2743402496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017163636363636364,
      "loss": 2.4457,
      "theoretical_loss": 3.3438719724101116,
      "tokens_seen": 2743468032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017161616161616162,
      "loss": 2.7806,
      "theoretical_loss": 3.343865718780604,
      "tokens_seen": 2743533568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001715959595959596,
      "loss": 2.3346,
      "theoretical_loss": 3.3438594653423044,
      "tokens_seen": 2743599104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001715757575757576,
      "loss": 2.4174,
      "theoretical_loss": 3.3438532120952016,
      "tokens_seen": 2743664640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017155555555555555,
      "loss": 2.2891,
      "theoretical_loss": 3.3438469590392854,
      "tokens_seen": 2743730176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017153535353535356,
      "loss": 2.4593,
      "theoretical_loss": 3.343840706174546,
      "tokens_seen": 2743795712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001715151515151515,
      "loss": 2.2727,
      "theoretical_loss": 3.3438344535009725,
      "tokens_seen": 2743861248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001714949494949495,
      "loss": 2.6239,
      "theoretical_loss": 3.3438282010185545,
      "tokens_seen": 2743926784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017147474747474748,
      "loss": 2.3674,
      "theoretical_loss": 3.343821948727282,
      "tokens_seen": 2743992320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017145454545454544,
      "loss": 2.4576,
      "theoretical_loss": 3.3438156966271433,
      "tokens_seen": 2744057856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017143434343434345,
      "loss": 2.4897,
      "theoretical_loss": 3.3438094447181292,
      "tokens_seen": 2744123392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001714141414141414,
      "loss": 2.5255,
      "theoretical_loss": 3.3438031930002294,
      "tokens_seen": 2744188928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017139393939393942,
      "loss": 2.5069,
      "theoretical_loss": 3.3437969414734328,
      "tokens_seen": 2744254464
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017137373737373737,
      "loss": 2.5154,
      "theoretical_loss": 3.3437906901377294,
      "tokens_seen": 2744320000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017135353535353536,
      "loss": 2.5421,
      "theoretical_loss": 3.343784438993109,
      "tokens_seen": 2744385536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017133333333333334,
      "loss": 2.4949,
      "theoretical_loss": 3.3437781880395603,
      "tokens_seen": 2744451072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001713131313131313,
      "loss": 2.6241,
      "theoretical_loss": 3.3437719372770736,
      "tokens_seen": 2744516608
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1505281,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.427666664123535,
      "objective/train/theoretical_loss": 3.3437656867056385,
      "objective/train/tokens_used": 1103441376,
      "theoretical_loss": 3.3437656867056385,
      "tokens_seen": 2744582144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001712929292929293,
      "loss": 2.2352,
      "theoretical_loss": 3.3437656867056385,
      "tokens_seen": 2744582144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017127272727272727,
      "loss": 2.6694,
      "theoretical_loss": 3.3437594363252443,
      "tokens_seen": 2744647680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017125252525252525,
      "loss": 2.4521,
      "theoretical_loss": 3.343753186135881,
      "tokens_seen": 2744713216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017123232323232323,
      "loss": 2.6045,
      "theoretical_loss": 3.343746936137538,
      "tokens_seen": 2744778752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017121212121212122,
      "loss": 2.4601,
      "theoretical_loss": 3.3437406863302046,
      "tokens_seen": 2744844288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001711919191919192,
      "loss": 2.2838,
      "theoretical_loss": 3.343734436713871,
      "tokens_seen": 2744909824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017117171717171716,
      "loss": 2.4298,
      "theoretical_loss": 3.3437281872885265,
      "tokens_seen": 2744975360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017115151515151514,
      "loss": 2.377,
      "theoretical_loss": 3.3437219380541605,
      "tokens_seen": 2745040896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017113131313131313,
      "loss": 2.4791,
      "theoretical_loss": 3.3437156890107627,
      "tokens_seen": 2745106432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001711111111111111,
      "loss": 2.6776,
      "theoretical_loss": 3.3437094401583227,
      "tokens_seen": 2745171968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001710909090909091,
      "loss": 2.3895,
      "theoretical_loss": 3.3437031914968305,
      "tokens_seen": 2745237504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017107070707070708,
      "loss": 2.4561,
      "theoretical_loss": 3.3436969430262753,
      "tokens_seen": 2745303040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017105050505050506,
      "loss": 2.4472,
      "theoretical_loss": 3.343690694746647,
      "tokens_seen": 2745368576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017103030303030305,
      "loss": 2.3924,
      "theoretical_loss": 3.3436844466579347,
      "tokens_seen": 2745434112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000171010101010101,
      "loss": 2.4105,
      "theoretical_loss": 3.3436781987601285,
      "tokens_seen": 2745499648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000170989898989899,
      "loss": 2.4297,
      "theoretical_loss": 3.343671951053218,
      "tokens_seen": 2745565184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017096969696969697,
      "loss": 2.2014,
      "theoretical_loss": 3.3436657035371926,
      "tokens_seen": 2745630720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017094949494949496,
      "loss": 2.3567,
      "theoretical_loss": 3.3436594562120416,
      "tokens_seen": 2745696256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017092929292929294,
      "loss": 2.5734,
      "theoretical_loss": 3.3436532090777553,
      "tokens_seen": 2745761792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001709090909090909,
      "loss": 2.378,
      "theoretical_loss": 3.343646962134323,
      "tokens_seen": 2745827328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001708888888888889,
      "loss": 2.5957,
      "theoretical_loss": 3.343640715381734,
      "tokens_seen": 2745892864
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017086868686868686,
      "loss": 2.3871,
      "theoretical_loss": 3.3436344688199786,
      "tokens_seen": 2745958400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017084848484848488,
      "loss": 2.5008,
      "theoretical_loss": 3.343628222449046,
      "tokens_seen": 2746023936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017082828282828283,
      "loss": 2.4392,
      "theoretical_loss": 3.343621976268926,
      "tokens_seen": 2746089472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001708080808080808,
      "loss": 2.4469,
      "theoretical_loss": 3.3436157302796072,
      "tokens_seen": 2746155008
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1506636,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.252068519592285,
      "objective/train/theoretical_loss": 3.343609484481081,
      "objective/train/tokens_used": 1105079776,
      "theoretical_loss": 3.343609484481081,
      "tokens_seen": 2746220544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001707878787878788,
      "loss": 2.4955,
      "theoretical_loss": 3.343609484481081,
      "tokens_seen": 2746220544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017076767676767676,
      "loss": 2.8094,
      "theoretical_loss": 3.343603238873336,
      "tokens_seen": 2746286080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017074747474747477,
      "loss": 2.642,
      "theoretical_loss": 3.3435969934563614,
      "tokens_seen": 2746351616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017072727272727273,
      "loss": 2.5326,
      "theoretical_loss": 3.343590748230148,
      "tokens_seen": 2746417152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001707070707070707,
      "loss": 2.5533,
      "theoretical_loss": 3.3435845031946845,
      "tokens_seen": 2746482688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001706868686868687,
      "loss": 2.4508,
      "theoretical_loss": 3.343578258349961,
      "tokens_seen": 2746548224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017066666666666668,
      "loss": 2.5825,
      "theoretical_loss": 3.3435720136959666,
      "tokens_seen": 2746613760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017064646464646466,
      "loss": 2.2389,
      "theoretical_loss": 3.343565769232691,
      "tokens_seen": 2746679296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017062626262626262,
      "loss": 2.4324,
      "theoretical_loss": 3.3435595249601247,
      "tokens_seen": 2746744832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001706060606060606,
      "loss": 2.5639,
      "theoretical_loss": 3.343553280878256,
      "tokens_seen": 2746810368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017058585858585859,
      "loss": 2.59,
      "theoretical_loss": 3.3435470369870757,
      "tokens_seen": 2746875904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017056565656565657,
      "loss": 2.6188,
      "theoretical_loss": 3.3435407932865733,
      "tokens_seen": 2746941440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017054545454545455,
      "loss": 2.5078,
      "theoretical_loss": 3.3435345497767375,
      "tokens_seen": 2747006976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017052525252525254,
      "loss": 2.3648,
      "theoretical_loss": 3.3435283064575585,
      "tokens_seen": 2747072512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001705050505050505,
      "loss": 2.3924,
      "theoretical_loss": 3.343522063329026,
      "tokens_seen": 2747138048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001704848484848485,
      "loss": 2.5121,
      "theoretical_loss": 3.3435158203911297,
      "tokens_seen": 2747203584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017046464646464646,
      "loss": 2.4281,
      "theoretical_loss": 3.343509577643859,
      "tokens_seen": 2747269120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017044444444444445,
      "loss": 2.6118,
      "theoretical_loss": 3.3435033350872034,
      "tokens_seen": 2747334656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017042424242424243,
      "loss": 2.5611,
      "theoretical_loss": 3.343497092721153,
      "tokens_seen": 2747400192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001704040404040404,
      "loss": 2.527,
      "theoretical_loss": 3.3434908505456975,
      "tokens_seen": 2747465728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001703838383838384,
      "loss": 2.4903,
      "theoretical_loss": 3.343484608560826,
      "tokens_seen": 2747531264
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017036363636363636,
      "loss": 2.4251,
      "theoretical_loss": 3.3434783667665284,
      "tokens_seen": 2747596800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017034343434343437,
      "loss": 2.7304,
      "theoretical_loss": 3.343472125162794,
      "tokens_seen": 2747662336
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017032323232323232,
      "loss": 2.4272,
      "theoretical_loss": 3.343465883749613,
      "tokens_seen": 2747727872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017030303030303028,
      "loss": 2.3809,
      "theoretical_loss": 3.343459642526975,
      "tokens_seen": 2747793408
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1507074,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4767045974731445,
      "objective/train/theoretical_loss": 3.3434534014948687,
      "objective/train/tokens_used": 1106718176,
      "theoretical_loss": 3.3434534014948687,
      "tokens_seen": 2747858944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001702828282828283,
      "loss": 2.473,
      "theoretical_loss": 3.3434534014948687,
      "tokens_seen": 2747858944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017026262626262625,
      "loss": 2.5404,
      "theoretical_loss": 3.343447160653285,
      "tokens_seen": 2747924480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017024242424242426,
      "loss": 2.6679,
      "theoretical_loss": 3.343440920002213,
      "tokens_seen": 2747990016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017022222222222222,
      "loss": 2.5049,
      "theoretical_loss": 3.343434679541642,
      "tokens_seen": 2748055552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017020202020202023,
      "loss": 2.6454,
      "theoretical_loss": 3.3434284392715625,
      "tokens_seen": 2748121088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017018181818181818,
      "loss": 2.4172,
      "theoretical_loss": 3.3434221991919633,
      "tokens_seen": 2748186624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017016161616161617,
      "loss": 2.694,
      "theoretical_loss": 3.343415959302834,
      "tokens_seen": 2748252160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017014141414141415,
      "loss": 2.4003,
      "theoretical_loss": 3.343409719604165,
      "tokens_seen": 2748317696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001701212121212121,
      "loss": 2.5977,
      "theoretical_loss": 3.3434034800959456,
      "tokens_seen": 2748383232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017010101010101012,
      "loss": 2.5354,
      "theoretical_loss": 3.3433972407781654,
      "tokens_seen": 2748448768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017008080808080808,
      "loss": 2.3547,
      "theoretical_loss": 3.343391001650814,
      "tokens_seen": 2748514304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017006060606060606,
      "loss": 2.4886,
      "theoretical_loss": 3.343384762713881,
      "tokens_seen": 2748579840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017004040404040404,
      "loss": 2.3441,
      "theoretical_loss": 3.343378523967356,
      "tokens_seen": 2748645376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017002020202020203,
      "loss": 2.563,
      "theoretical_loss": 3.3433722854112293,
      "tokens_seen": 2748710912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017,
      "loss": 2.5884,
      "theoretical_loss": 3.34336604704549,
      "tokens_seen": 2748776448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000169979797979798,
      "loss": 2.2748,
      "theoretical_loss": 3.3433598088701273,
      "tokens_seen": 2748841984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016995959595959595,
      "loss": 2.4044,
      "theoretical_loss": 3.3433535708851316,
      "tokens_seen": 2748907520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016993939393939394,
      "loss": 2.5009,
      "theoretical_loss": 3.343347333090492,
      "tokens_seen": 2748973056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016991919191919192,
      "loss": 2.6287,
      "theoretical_loss": 3.3433410954861986,
      "tokens_seen": 2749038592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001698989898989899,
      "loss": 2.5195,
      "theoretical_loss": 3.3433348580722413,
      "tokens_seen": 2749104128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001698787878787879,
      "loss": 2.4808,
      "theoretical_loss": 3.343328620848609,
      "tokens_seen": 2749169664
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016985858585858585,
      "loss": 2.5336,
      "theoretical_loss": 3.3433223838152917,
      "tokens_seen": 2749235200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016983838383838386,
      "loss": 2.5648,
      "theoretical_loss": 3.343316146972279,
      "tokens_seen": 2749300736
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016981818181818181,
      "loss": 2.565,
      "theoretical_loss": 3.3433099103195607,
      "tokens_seen": 2749366272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016979797979797982,
      "loss": 2.3842,
      "theoretical_loss": 3.3433036738571267,
      "tokens_seen": 2749431808
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1508309,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5114078521728516,
      "objective/train/theoretical_loss": 3.3432974375849662,
      "objective/train/tokens_used": 1108356576,
      "theoretical_loss": 3.3432974375849662,
      "tokens_seen": 2749497344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016977777777777778,
      "loss": 2.4407,
      "theoretical_loss": 3.3432974375849662,
      "tokens_seen": 2749497344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016975757575757574,
      "loss": 2.3782,
      "theoretical_loss": 3.3432912015030687,
      "tokens_seen": 2749562880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016973737373737375,
      "loss": 2.4324,
      "theoretical_loss": 3.3432849656114243,
      "tokens_seen": 2749628416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001697171717171717,
      "loss": 2.5037,
      "theoretical_loss": 3.343278729910023,
      "tokens_seen": 2749693952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016969696969696972,
      "loss": 2.3937,
      "theoretical_loss": 3.343272494398853,
      "tokens_seen": 2749759488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016967676767676767,
      "loss": 2.5391,
      "theoretical_loss": 3.3432662590779056,
      "tokens_seen": 2749825024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016965656565656566,
      "loss": 2.518,
      "theoretical_loss": 3.34326002394717,
      "tokens_seen": 2749890560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016963636363636364,
      "loss": 2.5049,
      "theoretical_loss": 3.3432537890066354,
      "tokens_seen": 2749956096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016961616161616163,
      "loss": 2.5801,
      "theoretical_loss": 3.343247554256292,
      "tokens_seen": 2750021632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001695959595959596,
      "loss": 2.4574,
      "theoretical_loss": 3.3432413196961286,
      "tokens_seen": 2750087168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016957575757575757,
      "loss": 2.5269,
      "theoretical_loss": 3.343235085326136,
      "tokens_seen": 2750152704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016955555555555555,
      "loss": 2.3649,
      "theoretical_loss": 3.343228851146303,
      "tokens_seen": 2750218240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016953535353535353,
      "loss": 2.583,
      "theoretical_loss": 3.34322261715662,
      "tokens_seen": 2750283776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016951515151515152,
      "loss": 2.5673,
      "theoretical_loss": 3.343216383357076,
      "tokens_seen": 2750349312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001694949494949495,
      "loss": 2.4754,
      "theoretical_loss": 3.343210149747661,
      "tokens_seen": 2750414848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001694747474747475,
      "loss": 2.6453,
      "theoretical_loss": 3.3432039163283647,
      "tokens_seen": 2750480384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016945454545454544,
      "loss": 2.4965,
      "theoretical_loss": 3.3431976830991768,
      "tokens_seen": 2750545920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016943434343434343,
      "loss": 2.4504,
      "theoretical_loss": 3.3431914500600866,
      "tokens_seen": 2750611456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001694141414141414,
      "loss": 2.5478,
      "theoretical_loss": 3.343185217211084,
      "tokens_seen": 2750676992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001693939393939394,
      "loss": 2.5805,
      "theoretical_loss": 3.343178984552159,
      "tokens_seen": 2750742528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016937373737373738,
      "loss": 2.542,
      "theoretical_loss": 3.3431727520833006,
      "tokens_seen": 2750808064
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016935353535353536,
      "loss": 2.4412,
      "theoretical_loss": 3.343166519804499,
      "tokens_seen": 2750873600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016933333333333335,
      "loss": 2.3483,
      "theoretical_loss": 3.3431602877157442,
      "tokens_seen": 2750939136
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001693131313131313,
      "loss": 2.4029,
      "theoretical_loss": 3.343154055817025,
      "tokens_seen": 2751004672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016929292929292932,
      "loss": 2.6504,
      "theoretical_loss": 3.343147824108332,
      "tokens_seen": 2751070208
    },
    {
      "epoch": 0.66,
      "objective/train/docs_used": 1509529,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8860881328582764,
      "objective/train/theoretical_loss": 3.343141592589654,
      "objective/train/tokens_used": 1109994976,
      "theoretical_loss": 3.343141592589654,
      "tokens_seen": 2751135744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016927272727272727,
      "loss": 2.7355,
      "theoretical_loss": 3.343141592589654,
      "tokens_seen": 2751135744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016925252525252526,
      "loss": 2.5975,
      "theoretical_loss": 3.3431353612609804,
      "tokens_seen": 2751201280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016923232323232324,
      "loss": 2.6238,
      "theoretical_loss": 3.3431291301223025,
      "tokens_seen": 2751266816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001692121212121212,
      "loss": 2.5902,
      "theoretical_loss": 3.3431228991736086,
      "tokens_seen": 2751332352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001691919191919192,
      "loss": 2.5332,
      "theoretical_loss": 3.343116668414889,
      "tokens_seen": 2751397888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016917171717171716,
      "loss": 2.3857,
      "theoretical_loss": 3.343110437846133,
      "tokens_seen": 2751463424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016915151515151518,
      "loss": 2.5115,
      "theoretical_loss": 3.3431042074673307,
      "tokens_seen": 2751528960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016913131313131313,
      "loss": 2.6071,
      "theoretical_loss": 3.3430979772784712,
      "tokens_seen": 2751594496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016911111111111112,
      "loss": 2.8059,
      "theoretical_loss": 3.3430917472795447,
      "tokens_seen": 2751660032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001690909090909091,
      "loss": 2.4418,
      "theoretical_loss": 3.343085517470541,
      "tokens_seen": 2751725568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016907070707070706,
      "loss": 2.4796,
      "theoretical_loss": 3.3430792878514493,
      "tokens_seen": 2751791104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016905050505050507,
      "loss": 2.3931,
      "theoretical_loss": 3.3430730584222594,
      "tokens_seen": 2751856640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016903030303030303,
      "loss": 2.5562,
      "theoretical_loss": 3.343066829182961,
      "tokens_seen": 2751922176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000169010101010101,
      "loss": 2.6501,
      "theoretical_loss": 3.343060600133544,
      "tokens_seen": 2751987712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000168989898989899,
      "loss": 2.4794,
      "theoretical_loss": 3.343054371273998,
      "tokens_seen": 2752053248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016896969696969698,
      "loss": 2.339,
      "theoretical_loss": 3.343048142604313,
      "tokens_seen": 2752118784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016894949494949496,
      "loss": 2.6137,
      "theoretical_loss": 3.343041914124478,
      "tokens_seen": 2752184320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016892929292929295,
      "loss": 2.7218,
      "theoretical_loss": 3.343035685834483,
      "tokens_seen": 2752249856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001689090909090909,
      "loss": 2.4793,
      "theoretical_loss": 3.3430294577343176,
      "tokens_seen": 2752315392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016888888888888889,
      "loss": 2.5701,
      "theoretical_loss": 3.343023229823972,
      "tokens_seen": 2752380928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016886868686868687,
      "loss": 2.5495,
      "theoretical_loss": 3.3430170021034353,
      "tokens_seen": 2752446464
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016884848484848485,
      "loss": 2.6591,
      "theoretical_loss": 3.3430107745726976,
      "tokens_seen": 2752512000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016882828282828284,
      "loss": 2.4868,
      "theoretical_loss": 3.3430045472317484,
      "tokens_seen": 2752577536
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001688080808080808,
      "loss": 2.3851,
      "theoretical_loss": 3.3429983200805773,
      "tokens_seen": 2752643072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001687878787878788,
      "loss": 2.2933,
      "theoretical_loss": 3.342992093119174,
      "tokens_seen": 2752708608
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1510029,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6230833530426025,
      "objective/train/theoretical_loss": 3.3429858663475285,
      "objective/train/tokens_used": 1111633376,
      "theoretical_loss": 3.3429858663475285,
      "tokens_seen": 2752774144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016876767676767676,
      "loss": 2.4983,
      "theoretical_loss": 3.3429858663475285,
      "tokens_seen": 2752774144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016874747474747477,
      "loss": 2.4781,
      "theoretical_loss": 3.3429796397656304,
      "tokens_seen": 2752839680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016872727272727273,
      "loss": 2.6614,
      "theoretical_loss": 3.342973413373469,
      "tokens_seen": 2752905216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001687070707070707,
      "loss": 2.4224,
      "theoretical_loss": 3.3429671871710345,
      "tokens_seen": 2752970752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001686868686868687,
      "loss": 2.4826,
      "theoretical_loss": 3.342960961158316,
      "tokens_seen": 2753036288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016866666666666666,
      "loss": 2.6242,
      "theoretical_loss": 3.3429547353353044,
      "tokens_seen": 2753101824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016864646464646467,
      "loss": 2.275,
      "theoretical_loss": 3.3429485097019884,
      "tokens_seen": 2753167360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016862626262626262,
      "loss": 2.1682,
      "theoretical_loss": 3.342942284258358,
      "tokens_seen": 2753232896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001686060606060606,
      "loss": 2.2587,
      "theoretical_loss": 3.3429360590044026,
      "tokens_seen": 2753298432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001685858585858586,
      "loss": 2.676,
      "theoretical_loss": 3.3429298339401123,
      "tokens_seen": 2753363968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016856565656565655,
      "loss": 2.6414,
      "theoretical_loss": 3.342923609065476,
      "tokens_seen": 2753429504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016854545454545456,
      "loss": 2.5263,
      "theoretical_loss": 3.342917384380485,
      "tokens_seen": 2753495040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016852525252525252,
      "loss": 2.3916,
      "theoretical_loss": 3.3429111598851278,
      "tokens_seen": 2753560576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016850505050505053,
      "loss": 2.6404,
      "theoretical_loss": 3.3429049355793943,
      "tokens_seen": 2753626112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016848484848484848,
      "loss": 2.7259,
      "theoretical_loss": 3.342898711463274,
      "tokens_seen": 2753691648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016846464646464647,
      "loss": 2.4837,
      "theoretical_loss": 3.342892487536757,
      "tokens_seen": 2753757184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016844444444444445,
      "loss": 2.5165,
      "theoretical_loss": 3.3428862637998336,
      "tokens_seen": 2753822720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016842424242424244,
      "loss": 2.3796,
      "theoretical_loss": 3.342880040252492,
      "tokens_seen": 2753888256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016840404040404042,
      "loss": 2.4224,
      "theoretical_loss": 3.3428738168947234,
      "tokens_seen": 2753953792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016838383838383838,
      "loss": 2.3061,
      "theoretical_loss": 3.342867593726516,
      "tokens_seen": 2754019328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016836363636363636,
      "loss": 2.2304,
      "theoretical_loss": 3.342861370747861,
      "tokens_seen": 2754084864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016834343434343434,
      "loss": 2.3142,
      "theoretical_loss": 3.3428551479587476,
      "tokens_seen": 2754150400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016832323232323233,
      "loss": 2.635,
      "theoretical_loss": 3.342848925359165,
      "tokens_seen": 2754215936
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001683030303030303,
      "loss": 2.4569,
      "theoretical_loss": 3.3428427029491035,
      "tokens_seen": 2754281472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001682828282828283,
      "loss": 2.5827,
      "theoretical_loss": 3.342836480728552,
      "tokens_seen": 2754347008
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1510960,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.503450632095337,
      "objective/train/theoretical_loss": 3.342830258697502,
      "objective/train/tokens_used": 1113271776,
      "theoretical_loss": 3.342830258697502,
      "tokens_seen": 2754412544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016826262626262625,
      "loss": 2.4446,
      "theoretical_loss": 3.342830258697502,
      "tokens_seen": 2754412544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016824242424242426,
      "loss": 2.5433,
      "theoretical_loss": 3.3428240368559416,
      "tokens_seen": 2754478080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016822222222222222,
      "loss": 2.6198,
      "theoretical_loss": 3.342817815203861,
      "tokens_seen": 2754543616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001682020202020202,
      "loss": 2.5607,
      "theoretical_loss": 3.3428115937412497,
      "tokens_seen": 2754609152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001681818181818182,
      "loss": 2.3436,
      "theoretical_loss": 3.3428053724680975,
      "tokens_seen": 2754674688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016816161616161615,
      "loss": 2.6481,
      "theoretical_loss": 3.342799151384395,
      "tokens_seen": 2754740224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016814141414141416,
      "loss": 2.458,
      "theoretical_loss": 3.3427929304901305,
      "tokens_seen": 2754805760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016812121212121211,
      "loss": 2.4262,
      "theoretical_loss": 3.342786709785295,
      "tokens_seen": 2754871296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016810101010101013,
      "loss": 2.4641,
      "theoretical_loss": 3.3427804892698774,
      "tokens_seen": 2754936832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016808080808080808,
      "loss": 2.5883,
      "theoretical_loss": 3.342774268943867,
      "tokens_seen": 2755002368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016806060606060607,
      "loss": 2.5174,
      "theoretical_loss": 3.342768048807255,
      "tokens_seen": 2755067904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016804040404040405,
      "loss": 2.5805,
      "theoretical_loss": 3.3427618288600303,
      "tokens_seen": 2755133440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000168020202020202,
      "loss": 2.6488,
      "theoretical_loss": 3.3427556091021824,
      "tokens_seen": 2755198976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016800000000000002,
      "loss": 2.5924,
      "theoretical_loss": 3.3427493895337013,
      "tokens_seen": 2755264512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016797979797979797,
      "loss": 2.6395,
      "theoretical_loss": 3.3427431701545767,
      "tokens_seen": 2755330048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016795959595959596,
      "loss": 2.675,
      "theoretical_loss": 3.3427369509647984,
      "tokens_seen": 2755395584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016793939393939394,
      "loss": 2.2821,
      "theoretical_loss": 3.342730731964356,
      "tokens_seen": 2755461120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016791919191919193,
      "loss": 2.7047,
      "theoretical_loss": 3.3427245131532395,
      "tokens_seen": 2755526656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001678989898989899,
      "loss": 2.3648,
      "theoretical_loss": 3.342718294531438,
      "tokens_seen": 2755592192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016787878787878787,
      "loss": 2.3327,
      "theoretical_loss": 3.342712076098942,
      "tokens_seen": 2755657728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016785858585858585,
      "loss": 2.4148,
      "theoretical_loss": 3.342705857855741,
      "tokens_seen": 2755723264
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016783838383838384,
      "loss": 2.4149,
      "theoretical_loss": 3.3426996398018245,
      "tokens_seen": 2755788800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016781818181818182,
      "loss": 2.5137,
      "theoretical_loss": 3.3426934219371827,
      "tokens_seen": 2755854336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001677979797979798,
      "loss": 2.6958,
      "theoretical_loss": 3.3426872042618045,
      "tokens_seen": 2755919872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001677777777777778,
      "loss": 2.3498,
      "theoretical_loss": 3.3426809867756804,
      "tokens_seen": 2755985408
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1512146,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4062039852142334,
      "objective/train/theoretical_loss": 3.3426747694788,
      "objective/train/tokens_used": 1114910176,
      "theoretical_loss": 3.3426747694788,
      "tokens_seen": 2756050944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016775757575757574,
      "loss": 2.5385,
      "theoretical_loss": 3.3426747694788,
      "tokens_seen": 2756050944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016773737373737376,
      "loss": 2.3607,
      "theoretical_loss": 3.342668552371153,
      "tokens_seen": 2756116480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001677171717171717,
      "loss": 2.3892,
      "theoretical_loss": 3.342662335452729,
      "tokens_seen": 2756182016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001676969696969697,
      "loss": 2.364,
      "theoretical_loss": 3.3426561187235175,
      "tokens_seen": 2756247552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016767676767676768,
      "loss": 2.49,
      "theoretical_loss": 3.342649902183509,
      "tokens_seen": 2756313088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016765656565656566,
      "loss": 2.6073,
      "theoretical_loss": 3.3426436858326927,
      "tokens_seen": 2756378624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016763636363636365,
      "loss": 2.5562,
      "theoretical_loss": 3.3426374696710583,
      "tokens_seen": 2756444160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001676161616161616,
      "loss": 2.6286,
      "theoretical_loss": 3.3426312536985954,
      "tokens_seen": 2756509696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016759595959595962,
      "loss": 2.6142,
      "theoretical_loss": 3.342625037915295,
      "tokens_seen": 2756575232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016757575757575757,
      "loss": 2.6647,
      "theoretical_loss": 3.3426188223211453,
      "tokens_seen": 2756640768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016755555555555558,
      "loss": 2.4265,
      "theoretical_loss": 3.3426126069161364,
      "tokens_seen": 2756706304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016753535353535354,
      "loss": 2.3922,
      "theoretical_loss": 3.3426063917002584,
      "tokens_seen": 2756771840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001675151515151515,
      "loss": 2.6707,
      "theoretical_loss": 3.342600176673501,
      "tokens_seen": 2756837376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001674949494949495,
      "loss": 2.6063,
      "theoretical_loss": 3.342593961835854,
      "tokens_seen": 2756902912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016747474747474747,
      "loss": 2.5734,
      "theoretical_loss": 3.342587747187307,
      "tokens_seen": 2756968448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016745454545454548,
      "loss": 2.535,
      "theoretical_loss": 3.34258153272785,
      "tokens_seen": 2757033984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016743434343434343,
      "loss": 2.7146,
      "theoretical_loss": 3.342575318457472,
      "tokens_seen": 2757099520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016741414141414142,
      "loss": 2.4252,
      "theoretical_loss": 3.342569104376164,
      "tokens_seen": 2757165056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001673939393939394,
      "loss": 2.3517,
      "theoretical_loss": 3.3425628904839146,
      "tokens_seen": 2757230592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016737373737373739,
      "loss": 2.6101,
      "theoretical_loss": 3.342556676780714,
      "tokens_seen": 2757296128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016735353535353537,
      "loss": 2.4752,
      "theoretical_loss": 3.342550463266552,
      "tokens_seen": 2757361664
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016733333333333333,
      "loss": 2.4862,
      "theoretical_loss": 3.3425442499414184,
      "tokens_seen": 2757427200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001673131313131313,
      "loss": 2.7177,
      "theoretical_loss": 3.3425380368053026,
      "tokens_seen": 2757492736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001672929292929293,
      "loss": 2.7104,
      "theoretical_loss": 3.342531823858195,
      "tokens_seen": 2757558272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016727272727272728,
      "loss": 2.5433,
      "theoretical_loss": 3.3425256111000845,
      "tokens_seen": 2757623808
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1512703,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.699195384979248,
      "objective/train/theoretical_loss": 3.342519398530962,
      "objective/train/tokens_used": 1116548576,
      "theoretical_loss": 3.342519398530962,
      "tokens_seen": 2757689344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016725252525252526,
      "loss": 2.6129,
      "theoretical_loss": 3.342519398530962,
      "tokens_seen": 2757689344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016723232323232325,
      "loss": 2.3816,
      "theoretical_loss": 3.342513186150816,
      "tokens_seen": 2757754880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001672121212121212,
      "loss": 2.4668,
      "theoretical_loss": 3.342506973959637,
      "tokens_seen": 2757820416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016719191919191921,
      "loss": 2.5441,
      "theoretical_loss": 3.3425007619574147,
      "tokens_seen": 2757885952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016717171717171717,
      "loss": 2.574,
      "theoretical_loss": 3.342494550144139,
      "tokens_seen": 2757951488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016715151515151515,
      "loss": 2.6698,
      "theoretical_loss": 3.342488338519799,
      "tokens_seen": 2758017024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016713131313131314,
      "loss": 2.493,
      "theoretical_loss": 3.342482127084385,
      "tokens_seen": 2758082560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001671111111111111,
      "loss": 2.5522,
      "theoretical_loss": 3.3424759158378876,
      "tokens_seen": 2758148096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001670909090909091,
      "loss": 2.6004,
      "theoretical_loss": 3.342469704780295,
      "tokens_seen": 2758213632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016707070707070706,
      "loss": 2.4535,
      "theoretical_loss": 3.3424634939115974,
      "tokens_seen": 2758279168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016705050505050507,
      "loss": 2.6645,
      "theoretical_loss": 3.3424572832317847,
      "tokens_seen": 2758344704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016703030303030303,
      "loss": 2.3888,
      "theoretical_loss": 3.342451072740847,
      "tokens_seen": 2758410240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000167010101010101,
      "loss": 2.5406,
      "theoretical_loss": 3.3424448624387737,
      "tokens_seen": 2758475776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000166989898989899,
      "loss": 2.5109,
      "theoretical_loss": 3.342438652325555,
      "tokens_seen": 2758541312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016696969696969696,
      "loss": 2.5602,
      "theoretical_loss": 3.34243244240118,
      "tokens_seen": 2758606848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016694949494949497,
      "loss": 2.466,
      "theoretical_loss": 3.342426232665639,
      "tokens_seen": 2758672384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016692929292929292,
      "loss": 2.5179,
      "theoretical_loss": 3.3424200231189216,
      "tokens_seen": 2758737920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001669090909090909,
      "loss": 2.7382,
      "theoretical_loss": 3.3424138137610173,
      "tokens_seen": 2758803456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001668888888888889,
      "loss": 2.6637,
      "theoretical_loss": 3.3424076045919167,
      "tokens_seen": 2758868992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016686868686868688,
      "loss": 2.4876,
      "theoretical_loss": 3.3424013956116085,
      "tokens_seen": 2758934528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016684848484848486,
      "loss": 2.655,
      "theoretical_loss": 3.342395186820083,
      "tokens_seen": 2759000064
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016682828282828282,
      "loss": 2.5795,
      "theoretical_loss": 3.3423889782173304,
      "tokens_seen": 2759065600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016680808080808083,
      "loss": 2.7132,
      "theoretical_loss": 3.34238276980334,
      "tokens_seen": 2759131136
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016678787878787878,
      "loss": 2.669,
      "theoretical_loss": 3.342376561578101,
      "tokens_seen": 2759196672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016676767676767677,
      "loss": 2.724,
      "theoretical_loss": 3.3423703535416043,
      "tokens_seen": 2759262208
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1514070,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4820828437805176,
      "objective/train/theoretical_loss": 3.342364145693839,
      "objective/train/tokens_used": 1118186976,
      "theoretical_loss": 3.342364145693839,
      "tokens_seen": 2759327744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016674747474747475,
      "loss": 2.5302,
      "theoretical_loss": 3.342364145693839,
      "tokens_seen": 2759327744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016672727272727274,
      "loss": 2.3592,
      "theoretical_loss": 3.3423579380347954,
      "tokens_seen": 2759393280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016670707070707072,
      "loss": 2.5375,
      "theoretical_loss": 3.342351730564463,
      "tokens_seen": 2759458816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001666868686868687,
      "loss": 2.4038,
      "theoretical_loss": 3.342345523282831,
      "tokens_seen": 2759524352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016666666666666666,
      "loss": 2.3906,
      "theoretical_loss": 3.34233931618989,
      "tokens_seen": 2759589888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016664646464646465,
      "loss": 2.5902,
      "theoretical_loss": 3.3423331092856294,
      "tokens_seen": 2759655424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016662626262626263,
      "loss": 2.6271,
      "theoretical_loss": 3.342326902570039,
      "tokens_seen": 2759720960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001666060606060606,
      "loss": 2.3666,
      "theoretical_loss": 3.342320696043109,
      "tokens_seen": 2759786496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001665858585858586,
      "loss": 2.484,
      "theoretical_loss": 3.3423144897048287,
      "tokens_seen": 2759852032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016656565656565655,
      "loss": 2.3156,
      "theoretical_loss": 3.342308283555188,
      "tokens_seen": 2759917568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016654545454545456,
      "loss": 2.7336,
      "theoretical_loss": 3.3423020775941765,
      "tokens_seen": 2759983104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016652525252525252,
      "loss": 2.3339,
      "theoretical_loss": 3.3422958718217846,
      "tokens_seen": 2760048640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016650505050505053,
      "loss": 2.6066,
      "theoretical_loss": 3.3422896662380017,
      "tokens_seen": 2760114176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001664848484848485,
      "loss": 2.5801,
      "theoretical_loss": 3.3422834608428174,
      "tokens_seen": 2760179712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016646464646464645,
      "loss": 2.7557,
      "theoretical_loss": 3.3422772556362217,
      "tokens_seen": 2760245248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016644444444444446,
      "loss": 2.4409,
      "theoretical_loss": 3.342271050618204,
      "tokens_seen": 2760310784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016642424242424241,
      "loss": 2.6982,
      "theoretical_loss": 3.3422648457887547,
      "tokens_seen": 2760376320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016640404040404043,
      "loss": 2.5672,
      "theoretical_loss": 3.342258641147864,
      "tokens_seen": 2760441856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016638383838383838,
      "loss": 2.2897,
      "theoretical_loss": 3.3422524366955204,
      "tokens_seen": 2760507392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016636363636363637,
      "loss": 2.6074,
      "theoretical_loss": 3.342246232431714,
      "tokens_seen": 2760572928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016634343434343435,
      "loss": 2.4495,
      "theoretical_loss": 3.3422400283564357,
      "tokens_seen": 2760638464
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016632323232323233,
      "loss": 2.5237,
      "theoretical_loss": 3.342233824469674,
      "tokens_seen": 2760704000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016630303030303032,
      "loss": 2.4297,
      "theoretical_loss": 3.34222762077142,
      "tokens_seen": 2760769536
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016628282828282828,
      "loss": 2.5756,
      "theoretical_loss": 3.342221417261662,
      "tokens_seen": 2760835072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016626262626262626,
      "loss": 2.6041,
      "theoretical_loss": 3.3422152139403907,
      "tokens_seen": 2760900608
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1514349,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5181102752685547,
      "objective/train/theoretical_loss": 3.3422090108075957,
      "objective/train/tokens_used": 1119825376,
      "theoretical_loss": 3.3422090108075957,
      "tokens_seen": 2760966144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016624242424242424,
      "loss": 2.3828,
      "theoretical_loss": 3.3422090108075957,
      "tokens_seen": 2760966144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016622222222222223,
      "loss": 2.3228,
      "theoretical_loss": 3.342202807863267,
      "tokens_seen": 2761031680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001662020202020202,
      "loss": 2.522,
      "theoretical_loss": 3.342196605107394,
      "tokens_seen": 2761097216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001661818181818182,
      "loss": 2.6477,
      "theoretical_loss": 3.3421904025399667,
      "tokens_seen": 2761162752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016616161616161615,
      "loss": 2.4636,
      "theoretical_loss": 3.342184200160975,
      "tokens_seen": 2761228288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016614141414141414,
      "loss": 2.6654,
      "theoretical_loss": 3.342177997970409,
      "tokens_seen": 2761293824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016612121212121212,
      "loss": 2.58,
      "theoretical_loss": 3.3421717959682575,
      "tokens_seen": 2761359360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001661010101010101,
      "loss": 2.6119,
      "theoretical_loss": 3.3421655941545114,
      "tokens_seen": 2761424896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001660808080808081,
      "loss": 2.5517,
      "theoretical_loss": 3.34215939252916,
      "tokens_seen": 2761490432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016606060606060604,
      "loss": 2.8304,
      "theoretical_loss": 3.342153191092193,
      "tokens_seen": 2761555968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016604040404040406,
      "loss": 2.4527,
      "theoretical_loss": 3.3421469898436005,
      "tokens_seen": 2761621504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000166020202020202,
      "loss": 2.5967,
      "theoretical_loss": 3.3421407887833716,
      "tokens_seen": 2761687040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016600000000000002,
      "loss": 2.6755,
      "theoretical_loss": 3.3421345879114974,
      "tokens_seen": 2761752576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016597979797979798,
      "loss": 2.5588,
      "theoretical_loss": 3.342128387227967,
      "tokens_seen": 2761818112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016595959595959594,
      "loss": 2.6735,
      "theoretical_loss": 3.3421221867327695,
      "tokens_seen": 2761883648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016593939393939395,
      "loss": 2.3662,
      "theoretical_loss": 3.3421159864258962,
      "tokens_seen": 2761949184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001659191919191919,
      "loss": 2.3604,
      "theoretical_loss": 3.3421097863073355,
      "tokens_seen": 2762014720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016589898989898992,
      "loss": 2.4748,
      "theoretical_loss": 3.3421035863770783,
      "tokens_seen": 2762080256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016587878787878787,
      "loss": 2.4228,
      "theoretical_loss": 3.3420973866351136,
      "tokens_seen": 2762145792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016585858585858588,
      "loss": 2.5678,
      "theoretical_loss": 3.3420911870814316,
      "tokens_seen": 2762211328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016583838383838384,
      "loss": 2.4854,
      "theoretical_loss": 3.3420849877160226,
      "tokens_seen": 2762276864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016581818181818182,
      "loss": 2.4906,
      "theoretical_loss": 3.342078788538875,
      "tokens_seen": 2762342400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001657979797979798,
      "loss": 2.6658,
      "theoretical_loss": 3.3420725895499803,
      "tokens_seen": 2762407936
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016577777777777777,
      "loss": 2.5736,
      "theoretical_loss": 3.342066390749327,
      "tokens_seen": 2762473472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016575757575757578,
      "loss": 2.6851,
      "theoretical_loss": 3.3420601921369055,
      "tokens_seen": 2762539008
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1514959,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.672292470932007,
      "objective/train/theoretical_loss": 3.342053993712706,
      "objective/train/tokens_used": 1121463776,
      "theoretical_loss": 3.342053993712706,
      "tokens_seen": 2762604544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016573737373737373,
      "loss": 2.7192,
      "theoretical_loss": 3.342053993712706,
      "tokens_seen": 2762604544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016571717171717172,
      "loss": 2.548,
      "theoretical_loss": 3.3420477954767174,
      "tokens_seen": 2762670080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001656969696969697,
      "loss": 2.6235,
      "theoretical_loss": 3.3420415974289304,
      "tokens_seen": 2762735616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016567676767676769,
      "loss": 2.3474,
      "theoretical_loss": 3.342035399569334,
      "tokens_seen": 2762801152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016565656565656567,
      "loss": 2.6931,
      "theoretical_loss": 3.3420292018979185,
      "tokens_seen": 2762866688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016563636363636365,
      "loss": 2.6854,
      "theoretical_loss": 3.3420230044146737,
      "tokens_seen": 2762932224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001656161616161616,
      "loss": 2.3871,
      "theoretical_loss": 3.34201680711959,
      "tokens_seen": 2762997760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001655959595959596,
      "loss": 2.4446,
      "theoretical_loss": 3.3420106100126556,
      "tokens_seen": 2763063296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016557575757575758,
      "loss": 2.4995,
      "theoretical_loss": 3.342004413093862,
      "tokens_seen": 2763128832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016555555555555556,
      "loss": 2.5297,
      "theoretical_loss": 3.3419982163631983,
      "tokens_seen": 2763194368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016553535353535355,
      "loss": 2.6865,
      "theoretical_loss": 3.3419920198206543,
      "tokens_seen": 2763259904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001655151515151515,
      "loss": 2.5231,
      "theoretical_loss": 3.34198582346622,
      "tokens_seen": 2763325440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016549494949494951,
      "loss": 2.6046,
      "theoretical_loss": 3.341979627299885,
      "tokens_seen": 2763390976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016547474747474747,
      "loss": 2.5949,
      "theoretical_loss": 3.341973431321639,
      "tokens_seen": 2763456512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016545454545454548,
      "loss": 2.5555,
      "theoretical_loss": 3.341967235531473,
      "tokens_seen": 2763522048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016543434343434344,
      "loss": 2.6034,
      "theoretical_loss": 3.3419610399293753,
      "tokens_seen": 2763587584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001654141414141414,
      "loss": 2.5834,
      "theoretical_loss": 3.341954844515336,
      "tokens_seen": 2763653120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001653939393939394,
      "loss": 2.4433,
      "theoretical_loss": 3.341948649289346,
      "tokens_seen": 2763718656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016537373737373736,
      "loss": 2.4306,
      "theoretical_loss": 3.3419424542513942,
      "tokens_seen": 2763784192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016535353535353537,
      "loss": 2.8165,
      "theoretical_loss": 3.3419362594014705,
      "tokens_seen": 2763849728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016533333333333333,
      "loss": 2.4275,
      "theoretical_loss": 3.3419300647395653,
      "tokens_seen": 2763915264
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016531313131313132,
      "loss": 2.3649,
      "theoretical_loss": 3.341923870265668,
      "tokens_seen": 2763980800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001652929292929293,
      "loss": 2.444,
      "theoretical_loss": 3.341917675979768,
      "tokens_seen": 2764046336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016527272727272726,
      "loss": 2.6089,
      "theoretical_loss": 3.3419114818818563,
      "tokens_seen": 2764111872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016525252525252527,
      "loss": 2.6381,
      "theoretical_loss": 3.3419052879719215,
      "tokens_seen": 2764177408
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1515566,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4727749824523926,
      "objective/train/theoretical_loss": 3.3418990942499542,
      "objective/train/tokens_used": 1123102176,
      "theoretical_loss": 3.3418990942499542,
      "tokens_seen": 2764242944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016523232323232322,
      "loss": 2.4772,
      "theoretical_loss": 3.3418990942499542,
      "tokens_seen": 2764242944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001652121212121212,
      "loss": 2.5884,
      "theoretical_loss": 3.341892900715944,
      "tokens_seen": 2764308480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001651919191919192,
      "loss": 2.4037,
      "theoretical_loss": 3.3418867073698806,
      "tokens_seen": 2764374016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016517171717171718,
      "loss": 2.359,
      "theoretical_loss": 3.3418805142117547,
      "tokens_seen": 2764439552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016515151515151516,
      "loss": 2.5857,
      "theoretical_loss": 3.341874321241555,
      "tokens_seen": 2764505088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016513131313131314,
      "loss": 2.4537,
      "theoretical_loss": 3.3418681284592715,
      "tokens_seen": 2764570624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001651111111111111,
      "loss": 2.5181,
      "theoretical_loss": 3.3418619358648947,
      "tokens_seen": 2764636160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016509090909090908,
      "loss": 2.3551,
      "theoretical_loss": 3.3418557434584137,
      "tokens_seen": 2764701696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016507070707070707,
      "loss": 2.6376,
      "theoretical_loss": 3.341849551239819,
      "tokens_seen": 2764767232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016505050505050505,
      "loss": 2.2745,
      "theoretical_loss": 3.3418433592091,
      "tokens_seen": 2764832768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016503030303030304,
      "loss": 2.5105,
      "theoretical_loss": 3.341837167366247,
      "tokens_seen": 2764898304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016501010101010102,
      "loss": 2.54,
      "theoretical_loss": 3.3418309757112494,
      "tokens_seen": 2764963840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000164989898989899,
      "loss": 2.6592,
      "theoretical_loss": 3.3418247842440976,
      "tokens_seen": 2765029376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016496969696969696,
      "loss": 2.7491,
      "theoretical_loss": 3.341818592964781,
      "tokens_seen": 2765094912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016494949494949497,
      "loss": 2.3027,
      "theoretical_loss": 3.341812401873289,
      "tokens_seen": 2765160448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016492929292929293,
      "loss": 2.5257,
      "theoretical_loss": 3.341806210969612,
      "tokens_seen": 2765225984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001649090909090909,
      "loss": 2.3164,
      "theoretical_loss": 3.34180002025374,
      "tokens_seen": 2765291520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001648888888888889,
      "loss": 2.6219,
      "theoretical_loss": 3.341793829725663,
      "tokens_seen": 2765357056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016486868686868685,
      "loss": 2.6914,
      "theoretical_loss": 3.34178763938537,
      "tokens_seen": 2765422592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016484848484848487,
      "loss": 2.5494,
      "theoretical_loss": 3.341781449232852,
      "tokens_seen": 2765488128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016482828282828282,
      "loss": 2.6226,
      "theoretical_loss": 3.341775259268098,
      "tokens_seen": 2765553664
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016480808080808083,
      "loss": 2.5238,
      "theoretical_loss": 3.3417690694910975,
      "tokens_seen": 2765619200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001647878787878788,
      "loss": 2.291,
      "theoretical_loss": 3.3417628799018413,
      "tokens_seen": 2765684736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016476767676767677,
      "loss": 2.5589,
      "theoretical_loss": 3.3417566905003193,
      "tokens_seen": 2765750272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016474747474747476,
      "loss": 2.4972,
      "theoretical_loss": 3.3417505012865205,
      "tokens_seen": 2765815808
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1516731,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.072603702545166,
      "objective/train/theoretical_loss": 3.3417443122604356,
      "objective/train/tokens_used": 1124740576,
      "theoretical_loss": 3.3417443122604356,
      "tokens_seen": 2765881344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016472727272727271,
      "loss": 2.5783,
      "theoretical_loss": 3.3417443122604356,
      "tokens_seen": 2765881344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016470707070707073,
      "loss": 2.5429,
      "theoretical_loss": 3.3417381234220533,
      "tokens_seen": 2765946880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016468686868686868,
      "loss": 2.6406,
      "theoretical_loss": 3.341731934771365,
      "tokens_seen": 2766012416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016466666666666667,
      "loss": 2.4154,
      "theoretical_loss": 3.3417257463083594,
      "tokens_seen": 2766077952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016464646464646465,
      "loss": 2.4299,
      "theoretical_loss": 3.3417195580330272,
      "tokens_seen": 2766143488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016462626262626263,
      "loss": 2.4882,
      "theoretical_loss": 3.3417133699453574,
      "tokens_seen": 2766209024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016460606060606062,
      "loss": 2.6944,
      "theoretical_loss": 3.34170718204534,
      "tokens_seen": 2766274560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016458585858585858,
      "loss": 2.5948,
      "theoretical_loss": 3.341700994332966,
      "tokens_seen": 2766340096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016456565656565656,
      "loss": 2.5758,
      "theoretical_loss": 3.341694806808224,
      "tokens_seen": 2766405632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016454545454545454,
      "loss": 2.4174,
      "theoretical_loss": 3.341688619471104,
      "tokens_seen": 2766471168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016452525252525253,
      "loss": 2.4039,
      "theoretical_loss": 3.341682432321597,
      "tokens_seen": 2766536704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001645050505050505,
      "loss": 2.2379,
      "theoretical_loss": 3.3416762453596913,
      "tokens_seen": 2766602240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001644848484848485,
      "loss": 2.605,
      "theoretical_loss": 3.341670058585377,
      "tokens_seen": 2766667776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016446464646464645,
      "loss": 2.6538,
      "theoretical_loss": 3.3416638719986453,
      "tokens_seen": 2766733312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016444444444444446,
      "loss": 2.69,
      "theoretical_loss": 3.3416576855994853,
      "tokens_seen": 2766798848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016442424242424242,
      "loss": 2.8116,
      "theoretical_loss": 3.3416514993878863,
      "tokens_seen": 2766864384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001644040404040404,
      "loss": 2.4402,
      "theoretical_loss": 3.3416453133638386,
      "tokens_seen": 2766929920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001643838383838384,
      "loss": 2.383,
      "theoretical_loss": 3.3416391275273325,
      "tokens_seen": 2766995456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016436363636363634,
      "loss": 2.5483,
      "theoretical_loss": 3.3416329418783572,
      "tokens_seen": 2767060992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016434343434343436,
      "loss": 2.5297,
      "theoretical_loss": 3.341626756416903,
      "tokens_seen": 2767126528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001643232323232323,
      "loss": 2.5741,
      "theoretical_loss": 3.34162057114296,
      "tokens_seen": 2767192064
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016430303030303032,
      "loss": 2.4842,
      "theoretical_loss": 3.3416143860565173,
      "tokens_seen": 2767257600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016428282828282828,
      "loss": 2.4207,
      "theoretical_loss": 3.341608201157565,
      "tokens_seen": 2767323136
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016426262626262626,
      "loss": 2.5274,
      "theoretical_loss": 3.341602016446094,
      "tokens_seen": 2767388672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016424242424242425,
      "loss": 2.4424,
      "theoretical_loss": 3.3415958319220924,
      "tokens_seen": 2767454208
    },
    {
      "epoch": 0.67,
      "objective/train/docs_used": 1517810,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.459321975708008,
      "objective/train/theoretical_loss": 3.3415896475855513,
      "objective/train/tokens_used": 1126378976,
      "theoretical_loss": 3.3415896475855513,
      "tokens_seen": 2767519744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001642222222222222,
      "loss": 2.6265,
      "theoretical_loss": 3.3415896475855513,
      "tokens_seen": 2767519744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016420202020202022,
      "loss": 2.5564,
      "theoretical_loss": 3.3415834634364607,
      "tokens_seen": 2767585280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016418181818181817,
      "loss": 2.4528,
      "theoretical_loss": 3.34157727947481,
      "tokens_seen": 2767650816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016416161616161618,
      "loss": 2.5266,
      "theoretical_loss": 3.341571095700589,
      "tokens_seen": 2767716352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016414141414141414,
      "loss": 2.5455,
      "theoretical_loss": 3.341564912113788,
      "tokens_seen": 2767781888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016412121212121213,
      "loss": 2.5253,
      "theoretical_loss": 3.3415587287143964,
      "tokens_seen": 2767847424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001641010101010101,
      "loss": 2.6848,
      "theoretical_loss": 3.341552545502404,
      "tokens_seen": 2767912960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001640808080808081,
      "loss": 2.5037,
      "theoretical_loss": 3.3415463624778017,
      "tokens_seen": 2767978496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016406060606060608,
      "loss": 2.5202,
      "theoretical_loss": 3.341540179640578,
      "tokens_seen": 2768044032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016404040404040403,
      "loss": 2.5081,
      "theoretical_loss": 3.341533996990724,
      "tokens_seen": 2768109568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016402020202020202,
      "loss": 2.4229,
      "theoretical_loss": 3.341527814528229,
      "tokens_seen": 2768175104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000164,
      "loss": 2.5212,
      "theoretical_loss": 3.3415216322530825,
      "tokens_seen": 2768240640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016397979797979799,
      "loss": 2.3994,
      "theoretical_loss": 3.341515450165275,
      "tokens_seen": 2768306176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016395959595959597,
      "loss": 2.5431,
      "theoretical_loss": 3.3415092682647964,
      "tokens_seen": 2768371712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016393939393939395,
      "loss": 2.5385,
      "theoretical_loss": 3.3415030865516364,
      "tokens_seen": 2768437248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001639191919191919,
      "loss": 2.497,
      "theoretical_loss": 3.3414969050257852,
      "tokens_seen": 2768502784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016389898989898992,
      "loss": 2.5348,
      "theoretical_loss": 3.341490723687232,
      "tokens_seen": 2768568320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016387878787878788,
      "loss": 2.6457,
      "theoretical_loss": 3.341484542535967,
      "tokens_seen": 2768633856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016385858585858586,
      "loss": 2.5542,
      "theoretical_loss": 3.3414783615719807,
      "tokens_seen": 2768699392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016383838383838385,
      "loss": 2.4564,
      "theoretical_loss": 3.341472180795262,
      "tokens_seen": 2768764928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001638181818181818,
      "loss": 2.3231,
      "theoretical_loss": 3.341466000205801,
      "tokens_seen": 2768830464
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016379797979797981,
      "loss": 2.6891,
      "theoretical_loss": 3.3414598198035885,
      "tokens_seen": 2768896000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016377777777777777,
      "loss": 2.6295,
      "theoretical_loss": 3.3414536395886136,
      "tokens_seen": 2768961536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016375757575757578,
      "loss": 2.6315,
      "theoretical_loss": 3.3414474595608663,
      "tokens_seen": 2769027072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016373737373737374,
      "loss": 2.4431,
      "theoretical_loss": 3.3414412797203363,
      "tokens_seen": 2769092608
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1518443,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.257474660873413,
      "objective/train/theoretical_loss": 3.341435100067014,
      "objective/train/tokens_used": 1128017376,
      "theoretical_loss": 3.341435100067014,
      "tokens_seen": 2769158144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001637171717171717,
      "loss": 2.4977,
      "theoretical_loss": 3.341435100067014,
      "tokens_seen": 2769158144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001636969696969697,
      "loss": 2.6552,
      "theoretical_loss": 3.3414289206008885,
      "tokens_seen": 2769223680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016367676767676766,
      "loss": 2.2741,
      "theoretical_loss": 3.341422741321951,
      "tokens_seen": 2769289216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016365656565656568,
      "loss": 2.4318,
      "theoretical_loss": 3.34141656223019,
      "tokens_seen": 2769354752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016363636363636363,
      "loss": 2.5306,
      "theoretical_loss": 3.341410383325597,
      "tokens_seen": 2769420288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016361616161616162,
      "loss": 2.3239,
      "theoretical_loss": 3.34140420460816,
      "tokens_seen": 2769485824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001635959595959596,
      "loss": 2.6397,
      "theoretical_loss": 3.34139802607787,
      "tokens_seen": 2769551360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016357575757575758,
      "loss": 2.5651,
      "theoretical_loss": 3.341391847734717,
      "tokens_seen": 2769616896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016355555555555557,
      "loss": 2.5907,
      "theoretical_loss": 3.3413856695786905,
      "tokens_seen": 2769682432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016353535353535352,
      "loss": 2.3563,
      "theoretical_loss": 3.3413794916097808,
      "tokens_seen": 2769747968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001635151515151515,
      "loss": 2.4961,
      "theoretical_loss": 3.341373313827977,
      "tokens_seen": 2769813504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001634949494949495,
      "loss": 2.501,
      "theoretical_loss": 3.34136713623327,
      "tokens_seen": 2769879040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016347474747474748,
      "loss": 2.2876,
      "theoretical_loss": 3.3413609588256494,
      "tokens_seen": 2769944576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016345454545454546,
      "loss": 2.6528,
      "theoretical_loss": 3.341354781605105,
      "tokens_seen": 2770010112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016343434343434344,
      "loss": 2.4382,
      "theoretical_loss": 3.341348604571626,
      "tokens_seen": 2770075648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001634141414141414,
      "loss": 2.7224,
      "theoretical_loss": 3.3413424277252037,
      "tokens_seen": 2770141184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001633939393939394,
      "loss": 2.5647,
      "theoretical_loss": 3.3413362510658273,
      "tokens_seen": 2770206720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016337373737373737,
      "loss": 2.5841,
      "theoretical_loss": 3.341330074593486,
      "tokens_seen": 2770272256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016335353535353535,
      "loss": 2.6594,
      "theoretical_loss": 3.3413238983081714,
      "tokens_seen": 2770337792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016333333333333334,
      "loss": 2.5856,
      "theoretical_loss": 3.341317722209872,
      "tokens_seen": 2770403328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016331313131313132,
      "loss": 2.5069,
      "theoretical_loss": 3.341311546298578,
      "tokens_seen": 2770468864
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001632929292929293,
      "loss": 2.6656,
      "theoretical_loss": 3.34130537057428,
      "tokens_seen": 2770534400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016327272727272726,
      "loss": 2.6297,
      "theoretical_loss": 3.341299195036967,
      "tokens_seen": 2770599936
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016325252525252527,
      "loss": 2.7451,
      "theoretical_loss": 3.3412930196866295,
      "tokens_seen": 2770665472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016323232323232323,
      "loss": 2.5948,
      "theoretical_loss": 3.341286844523257,
      "tokens_seen": 2770731008
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1519063,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.488872528076172,
      "objective/train/theoretical_loss": 3.3412806695468396,
      "objective/train/tokens_used": 1129655776,
      "theoretical_loss": 3.3412806695468396,
      "tokens_seen": 2770796544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016321212121212124,
      "loss": 2.547,
      "theoretical_loss": 3.3412806695468396,
      "tokens_seen": 2770796544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001631919191919192,
      "loss": 2.6067,
      "theoretical_loss": 3.3412744947573674,
      "tokens_seen": 2770862080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016317171717171715,
      "loss": 2.4995,
      "theoretical_loss": 3.34126832015483,
      "tokens_seen": 2770927616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016315151515151517,
      "loss": 2.5637,
      "theoretical_loss": 3.341262145739218,
      "tokens_seen": 2770993152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016313131313131312,
      "loss": 2.6142,
      "theoretical_loss": 3.3412559715105203,
      "tokens_seen": 2771058688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016311111111111113,
      "loss": 2.4756,
      "theoretical_loss": 3.3412497974687274,
      "tokens_seen": 2771124224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001630909090909091,
      "loss": 2.3863,
      "theoretical_loss": 3.3412436236138294,
      "tokens_seen": 2771189760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016307070707070707,
      "loss": 2.3899,
      "theoretical_loss": 3.3412374499458157,
      "tokens_seen": 2771255296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016305050505050506,
      "loss": 2.64,
      "theoretical_loss": 3.3412312764646765,
      "tokens_seen": 2771320832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016303030303030304,
      "loss": 2.3047,
      "theoretical_loss": 3.3412251031704026,
      "tokens_seen": 2771386368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016301010101010103,
      "loss": 2.431,
      "theoretical_loss": 3.3412189300629818,
      "tokens_seen": 2771451904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016298989898989898,
      "loss": 2.5177,
      "theoretical_loss": 3.341212757142406,
      "tokens_seen": 2771517440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016296969696969697,
      "loss": 2.6589,
      "theoretical_loss": 3.3412065844086642,
      "tokens_seen": 2771582976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016294949494949495,
      "loss": 2.618,
      "theoretical_loss": 3.3412004118617467,
      "tokens_seen": 2771648512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016292929292929294,
      "loss": 2.543,
      "theoretical_loss": 3.341194239501643,
      "tokens_seen": 2771714048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016290909090909092,
      "loss": 2.6723,
      "theoretical_loss": 3.3411880673283436,
      "tokens_seen": 2771779584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001628888888888889,
      "loss": 2.6841,
      "theoretical_loss": 3.341181895341838,
      "tokens_seen": 2771845120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016286868686868686,
      "loss": 2.5804,
      "theoretical_loss": 3.3411757235421162,
      "tokens_seen": 2771910656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016284848484848484,
      "loss": 2.3863,
      "theoretical_loss": 3.3411695519291684,
      "tokens_seen": 2771976192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016282828282828283,
      "loss": 2.529,
      "theoretical_loss": 3.341163380502984,
      "tokens_seen": 2772041728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001628080808080808,
      "loss": 2.6733,
      "theoretical_loss": 3.3411572092635535,
      "tokens_seen": 2772107264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001627878787878788,
      "loss": 2.6142,
      "theoretical_loss": 3.3411510382108665,
      "tokens_seen": 2772172800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016276767676767675,
      "loss": 2.6541,
      "theoretical_loss": 3.341144867344913,
      "tokens_seen": 2772238336
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016274747474747476,
      "loss": 2.3849,
      "theoretical_loss": 3.3411386966656833,
      "tokens_seen": 2772303872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016272727272727272,
      "loss": 2.7101,
      "theoretical_loss": 3.3411325261731664,
      "tokens_seen": 2772369408
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1519644,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.670851945877075,
      "objective/train/theoretical_loss": 3.341126355867353,
      "objective/train/tokens_used": 1131294176,
      "theoretical_loss": 3.341126355867353,
      "tokens_seen": 2772434944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016270707070707073,
      "loss": 2.6129,
      "theoretical_loss": 3.341126355867353,
      "tokens_seen": 2772434944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001626868686868687,
      "loss": 2.63,
      "theoretical_loss": 3.341120185748233,
      "tokens_seen": 2772500480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016266666666666665,
      "loss": 2.8461,
      "theoretical_loss": 3.341114015815796,
      "tokens_seen": 2772566016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016264646464646466,
      "loss": 2.7287,
      "theoretical_loss": 3.3411078460700327,
      "tokens_seen": 2772631552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001626262626262626,
      "loss": 2.5764,
      "theoretical_loss": 3.341101676510932,
      "tokens_seen": 2772697088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016260606060606062,
      "loss": 2.5869,
      "theoretical_loss": 3.3410955071384842,
      "tokens_seen": 2772762624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016258585858585858,
      "loss": 2.5042,
      "theoretical_loss": 3.3410893379526794,
      "tokens_seen": 2772828160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016256565656565656,
      "loss": 2.6667,
      "theoretical_loss": 3.341083168953508,
      "tokens_seen": 2772893696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016254545454545455,
      "loss": 2.2917,
      "theoretical_loss": 3.341077000140959,
      "tokens_seen": 2772959232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016252525252525253,
      "loss": 2.5474,
      "theoretical_loss": 3.341070831515023,
      "tokens_seen": 2773024768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016250505050505052,
      "loss": 2.5906,
      "theoretical_loss": 3.3410646630756897,
      "tokens_seen": 2773090304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016248484848484847,
      "loss": 2.651,
      "theoretical_loss": 3.341058494822949,
      "tokens_seen": 2773155840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016246464646464648,
      "loss": 2.4476,
      "theoretical_loss": 3.341052326756791,
      "tokens_seen": 2773221376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016244444444444444,
      "loss": 2.5894,
      "theoretical_loss": 3.3410461588772056,
      "tokens_seen": 2773286912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016242424242424243,
      "loss": 2.5081,
      "theoretical_loss": 3.341039991184183,
      "tokens_seen": 2773352448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001624040404040404,
      "loss": 2.4556,
      "theoretical_loss": 3.341033823677712,
      "tokens_seen": 2773417984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001623838383838384,
      "loss": 2.5536,
      "theoretical_loss": 3.3410276563577845,
      "tokens_seen": 2773483520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016236363636363638,
      "loss": 2.6117,
      "theoretical_loss": 3.3410214892243886,
      "tokens_seen": 2773549056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016234343434343436,
      "loss": 2.6355,
      "theoretical_loss": 3.3410153222775154,
      "tokens_seen": 2773614592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016232323232323232,
      "loss": 2.4232,
      "theoretical_loss": 3.3410091555171544,
      "tokens_seen": 2773680128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001623030303030303,
      "loss": 2.5252,
      "theoretical_loss": 3.3410029889432957,
      "tokens_seen": 2773745664
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016228282828282829,
      "loss": 2.5139,
      "theoretical_loss": 3.3409968225559292,
      "tokens_seen": 2773811200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016226262626262627,
      "loss": 2.6772,
      "theoretical_loss": 3.3409906563550447,
      "tokens_seen": 2773876736
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016224242424242425,
      "loss": 2.3691,
      "theoretical_loss": 3.3409844903406323,
      "tokens_seen": 2773942272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001622222222222222,
      "loss": 2.5815,
      "theoretical_loss": 3.340978324512682,
      "tokens_seen": 2774007808
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1520668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0536701679229736,
      "objective/train/theoretical_loss": 3.3409721588711836,
      "objective/train/tokens_used": 1132932576,
      "theoretical_loss": 3.3409721588711836,
      "tokens_seen": 2774073344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016220202020202022,
      "loss": 2.8458,
      "theoretical_loss": 3.3409721588711836,
      "tokens_seen": 2774073344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016218181818181818,
      "loss": 2.7143,
      "theoretical_loss": 3.3409659934161273,
      "tokens_seen": 2774138880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001621616161616162,
      "loss": 2.5732,
      "theoretical_loss": 3.340959828147503,
      "tokens_seen": 2774204416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016214141414141415,
      "loss": 2.7027,
      "theoretical_loss": 3.3409536630653003,
      "tokens_seen": 2774269952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001621212121212121,
      "loss": 2.4665,
      "theoretical_loss": 3.3409474981695095,
      "tokens_seen": 2774335488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016210101010101011,
      "loss": 2.4467,
      "theoretical_loss": 3.340941333460121,
      "tokens_seen": 2774401024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016208080808080807,
      "loss": 2.326,
      "theoretical_loss": 3.3409351689371234,
      "tokens_seen": 2774466560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016206060606060608,
      "loss": 2.5686,
      "theoretical_loss": 3.340929004600508,
      "tokens_seen": 2774532096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016204040404040404,
      "loss": 2.549,
      "theoretical_loss": 3.340922840450264,
      "tokens_seen": 2774597632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016202020202020202,
      "loss": 2.3004,
      "theoretical_loss": 3.340916676486382,
      "tokens_seen": 2774663168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000162,
      "loss": 2.6201,
      "theoretical_loss": 3.3409105127088514,
      "tokens_seen": 2774728704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016197979797979796,
      "loss": 2.7913,
      "theoretical_loss": 3.340904349117663,
      "tokens_seen": 2774794240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016195959595959598,
      "loss": 2.3966,
      "theoretical_loss": 3.3408981857128053,
      "tokens_seen": 2774859776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016193939393939393,
      "loss": 2.5628,
      "theoretical_loss": 3.3408920224942693,
      "tokens_seen": 2774925312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016191919191919192,
      "loss": 2.6146,
      "theoretical_loss": 3.3408858594620447,
      "tokens_seen": 2774990848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001618989898989899,
      "loss": 2.554,
      "theoretical_loss": 3.3408796966161223,
      "tokens_seen": 2775056384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016187878787878788,
      "loss": 2.4501,
      "theoretical_loss": 3.3408735339564903,
      "tokens_seen": 2775121920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016185858585858587,
      "loss": 2.5027,
      "theoretical_loss": 3.3408673714831405,
      "tokens_seen": 2775187456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016183838383838385,
      "loss": 2.6133,
      "theoretical_loss": 3.3408612091960617,
      "tokens_seen": 2775252992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001618181818181818,
      "loss": 2.4547,
      "theoretical_loss": 3.340855047095244,
      "tokens_seen": 2775318528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001617979797979798,
      "loss": 2.2018,
      "theoretical_loss": 3.340848885180678,
      "tokens_seen": 2775384064
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016177777777777778,
      "loss": 2.567,
      "theoretical_loss": 3.340842723452353,
      "tokens_seen": 2775449600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016175757575757576,
      "loss": 2.5154,
      "theoretical_loss": 3.340836561910259,
      "tokens_seen": 2775515136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016173737373737374,
      "loss": 2.4335,
      "theoretical_loss": 3.340830400554387,
      "tokens_seen": 2775580672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001617171717171717,
      "loss": 2.6162,
      "theoretical_loss": 3.340824239384726,
      "tokens_seen": 2775646208
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1521377,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4955875873565674,
      "objective/train/theoretical_loss": 3.3408180784012655,
      "objective/train/tokens_used": 1134570976,
      "theoretical_loss": 3.3408180784012655,
      "tokens_seen": 2775711744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001616969696969697,
      "loss": 2.6075,
      "theoretical_loss": 3.3408180784012655,
      "tokens_seen": 2775711744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016167676767676767,
      "loss": 2.5113,
      "theoretical_loss": 3.3408119176039963,
      "tokens_seen": 2775777280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016165656565656568,
      "loss": 2.5171,
      "theoretical_loss": 3.340805756992909,
      "tokens_seen": 2775842816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016163636363636364,
      "loss": 2.3704,
      "theoretical_loss": 3.340799596567992,
      "tokens_seen": 2775908352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016161616161616162,
      "loss": 2.3761,
      "theoretical_loss": 3.3407934363292364,
      "tokens_seen": 2775973888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001615959595959596,
      "loss": 2.4156,
      "theoretical_loss": 3.3407872762766315,
      "tokens_seen": 2776039424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016157575757575756,
      "loss": 2.4299,
      "theoretical_loss": 3.340781116410168,
      "tokens_seen": 2776104960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016155555555555557,
      "loss": 2.4115,
      "theoretical_loss": 3.3407749567298355,
      "tokens_seen": 2776170496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016153535353535353,
      "loss": 2.2878,
      "theoretical_loss": 3.340768797235624,
      "tokens_seen": 2776236032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016151515151515154,
      "loss": 2.4185,
      "theoretical_loss": 3.3407626379275235,
      "tokens_seen": 2776301568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001614949494949495,
      "loss": 2.601,
      "theoretical_loss": 3.340756478805524,
      "tokens_seen": 2776367104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016147474747474748,
      "loss": 2.4135,
      "theoretical_loss": 3.340750319869615,
      "tokens_seen": 2776432640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016145454545454547,
      "loss": 2.4905,
      "theoretical_loss": 3.3407441611197872,
      "tokens_seen": 2776498176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016143434343434342,
      "loss": 2.6558,
      "theoretical_loss": 3.3407380025560305,
      "tokens_seen": 2776563712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016141414141414143,
      "loss": 2.7102,
      "theoretical_loss": 3.3407318441783347,
      "tokens_seen": 2776629248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001613939393939394,
      "loss": 2.383,
      "theoretical_loss": 3.3407256859866896,
      "tokens_seen": 2776694784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016137373737373737,
      "loss": 2.2967,
      "theoretical_loss": 3.3407195279810855,
      "tokens_seen": 2776760320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016135353535353536,
      "loss": 2.6323,
      "theoretical_loss": 3.340713370161512,
      "tokens_seen": 2776825856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016133333333333334,
      "loss": 2.6789,
      "theoretical_loss": 3.3407072125279598,
      "tokens_seen": 2776891392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016131313131313133,
      "loss": 2.4848,
      "theoretical_loss": 3.3407010550804186,
      "tokens_seen": 2776956928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016129292929292928,
      "loss": 2.6104,
      "theoretical_loss": 3.340694897818878,
      "tokens_seen": 2777022464
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016127272727272727,
      "loss": 2.5611,
      "theoretical_loss": 3.340688740743328,
      "tokens_seen": 2777088000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016125252525252525,
      "loss": 2.5374,
      "theoretical_loss": 3.340682583853759,
      "tokens_seen": 2777153536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016123232323232324,
      "loss": 2.3621,
      "theoretical_loss": 3.3406764271501608,
      "tokens_seen": 2777219072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016121212121212122,
      "loss": 2.4859,
      "theoretical_loss": 3.3406702706325233,
      "tokens_seen": 2777284608
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1522539,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6646552085876465,
      "objective/train/theoretical_loss": 3.3406641143008367,
      "objective/train/tokens_used": 1136209376,
      "theoretical_loss": 3.3406641143008367,
      "tokens_seen": 2777350144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001611919191919192,
      "loss": 2.5536,
      "theoretical_loss": 3.3406641143008367,
      "tokens_seen": 2777350144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016117171717171716,
      "loss": 2.4847,
      "theoretical_loss": 3.340657958155091,
      "tokens_seen": 2777415680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016115151515151517,
      "loss": 2.5023,
      "theoretical_loss": 3.3406518021952762,
      "tokens_seen": 2777481216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016113131313131313,
      "loss": 2.723,
      "theoretical_loss": 3.340645646421382,
      "tokens_seen": 2777546752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001611111111111111,
      "loss": 2.3839,
      "theoretical_loss": 3.3406394908333987,
      "tokens_seen": 2777612288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001610909090909091,
      "loss": 2.5015,
      "theoretical_loss": 3.340633335431316,
      "tokens_seen": 2777677824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016107070707070705,
      "loss": 2.6068,
      "theoretical_loss": 3.340627180215124,
      "tokens_seen": 2777743360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016105050505050506,
      "loss": 2.6152,
      "theoretical_loss": 3.3406210251848134,
      "tokens_seen": 2777808896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016103030303030302,
      "loss": 2.6447,
      "theoretical_loss": 3.3406148703403735,
      "tokens_seen": 2777874432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016101010101010103,
      "loss": 2.7729,
      "theoretical_loss": 3.3406087156817943,
      "tokens_seen": 2777939968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160989898989899,
      "loss": 2.3859,
      "theoretical_loss": 3.3406025612090655,
      "tokens_seen": 2778005504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016096969696969697,
      "loss": 2.4838,
      "theoretical_loss": 3.340596406922178,
      "tokens_seen": 2778071040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016094949494949496,
      "loss": 2.4317,
      "theoretical_loss": 3.340590252821121,
      "tokens_seen": 2778136576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001609292929292929,
      "loss": 2.475,
      "theoretical_loss": 3.340584098905885,
      "tokens_seen": 2778202112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016090909090909092,
      "loss": 2.2991,
      "theoretical_loss": 3.3405779451764595,
      "tokens_seen": 2778267648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016088888888888888,
      "loss": 2.6116,
      "theoretical_loss": 3.3405717916328355,
      "tokens_seen": 2778333184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016086868686868687,
      "loss": 2.597,
      "theoretical_loss": 3.3405656382750015,
      "tokens_seen": 2778398720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016084848484848485,
      "loss": 2.5592,
      "theoretical_loss": 3.3405594851029488,
      "tokens_seen": 2778464256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016082828282828283,
      "loss": 2.4637,
      "theoretical_loss": 3.3405533321166674,
      "tokens_seen": 2778529792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016080808080808082,
      "loss": 2.4627,
      "theoretical_loss": 3.3405471793161463,
      "tokens_seen": 2778595328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001607878787878788,
      "loss": 2.4488,
      "theoretical_loss": 3.340541026701376,
      "tokens_seen": 2778660864
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016076767676767679,
      "loss": 2.5985,
      "theoretical_loss": 3.340534874272347,
      "tokens_seen": 2778726400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016074747474747474,
      "loss": 2.5007,
      "theoretical_loss": 3.3405287220290485,
      "tokens_seen": 2778791936
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016072727272727273,
      "loss": 2.3902,
      "theoretical_loss": 3.3405225699714713,
      "tokens_seen": 2778857472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001607070707070707,
      "loss": 2.6528,
      "theoretical_loss": 3.3405164180996048,
      "tokens_seen": 2778923008
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1523044,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.768937587738037,
      "objective/train/theoretical_loss": 3.340510266413439,
      "objective/train/tokens_used": 1137847776,
      "theoretical_loss": 3.340510266413439,
      "tokens_seen": 2778988544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001606868686868687,
      "loss": 2.6315,
      "theoretical_loss": 3.340510266413439,
      "tokens_seen": 2778988544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016066666666666668,
      "loss": 2.4005,
      "theoretical_loss": 3.3405041149129646,
      "tokens_seen": 2779054080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016064646464646466,
      "loss": 2.6206,
      "theoretical_loss": 3.340497963598171,
      "tokens_seen": 2779119616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016062626262626262,
      "loss": 2.4944,
      "theoretical_loss": 3.3404918124690486,
      "tokens_seen": 2779185152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016060606060606063,
      "loss": 2.5336,
      "theoretical_loss": 3.340485661525587,
      "tokens_seen": 2779250688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001605858585858586,
      "loss": 2.4623,
      "theoretical_loss": 3.3404795107677763,
      "tokens_seen": 2779316224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016056565656565657,
      "loss": 2.4734,
      "theoretical_loss": 3.340473360195607,
      "tokens_seen": 2779381760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016054545454545455,
      "loss": 2.6917,
      "theoretical_loss": 3.340467209809068,
      "tokens_seen": 2779447296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001605252525252525,
      "loss": 2.6046,
      "theoretical_loss": 3.3404610596081508,
      "tokens_seen": 2779512832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016050505050505052,
      "loss": 2.5243,
      "theoretical_loss": 3.340454909592845,
      "tokens_seen": 2779578368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016048484848484848,
      "loss": 2.5576,
      "theoretical_loss": 3.34044875976314,
      "tokens_seen": 2779643904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001604646464646465,
      "loss": 2.3988,
      "theoretical_loss": 3.340442610119026,
      "tokens_seen": 2779709440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016044444444444445,
      "loss": 2.6069,
      "theoretical_loss": 3.340436460660493,
      "tokens_seen": 2779774976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001604242424242424,
      "loss": 2.463,
      "theoretical_loss": 3.3404303113875313,
      "tokens_seen": 2779840512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016040404040404042,
      "loss": 2.4132,
      "theoretical_loss": 3.340424162300131,
      "tokens_seen": 2779906048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016038383838383837,
      "loss": 2.3336,
      "theoretical_loss": 3.340418013398282,
      "tokens_seen": 2779971584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016036363636363638,
      "loss": 2.483,
      "theoretical_loss": 3.340411864681974,
      "tokens_seen": 2780037120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016034343434343434,
      "loss": 2.5414,
      "theoretical_loss": 3.3404057161511975,
      "tokens_seen": 2780102656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016032323232323232,
      "loss": 2.5594,
      "theoretical_loss": 3.3403995678059424,
      "tokens_seen": 2780168192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001603030303030303,
      "loss": 2.6549,
      "theoretical_loss": 3.340393419646199,
      "tokens_seen": 2780233728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001602828282828283,
      "loss": 2.7012,
      "theoretical_loss": 3.3403872716719563,
      "tokens_seen": 2780299264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016026262626262628,
      "loss": 2.8683,
      "theoretical_loss": 3.3403811238832053,
      "tokens_seen": 2780364800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016024242424242423,
      "loss": 2.6444,
      "theoretical_loss": 3.3403749762799357,
      "tokens_seen": 2780430336
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016022222222222222,
      "loss": 2.652,
      "theoretical_loss": 3.340368828862138,
      "tokens_seen": 2780495872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001602020202020202,
      "loss": 2.4963,
      "theoretical_loss": 3.3403626816298013,
      "tokens_seen": 2780561408
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1524063,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6973204612731934,
      "objective/train/theoretical_loss": 3.3403565345829165,
      "objective/train/tokens_used": 1139486176,
      "theoretical_loss": 3.3403565345829165,
      "tokens_seen": 2780626944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016018181818181818,
      "loss": 2.5187,
      "theoretical_loss": 3.3403565345829165,
      "tokens_seen": 2780626944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016016161616161617,
      "loss": 2.5263,
      "theoretical_loss": 3.340350387721473,
      "tokens_seen": 2780692480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016014141414141415,
      "loss": 2.6961,
      "theoretical_loss": 3.3403442410454613,
      "tokens_seen": 2780758016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001601212121212121,
      "loss": 2.5047,
      "theoretical_loss": 3.3403380945548715,
      "tokens_seen": 2780823552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016010101010101012,
      "loss": 2.5578,
      "theoretical_loss": 3.340331948249693,
      "tokens_seen": 2780889088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016008080808080808,
      "loss": 2.6112,
      "theoretical_loss": 3.3403258021299163,
      "tokens_seen": 2780954624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016006060606060606,
      "loss": 2.7148,
      "theoretical_loss": 3.3403196561955317,
      "tokens_seen": 2781020160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016004040404040405,
      "loss": 2.5043,
      "theoretical_loss": 3.3403135104465287,
      "tokens_seen": 2781085696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160020202020202,
      "loss": 2.4761,
      "theoretical_loss": 3.3403073648828974,
      "tokens_seen": 2781151232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016,
      "loss": 2.3614,
      "theoretical_loss": 3.340301219504628,
      "tokens_seen": 2781216768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015997979797979797,
      "loss": 2.5628,
      "theoretical_loss": 3.340295074311711,
      "tokens_seen": 2781282304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015995959595959598,
      "loss": 2.5824,
      "theoretical_loss": 3.340288929304135,
      "tokens_seen": 2781347840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015993939393939394,
      "loss": 2.4464,
      "theoretical_loss": 3.340282784481892,
      "tokens_seen": 2781413376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015991919191919192,
      "loss": 2.4653,
      "theoretical_loss": 3.340276639844971,
      "tokens_seen": 2781478912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001598989898989899,
      "loss": 2.5715,
      "theoretical_loss": 3.3402704953933617,
      "tokens_seen": 2781544448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015987878787878786,
      "loss": 2.5535,
      "theoretical_loss": 3.340264351127055,
      "tokens_seen": 2781609984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015985858585858587,
      "loss": 2.611,
      "theoretical_loss": 3.34025820704604,
      "tokens_seen": 2781675520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015983838383838383,
      "loss": 2.6541,
      "theoretical_loss": 3.3402520631503076,
      "tokens_seen": 2781741056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015981818181818184,
      "loss": 2.5694,
      "theoretical_loss": 3.3402459194398473,
      "tokens_seen": 2781806592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001597979797979798,
      "loss": 2.7202,
      "theoretical_loss": 3.3402397759146494,
      "tokens_seen": 2781872128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015977777777777778,
      "loss": 2.389,
      "theoretical_loss": 3.3402336325747037,
      "tokens_seen": 2781937664
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015975757575757577,
      "loss": 2.4132,
      "theoretical_loss": 3.340227489420001,
      "tokens_seen": 2782003200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015973737373737372,
      "loss": 2.7229,
      "theoretical_loss": 3.34022134645053,
      "tokens_seen": 2782068736
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015971717171717173,
      "loss": 2.52,
      "theoretical_loss": 3.340215203666282,
      "tokens_seen": 2782134272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001596969696969697,
      "loss": 2.5388,
      "theoretical_loss": 3.340209061067247,
      "tokens_seen": 2782199808
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1525440,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.696449041366577,
      "objective/train/theoretical_loss": 3.340202918653414,
      "objective/train/tokens_used": 1141124576,
      "theoretical_loss": 3.340202918653414,
      "tokens_seen": 2782265344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015967676767676768,
      "loss": 2.4945,
      "theoretical_loss": 3.340202918653414,
      "tokens_seen": 2782265344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015965656565656566,
      "loss": 2.4624,
      "theoretical_loss": 3.340196776424774,
      "tokens_seen": 2782330880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015963636363636364,
      "loss": 2.5951,
      "theoretical_loss": 3.3401906343813166,
      "tokens_seen": 2782396416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015961616161616163,
      "loss": 2.6573,
      "theoretical_loss": 3.3401844925230324,
      "tokens_seen": 2782461952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001595959595959596,
      "loss": 2.4068,
      "theoretical_loss": 3.3401783508499108,
      "tokens_seen": 2782527488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015957575757575757,
      "loss": 2.5661,
      "theoretical_loss": 3.340172209361942,
      "tokens_seen": 2782593024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015955555555555555,
      "loss": 2.7027,
      "theoretical_loss": 3.3401660680591165,
      "tokens_seen": 2782658560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015953535353535354,
      "loss": 2.6991,
      "theoretical_loss": 3.3401599269414235,
      "tokens_seen": 2782724096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015951515151515152,
      "loss": 2.5601,
      "theoretical_loss": 3.3401537860088544,
      "tokens_seen": 2782789632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001594949494949495,
      "loss": 2.575,
      "theoretical_loss": 3.340147645261398,
      "tokens_seen": 2782855168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015947474747474746,
      "loss": 2.3982,
      "theoretical_loss": 3.340141504699045,
      "tokens_seen": 2782920704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015945454545454547,
      "loss": 2.5263,
      "theoretical_loss": 3.340135364321785,
      "tokens_seen": 2782986240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015943434343434343,
      "loss": 2.2968,
      "theoretical_loss": 3.3401292241296083,
      "tokens_seen": 2783051776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015941414141414144,
      "loss": 2.5574,
      "theoretical_loss": 3.340123084122505,
      "tokens_seen": 2783117312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001593939393939394,
      "loss": 2.6271,
      "theoretical_loss": 3.3401169443004655,
      "tokens_seen": 2783182848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015937373737373735,
      "loss": 2.7247,
      "theoretical_loss": 3.3401108046634795,
      "tokens_seen": 2783248384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015935353535353536,
      "loss": 2.3931,
      "theoretical_loss": 3.340104665211537,
      "tokens_seen": 2783313920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015933333333333332,
      "loss": 2.7295,
      "theoretical_loss": 3.3400985259446285,
      "tokens_seen": 2783379456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015931313131313133,
      "loss": 2.501,
      "theoretical_loss": 3.340092386862743,
      "tokens_seen": 2783444992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001592929292929293,
      "loss": 2.4324,
      "theoretical_loss": 3.340086247965872,
      "tokens_seen": 2783510528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015927272727272727,
      "loss": 2.421,
      "theoretical_loss": 3.340080109254005,
      "tokens_seen": 2783576064
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015925252525252526,
      "loss": 2.8095,
      "theoretical_loss": 3.3400739707271314,
      "tokens_seen": 2783641600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015923232323232324,
      "loss": 2.4783,
      "theoretical_loss": 3.340067832385242,
      "tokens_seen": 2783707136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015921212121212122,
      "loss": 2.5353,
      "theoretical_loss": 3.3400616942283268,
      "tokens_seen": 2783772672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015919191919191918,
      "loss": 2.395,
      "theoretical_loss": 3.3400555562563756,
      "tokens_seen": 2783838208
    },
    {
      "epoch": 0.68,
      "objective/train/docs_used": 1526156,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.48842716217041,
      "objective/train/theoretical_loss": 3.3400494184693788,
      "objective/train/tokens_used": 1142762976,
      "theoretical_loss": 3.3400494184693788,
      "tokens_seen": 2783903744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015917171717171717,
      "loss": 2.645,
      "theoretical_loss": 3.3400494184693788,
      "tokens_seen": 2783903744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015915151515151515,
      "loss": 2.726,
      "theoretical_loss": 3.340043280867326,
      "tokens_seen": 2783969280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015913131313131313,
      "loss": 2.6242,
      "theoretical_loss": 3.340037143450208,
      "tokens_seen": 2784034816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015911111111111112,
      "loss": 2.4032,
      "theoretical_loss": 3.340031006218014,
      "tokens_seen": 2784100352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001590909090909091,
      "loss": 2.56,
      "theoretical_loss": 3.3400248691707346,
      "tokens_seen": 2784165888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015907070707070706,
      "loss": 2.7339,
      "theoretical_loss": 3.34001873230836,
      "tokens_seen": 2784231424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015905050505050507,
      "loss": 2.6934,
      "theoretical_loss": 3.34001259563088,
      "tokens_seen": 2784296960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015903030303030303,
      "loss": 2.6446,
      "theoretical_loss": 3.3400064591382845,
      "tokens_seen": 2784362496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000159010101010101,
      "loss": 2.6008,
      "theoretical_loss": 3.340000322830564,
      "tokens_seen": 2784428032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000158989898989899,
      "loss": 2.4076,
      "theoretical_loss": 3.3399941867077088,
      "tokens_seen": 2784493568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015896969696969698,
      "loss": 2.6223,
      "theoretical_loss": 3.3399880507697084,
      "tokens_seen": 2784559104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015894949494949496,
      "loss": 2.6277,
      "theoretical_loss": 3.3399819150165526,
      "tokens_seen": 2784624640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015892929292929292,
      "loss": 2.5007,
      "theoretical_loss": 3.3399757794482325,
      "tokens_seen": 2784690176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015890909090909093,
      "loss": 2.3649,
      "theoretical_loss": 3.3399696440647375,
      "tokens_seen": 2784755712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001588888888888889,
      "loss": 2.6687,
      "theoretical_loss": 3.3399635088660578,
      "tokens_seen": 2784821248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001588686868686869,
      "loss": 2.5197,
      "theoretical_loss": 3.3399573738521835,
      "tokens_seen": 2784886784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015884848484848485,
      "loss": 2.5765,
      "theoretical_loss": 3.3399512390231045,
      "tokens_seen": 2784952320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001588282828282828,
      "loss": 2.4472,
      "theoretical_loss": 3.3399451043788115,
      "tokens_seen": 2785017856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015880808080808082,
      "loss": 2.5299,
      "theoretical_loss": 3.339938969919294,
      "tokens_seen": 2785083392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015878787878787878,
      "loss": 2.6906,
      "theoretical_loss": 3.3399328356445417,
      "tokens_seen": 2785148928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001587676767676768,
      "loss": 2.3887,
      "theoretical_loss": 3.3399267015545457,
      "tokens_seen": 2785214464
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015874747474747475,
      "loss": 2.6104,
      "theoretical_loss": 3.339920567649296,
      "tokens_seen": 2785280000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015872727272727273,
      "loss": 2.5876,
      "theoretical_loss": 3.3399144339287816,
      "tokens_seen": 2785345536
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015870707070707072,
      "loss": 2.454,
      "theoretical_loss": 3.339908300392994,
      "tokens_seen": 2785411072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015868686868686867,
      "loss": 2.5383,
      "theoretical_loss": 3.339902167041922,
      "tokens_seen": 2785476608
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1527495,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.147812843322754,
      "objective/train/theoretical_loss": 3.339896033875556,
      "objective/train/tokens_used": 1144401376,
      "theoretical_loss": 3.339896033875556,
      "tokens_seen": 2785542144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015866666666666668,
      "loss": 2.5366,
      "theoretical_loss": 3.339896033875556,
      "tokens_seen": 2785542144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015864646464646464,
      "loss": 2.5483,
      "theoretical_loss": 3.3398899008938874,
      "tokens_seen": 2785607680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015862626262626262,
      "loss": 2.4007,
      "theoretical_loss": 3.3398837680969047,
      "tokens_seen": 2785673216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001586060606060606,
      "loss": 2.5627,
      "theoretical_loss": 3.3398776354845987,
      "tokens_seen": 2785738752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001585858585858586,
      "loss": 2.5758,
      "theoretical_loss": 3.3398715030569592,
      "tokens_seen": 2785804288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015856565656565658,
      "loss": 2.533,
      "theoretical_loss": 3.3398653708139765,
      "tokens_seen": 2785869824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015854545454545456,
      "loss": 2.517,
      "theoretical_loss": 3.3398592387556407,
      "tokens_seen": 2785935360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015852525252525252,
      "loss": 2.5876,
      "theoretical_loss": 3.3398531068819417,
      "tokens_seen": 2786000896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001585050505050505,
      "loss": 2.3341,
      "theoretical_loss": 3.33984697519287,
      "tokens_seen": 2786066432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015848484848484848,
      "loss": 2.4933,
      "theoretical_loss": 3.3398408436884153,
      "tokens_seen": 2786131968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015846464646464647,
      "loss": 2.5171,
      "theoretical_loss": 3.3398347123685674,
      "tokens_seen": 2786197504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015844444444444445,
      "loss": 2.5912,
      "theoretical_loss": 3.3398285812333173,
      "tokens_seen": 2786263040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001584242424242424,
      "loss": 2.4985,
      "theoretical_loss": 3.3398224502826546,
      "tokens_seen": 2786328576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015840404040404042,
      "loss": 2.5726,
      "theoretical_loss": 3.3398163195165695,
      "tokens_seen": 2786394112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015838383838383838,
      "loss": 2.3558,
      "theoretical_loss": 3.339810188935052,
      "tokens_seen": 2786459648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001583636363636364,
      "loss": 2.3956,
      "theoretical_loss": 3.3398040585380926,
      "tokens_seen": 2786525184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015834343434343435,
      "loss": 2.5117,
      "theoretical_loss": 3.3397979283256802,
      "tokens_seen": 2786590720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001583232323232323,
      "loss": 2.4924,
      "theoretical_loss": 3.3397917982978065,
      "tokens_seen": 2786656256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001583030303030303,
      "loss": 2.5968,
      "theoretical_loss": 3.3397856684544607,
      "tokens_seen": 2786721792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015828282828282827,
      "loss": 2.5655,
      "theoretical_loss": 3.339779538795633,
      "tokens_seen": 2786787328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015826262626262628,
      "loss": 2.6844,
      "theoretical_loss": 3.3397734093213134,
      "tokens_seen": 2786852864
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015824242424242424,
      "loss": 2.4677,
      "theoretical_loss": 3.3397672800314924,
      "tokens_seen": 2786918400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015822222222222222,
      "loss": 2.6896,
      "theoretical_loss": 3.3397611509261598,
      "tokens_seen": 2786983936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001582020202020202,
      "loss": 2.5562,
      "theoretical_loss": 3.3397550220053054,
      "tokens_seen": 2787049472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001581818181818182,
      "loss": 2.5334,
      "theoretical_loss": 3.3397488932689203,
      "tokens_seen": 2787115008
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1527941,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6329824924468994,
      "objective/train/theoretical_loss": 3.339742764716994,
      "objective/train/tokens_used": 1146039776,
      "theoretical_loss": 3.339742764716994,
      "tokens_seen": 2787180544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015816161616161617,
      "loss": 2.5974,
      "theoretical_loss": 3.339742764716994,
      "tokens_seen": 2787180544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015814141414141413,
      "loss": 2.4753,
      "theoretical_loss": 3.3397366363495165,
      "tokens_seen": 2787246080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015812121212121214,
      "loss": 2.6397,
      "theoretical_loss": 3.3397305081664777,
      "tokens_seen": 2787311616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001581010101010101,
      "loss": 2.4848,
      "theoretical_loss": 3.339724380167868,
      "tokens_seen": 2787377152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015808080808080808,
      "loss": 2.4093,
      "theoretical_loss": 3.3397182523536784,
      "tokens_seen": 2787442688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015806060606060607,
      "loss": 2.4799,
      "theoretical_loss": 3.3397121247238974,
      "tokens_seen": 2787508224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015804040404040405,
      "loss": 2.6244,
      "theoretical_loss": 3.339705997278516,
      "tokens_seen": 2787573760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015802020202020203,
      "loss": 2.6077,
      "theoretical_loss": 3.3396998700175247,
      "tokens_seen": 2787639296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000158,
      "loss": 2.4154,
      "theoretical_loss": 3.339693742940913,
      "tokens_seen": 2787704832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015797979797979798,
      "loss": 2.7696,
      "theoretical_loss": 3.339687616048671,
      "tokens_seen": 2787770368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015795959595959596,
      "loss": 2.5866,
      "theoretical_loss": 3.339681489340789,
      "tokens_seen": 2787835904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015793939393939394,
      "loss": 2.5651,
      "theoretical_loss": 3.339675362817257,
      "tokens_seen": 2787901440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015791919191919193,
      "loss": 2.5116,
      "theoretical_loss": 3.339669236478065,
      "tokens_seen": 2787966976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001578989898989899,
      "loss": 2.578,
      "theoretical_loss": 3.3396631103232033,
      "tokens_seen": 2788032512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015787878787878787,
      "loss": 2.6013,
      "theoretical_loss": 3.3396569843526622,
      "tokens_seen": 2788098048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015785858585858588,
      "loss": 2.4345,
      "theoretical_loss": 3.339650858566432,
      "tokens_seen": 2788163584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015783838383838384,
      "loss": 2.5461,
      "theoretical_loss": 3.3396447329645023,
      "tokens_seen": 2788229120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015781818181818182,
      "loss": 2.4301,
      "theoretical_loss": 3.3396386075468634,
      "tokens_seen": 2788294656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001577979797979798,
      "loss": 2.5665,
      "theoretical_loss": 3.339632482313505,
      "tokens_seen": 2788360192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015777777777777776,
      "loss": 2.6078,
      "theoretical_loss": 3.339626357264418,
      "tokens_seen": 2788425728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015775757575757577,
      "loss": 2.4667,
      "theoretical_loss": 3.339620232399592,
      "tokens_seen": 2788491264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015773737373737373,
      "loss": 2.3245,
      "theoretical_loss": 3.339614107719018,
      "tokens_seen": 2788556800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015771717171717174,
      "loss": 2.7024,
      "theoretical_loss": 3.3396079832226846,
      "tokens_seen": 2788622336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001576969696969697,
      "loss": 2.3591,
      "theoretical_loss": 3.339601858910583,
      "tokens_seen": 2788687872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015767676767676768,
      "loss": 2.2236,
      "theoretical_loss": 3.339595734782703,
      "tokens_seen": 2788753408
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1529074,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.408823013305664,
      "objective/train/theoretical_loss": 3.3395896108390355,
      "objective/train/tokens_used": 1147678176,
      "theoretical_loss": 3.3395896108390355,
      "tokens_seen": 2788818944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015765656565656566,
      "loss": 2.5882,
      "theoretical_loss": 3.3395896108390355,
      "tokens_seen": 2788818944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015763636363636362,
      "loss": 2.4902,
      "theoretical_loss": 3.3395834870795698,
      "tokens_seen": 2788884480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015761616161616163,
      "loss": 2.5388,
      "theoretical_loss": 3.339577363504296,
      "tokens_seen": 2788950016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001575959595959596,
      "loss": 2.6776,
      "theoretical_loss": 3.339571240113204,
      "tokens_seen": 2789015552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015757575757575757,
      "loss": 2.5946,
      "theoretical_loss": 3.3395651169062845,
      "tokens_seen": 2789081088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015755555555555556,
      "loss": 2.6231,
      "theoretical_loss": 3.3395589938835277,
      "tokens_seen": 2789146624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015753535353535354,
      "loss": 2.6282,
      "theoretical_loss": 3.3395528710449236,
      "tokens_seen": 2789212160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015751515151515153,
      "loss": 2.602,
      "theoretical_loss": 3.339546748390462,
      "tokens_seen": 2789277696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001574949494949495,
      "loss": 2.6143,
      "theoretical_loss": 3.3395406259201335,
      "tokens_seen": 2789343232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015747474747474747,
      "loss": 2.4692,
      "theoretical_loss": 3.339534503633928,
      "tokens_seen": 2789408768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015745454545454545,
      "loss": 2.7083,
      "theoretical_loss": 3.339528381531835,
      "tokens_seen": 2789474304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015743434343434343,
      "loss": 2.4526,
      "theoretical_loss": 3.339522259613846,
      "tokens_seen": 2789539840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015741414141414142,
      "loss": 2.602,
      "theoretical_loss": 3.3395161378799507,
      "tokens_seen": 2789605376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001573939393939394,
      "loss": 2.4026,
      "theoretical_loss": 3.3395100163301383,
      "tokens_seen": 2789670912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015737373737373736,
      "loss": 2.3998,
      "theoretical_loss": 3.3395038949644,
      "tokens_seen": 2789736448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015735353535353537,
      "loss": 2.3635,
      "theoretical_loss": 3.3394977737827256,
      "tokens_seen": 2789801984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015733333333333333,
      "loss": 2.5457,
      "theoretical_loss": 3.339491652785105,
      "tokens_seen": 2789867520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015731313131313134,
      "loss": 2.4846,
      "theoretical_loss": 3.3394855319715284,
      "tokens_seen": 2789933056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001572929292929293,
      "loss": 2.5002,
      "theoretical_loss": 3.3394794113419866,
      "tokens_seen": 2789998592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015727272727272728,
      "loss": 2.5573,
      "theoretical_loss": 3.3394732908964686,
      "tokens_seen": 2790064128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015725252525252526,
      "loss": 2.4706,
      "theoretical_loss": 3.3394671706349657,
      "tokens_seen": 2790129664
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015723232323232322,
      "loss": 2.5603,
      "theoretical_loss": 3.3394610505574676,
      "tokens_seen": 2790195200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015721212121212123,
      "loss": 2.6589,
      "theoretical_loss": 3.3394549306639636,
      "tokens_seen": 2790260736
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001571919191919192,
      "loss": 2.4695,
      "theoretical_loss": 3.3394488109544453,
      "tokens_seen": 2790326272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001571717171717172,
      "loss": 2.5342,
      "theoretical_loss": 3.3394426914289017,
      "tokens_seen": 2790391808
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1530340,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0540807247161865,
      "objective/train/theoretical_loss": 3.3394365720873243,
      "objective/train/tokens_used": 1149316576,
      "theoretical_loss": 3.3394365720873243,
      "tokens_seen": 2790457344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015715151515151516,
      "loss": 2.6073,
      "theoretical_loss": 3.3394365720873243,
      "tokens_seen": 2790457344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001571313131313131,
      "loss": 2.4533,
      "theoretical_loss": 3.3394304529297014,
      "tokens_seen": 2790522880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015711111111111112,
      "loss": 2.5406,
      "theoretical_loss": 3.3394243339560243,
      "tokens_seen": 2790588416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015709090909090908,
      "loss": 2.5347,
      "theoretical_loss": 3.3394182151662832,
      "tokens_seen": 2790653952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001570707070707071,
      "loss": 2.4503,
      "theoretical_loss": 3.339412096560468,
      "tokens_seen": 2790719488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015705050505050505,
      "loss": 2.5308,
      "theoretical_loss": 3.3394059781385685,
      "tokens_seen": 2790785024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015703030303030303,
      "loss": 2.4758,
      "theoretical_loss": 3.339399859900576,
      "tokens_seen": 2790850560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015701010101010102,
      "loss": 2.5824,
      "theoretical_loss": 3.339393741846479,
      "tokens_seen": 2790916096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000156989898989899,
      "loss": 2.3199,
      "theoretical_loss": 3.339387623976269,
      "tokens_seen": 2790981632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015696969696969698,
      "loss": 2.7757,
      "theoretical_loss": 3.3393815062899357,
      "tokens_seen": 2791047168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015694949494949494,
      "loss": 2.3719,
      "theoretical_loss": 3.3393753887874693,
      "tokens_seen": 2791112704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015692929292929292,
      "loss": 2.4635,
      "theoretical_loss": 3.3393692714688594,
      "tokens_seen": 2791178240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001569090909090909,
      "loss": 2.5353,
      "theoretical_loss": 3.339363154334097,
      "tokens_seen": 2791243776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001568888888888889,
      "loss": 2.4945,
      "theoretical_loss": 3.339357037383172,
      "tokens_seen": 2791309312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015686868686868688,
      "loss": 2.5727,
      "theoretical_loss": 3.339350920616074,
      "tokens_seen": 2791374848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015684848484848486,
      "loss": 2.4194,
      "theoretical_loss": 3.339344804032794,
      "tokens_seen": 2791440384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015682828282828282,
      "loss": 2.4938,
      "theoretical_loss": 3.339338687633322,
      "tokens_seen": 2791505920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015680808080808083,
      "loss": 2.5597,
      "theoretical_loss": 3.3393325714176476,
      "tokens_seen": 2791571456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015678787878787879,
      "loss": 2.2846,
      "theoretical_loss": 3.3393264553857613,
      "tokens_seen": 2791636992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015676767676767677,
      "loss": 2.485,
      "theoretical_loss": 3.3393203395376534,
      "tokens_seen": 2791702528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015674747474747475,
      "loss": 2.6209,
      "theoretical_loss": 3.339314223873314,
      "tokens_seen": 2791768064
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001567272727272727,
      "loss": 2.6967,
      "theoretical_loss": 3.339308108392733,
      "tokens_seen": 2791833600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015670707070707072,
      "loss": 2.6042,
      "theoretical_loss": 3.339301993095901,
      "tokens_seen": 2791899136
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015668686868686868,
      "loss": 2.4045,
      "theoretical_loss": 3.3392958779828077,
      "tokens_seen": 2791964672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001566666666666667,
      "loss": 2.5268,
      "theoretical_loss": 3.339289763053444,
      "tokens_seen": 2792030208
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1530963,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8091771602630615,
      "objective/train/theoretical_loss": 3.339283648307799,
      "objective/train/tokens_used": 1150954976,
      "theoretical_loss": 3.339283648307799,
      "tokens_seen": 2792095744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015664646464646465,
      "loss": 2.4426,
      "theoretical_loss": 3.339283648307799,
      "tokens_seen": 2792095744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015662626262626263,
      "loss": 2.7038,
      "theoretical_loss": 3.3392775337458636,
      "tokens_seen": 2792161280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015660606060606061,
      "loss": 2.4854,
      "theoretical_loss": 3.339271419367628,
      "tokens_seen": 2792226816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015658585858585857,
      "loss": 2.5044,
      "theoretical_loss": 3.3392653051730816,
      "tokens_seen": 2792292352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015656565656565658,
      "loss": 2.4555,
      "theoretical_loss": 3.339259191162216,
      "tokens_seen": 2792357888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015654545454545454,
      "loss": 2.2044,
      "theoretical_loss": 3.33925307733502,
      "tokens_seen": 2792423424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015652525252525252,
      "loss": 2.4351,
      "theoretical_loss": 3.339246963691484,
      "tokens_seen": 2792488960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001565050505050505,
      "loss": 2.4081,
      "theoretical_loss": 3.339240850231599,
      "tokens_seen": 2792554496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001564848484848485,
      "loss": 2.4825,
      "theoretical_loss": 3.3392347369553548,
      "tokens_seen": 2792620032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015646464646464647,
      "loss": 2.4679,
      "theoretical_loss": 3.339228623862741,
      "tokens_seen": 2792685568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015644444444444443,
      "loss": 2.6429,
      "theoretical_loss": 3.339222510953748,
      "tokens_seen": 2792751104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015642424242424244,
      "loss": 2.5189,
      "theoretical_loss": 3.3392163982283667,
      "tokens_seen": 2792816640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001564040404040404,
      "loss": 2.6314,
      "theoretical_loss": 3.3392102856865864,
      "tokens_seen": 2792882176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015638383838383838,
      "loss": 2.5986,
      "theoretical_loss": 3.3392041733283975,
      "tokens_seen": 2792947712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015636363636363637,
      "loss": 2.3353,
      "theoretical_loss": 3.3391980611537906,
      "tokens_seen": 2793013248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015634343434343435,
      "loss": 2.5565,
      "theoretical_loss": 3.3391919491627555,
      "tokens_seen": 2793078784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015632323232323234,
      "loss": 2.4338,
      "theoretical_loss": 3.3391858373552825,
      "tokens_seen": 2793144320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015630303030303032,
      "loss": 2.3473,
      "theoretical_loss": 3.3391797257313613,
      "tokens_seen": 2793209856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015628282828282828,
      "loss": 2.6957,
      "theoretical_loss": 3.339173614290983,
      "tokens_seen": 2793275392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015626262626262626,
      "loss": 2.6184,
      "theoretical_loss": 3.3391675030341372,
      "tokens_seen": 2793340928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015624242424242424,
      "loss": 2.2866,
      "theoretical_loss": 3.339161391960814,
      "tokens_seen": 2793406464
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015622222222222223,
      "loss": 2.6464,
      "theoretical_loss": 3.3391552810710037,
      "tokens_seen": 2793472000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001562020202020202,
      "loss": 2.4466,
      "theoretical_loss": 3.339149170364697,
      "tokens_seen": 2793537536
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015618181818181817,
      "loss": 2.5138,
      "theoretical_loss": 3.3391430598418834,
      "tokens_seen": 2793603072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015616161616161618,
      "loss": 2.5405,
      "theoretical_loss": 3.339136949502553,
      "tokens_seen": 2793668608
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1531861,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0454108715057373,
      "objective/train/theoretical_loss": 3.3391308393466965,
      "objective/train/tokens_used": 1152593376,
      "theoretical_loss": 3.3391308393466965,
      "tokens_seen": 2793734144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015614141414141414,
      "loss": 2.6523,
      "theoretical_loss": 3.3391308393466965,
      "tokens_seen": 2793734144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015612121212121215,
      "loss": 2.6731,
      "theoretical_loss": 3.339124729374304,
      "tokens_seen": 2793799680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001561010101010101,
      "loss": 2.6244,
      "theoretical_loss": 3.3391186195853657,
      "tokens_seen": 2793865216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015608080808080806,
      "loss": 2.426,
      "theoretical_loss": 3.3391125099798713,
      "tokens_seen": 2793930752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015606060606060607,
      "loss": 2.6614,
      "theoretical_loss": 3.3391064005578115,
      "tokens_seen": 2793996288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015604040404040403,
      "loss": 2.6031,
      "theoretical_loss": 3.3391002913191765,
      "tokens_seen": 2794061824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015602020202020204,
      "loss": 2.5909,
      "theoretical_loss": 3.339094182263956,
      "tokens_seen": 2794127360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000156,
      "loss": 2.8263,
      "theoretical_loss": 3.339088073392141,
      "tokens_seen": 2794192896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015597979797979798,
      "loss": 2.5441,
      "theoretical_loss": 3.339081964703721,
      "tokens_seen": 2794258432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015595959595959597,
      "loss": 2.6077,
      "theoretical_loss": 3.3390758561986864,
      "tokens_seen": 2794323968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015593939393939395,
      "loss": 2.5086,
      "theoretical_loss": 3.3390697478770273,
      "tokens_seen": 2794389504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015591919191919193,
      "loss": 2.5239,
      "theoretical_loss": 3.339063639738734,
      "tokens_seen": 2794455040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001558989898989899,
      "loss": 2.5868,
      "theoretical_loss": 3.339057531783797,
      "tokens_seen": 2794520576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015587878787878787,
      "loss": 2.601,
      "theoretical_loss": 3.339051424012206,
      "tokens_seen": 2794586112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015585858585858586,
      "loss": 2.372,
      "theoretical_loss": 3.339045316423951,
      "tokens_seen": 2794651648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015583838383838384,
      "loss": 2.5286,
      "theoretical_loss": 3.339039209019023,
      "tokens_seen": 2794717184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015581818181818183,
      "loss": 2.4864,
      "theoretical_loss": 3.339033101797412,
      "tokens_seen": 2794782720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001557979797979798,
      "loss": 2.4939,
      "theoretical_loss": 3.3390269947591076,
      "tokens_seen": 2794848256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015577777777777777,
      "loss": 2.6545,
      "theoretical_loss": 3.3390208879041006,
      "tokens_seen": 2794913792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015575757575757578,
      "loss": 2.4191,
      "theoretical_loss": 3.3390147812323807,
      "tokens_seen": 2794979328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015573737373737373,
      "loss": 2.5194,
      "theoretical_loss": 3.3390086747439387,
      "tokens_seen": 2795044864
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015571717171717172,
      "loss": 2.4851,
      "theoretical_loss": 3.3390025684387643,
      "tokens_seen": 2795110400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001556969696969697,
      "loss": 2.3099,
      "theoretical_loss": 3.3389964623168478,
      "tokens_seen": 2795175936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015567676767676766,
      "loss": 2.6324,
      "theoretical_loss": 3.3389903563781798,
      "tokens_seen": 2795241472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015565656565656567,
      "loss": 2.547,
      "theoretical_loss": 3.33898425062275,
      "tokens_seen": 2795307008
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1532487,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.366612434387207,
      "objective/train/theoretical_loss": 3.338978145050549,
      "objective/train/tokens_used": 1154231776,
      "theoretical_loss": 3.338978145050549,
      "tokens_seen": 2795372544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015563636363636363,
      "loss": 2.4968,
      "theoretical_loss": 3.338978145050549,
      "tokens_seen": 2795372544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015561616161616164,
      "loss": 2.2641,
      "theoretical_loss": 3.3389720396615665,
      "tokens_seen": 2795438080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001555959595959596,
      "loss": 2.6109,
      "theoretical_loss": 3.338965934455793,
      "tokens_seen": 2795503616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001555757575757576,
      "loss": 2.5505,
      "theoretical_loss": 3.338959829433219,
      "tokens_seen": 2795569152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015555555555555556,
      "loss": 2.5516,
      "theoretical_loss": 3.3389537245938343,
      "tokens_seen": 2795634688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015553535353535352,
      "loss": 2.4606,
      "theoretical_loss": 3.3389476199376293,
      "tokens_seen": 2795700224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015551515151515153,
      "loss": 2.5946,
      "theoretical_loss": 3.3389415154645943,
      "tokens_seen": 2795765760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001554949494949495,
      "loss": 2.554,
      "theoretical_loss": 3.3389354111747194,
      "tokens_seen": 2795831296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001554747474747475,
      "loss": 2.4172,
      "theoretical_loss": 3.3389293070679944,
      "tokens_seen": 2795896832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015545454545454546,
      "loss": 2.5793,
      "theoretical_loss": 3.33892320314441,
      "tokens_seen": 2795962368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015543434343434344,
      "loss": 2.4952,
      "theoretical_loss": 3.338917099403956,
      "tokens_seen": 2796027904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015541414141414142,
      "loss": 2.5795,
      "theoretical_loss": 3.3389109958466237,
      "tokens_seen": 2796093440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015539393939393938,
      "loss": 2.5075,
      "theoretical_loss": 3.338904892472402,
      "tokens_seen": 2796158976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001553737373737374,
      "loss": 2.4455,
      "theoretical_loss": 3.3388987892812816,
      "tokens_seen": 2796224512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015535353535353535,
      "loss": 2.5127,
      "theoretical_loss": 3.338892686273253,
      "tokens_seen": 2796290048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015533333333333333,
      "loss": 2.7631,
      "theoretical_loss": 3.338886583448306,
      "tokens_seen": 2796355584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015531313131313132,
      "loss": 2.5081,
      "theoretical_loss": 3.3388804808064307,
      "tokens_seen": 2796421120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001552929292929293,
      "loss": 2.4577,
      "theoretical_loss": 3.3388743783476182,
      "tokens_seen": 2796486656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015527272727272728,
      "loss": 2.4207,
      "theoretical_loss": 3.338868276071858,
      "tokens_seen": 2796552192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015525252525252527,
      "loss": 2.5342,
      "theoretical_loss": 3.33886217397914,
      "tokens_seen": 2796617728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015523232323232323,
      "loss": 2.7969,
      "theoretical_loss": 3.3388560720694556,
      "tokens_seen": 2796683264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001552121212121212,
      "loss": 2.5009,
      "theoretical_loss": 3.338849970342794,
      "tokens_seen": 2796748800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001551919191919192,
      "loss": 2.5744,
      "theoretical_loss": 3.3388438687991453,
      "tokens_seen": 2796814336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015517171717171718,
      "loss": 2.5787,
      "theoretical_loss": 3.3388377674385006,
      "tokens_seen": 2796879872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015515151515151516,
      "loss": 2.4585,
      "theoretical_loss": 3.3388316662608495,
      "tokens_seen": 2796945408
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1533684,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3293659687042236,
      "objective/train/theoretical_loss": 3.3388255652661822,
      "objective/train/tokens_used": 1155870176,
      "theoretical_loss": 3.3388255652661822,
      "tokens_seen": 2797010944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015513131313131312,
      "loss": 2.3592,
      "theoretical_loss": 3.3388255652661822,
      "tokens_seen": 2797010944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015511111111111113,
      "loss": 2.5073,
      "theoretical_loss": 3.338819464454489,
      "tokens_seen": 2797076480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015509090909090909,
      "loss": 2.6182,
      "theoretical_loss": 3.3388133638257607,
      "tokens_seen": 2797142016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001550707070707071,
      "loss": 2.6404,
      "theoretical_loss": 3.3388072633799872,
      "tokens_seen": 2797207552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015505050505050505,
      "loss": 2.5199,
      "theoretical_loss": 3.338801163117158,
      "tokens_seen": 2797273088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000155030303030303,
      "loss": 2.5114,
      "theoretical_loss": 3.338795063037264,
      "tokens_seen": 2797338624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015501010101010102,
      "loss": 2.5406,
      "theoretical_loss": 3.3387889631402956,
      "tokens_seen": 2797404160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015498989898989898,
      "loss": 2.4086,
      "theoretical_loss": 3.3387828634262426,
      "tokens_seen": 2797469696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000154969696969697,
      "loss": 2.6467,
      "theoretical_loss": 3.338776763895096,
      "tokens_seen": 2797535232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015494949494949495,
      "loss": 2.659,
      "theoretical_loss": 3.3387706645468445,
      "tokens_seen": 2797600768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015492929292929293,
      "loss": 2.4455,
      "theoretical_loss": 3.3387645653814797,
      "tokens_seen": 2797666304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015490909090909091,
      "loss": 2.6915,
      "theoretical_loss": 3.3387584663989913,
      "tokens_seen": 2797731840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001548888888888889,
      "loss": 2.75,
      "theoretical_loss": 3.3387523675993696,
      "tokens_seen": 2797797376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015486868686868688,
      "loss": 2.5955,
      "theoretical_loss": 3.338746268982605,
      "tokens_seen": 2797862912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015484848484848484,
      "loss": 2.4127,
      "theoretical_loss": 3.3387401705486877,
      "tokens_seen": 2797928448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015482828282828282,
      "loss": 2.4378,
      "theoretical_loss": 3.3387340722976075,
      "tokens_seen": 2797993984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001548080808080808,
      "loss": 2.6867,
      "theoretical_loss": 3.338727974229355,
      "tokens_seen": 2798059520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001547878787878788,
      "loss": 2.5887,
      "theoretical_loss": 3.3387218763439206,
      "tokens_seen": 2798125056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015476767676767677,
      "loss": 2.6311,
      "theoretical_loss": 3.338715778641294,
      "tokens_seen": 2798190592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015474747474747476,
      "loss": 2.3161,
      "theoretical_loss": 3.3387096811214665,
      "tokens_seen": 2798256128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015472727272727274,
      "loss": 2.7239,
      "theoretical_loss": 3.3387035837844268,
      "tokens_seen": 2798321664
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001547070707070707,
      "loss": 2.698,
      "theoretical_loss": 3.3386974866301666,
      "tokens_seen": 2798387200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015468686868686868,
      "loss": 2.5643,
      "theoretical_loss": 3.338691389658675,
      "tokens_seen": 2798452736
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015466666666666667,
      "loss": 2.5428,
      "theoretical_loss": 3.3386852928699433,
      "tokens_seen": 2798518272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015464646464646465,
      "loss": 2.2315,
      "theoretical_loss": 3.3386791962639606,
      "tokens_seen": 2798583808
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1534363,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5379207134246826,
      "objective/train/theoretical_loss": 3.338673099840718,
      "objective/train/tokens_used": 1157508576,
      "theoretical_loss": 3.338673099840718,
      "tokens_seen": 2798649344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015462626262626264,
      "loss": 2.4977,
      "theoretical_loss": 3.338673099840718,
      "tokens_seen": 2798649344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015460606060606062,
      "loss": 2.412,
      "theoretical_loss": 3.3386670036002055,
      "tokens_seen": 2798714880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015458585858585858,
      "loss": 2.5595,
      "theoretical_loss": 3.338660907542413,
      "tokens_seen": 2798780416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001545656565656566,
      "loss": 2.5129,
      "theoretical_loss": 3.3386548116673316,
      "tokens_seen": 2798845952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015454545454545454,
      "loss": 2.634,
      "theoretical_loss": 3.3386487159749505,
      "tokens_seen": 2798911488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015452525252525253,
      "loss": 2.633,
      "theoretical_loss": 3.3386426204652606,
      "tokens_seen": 2798977024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001545050505050505,
      "loss": 2.4416,
      "theoretical_loss": 3.338636525138252,
      "tokens_seen": 2799042560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015448484848484847,
      "loss": 2.4152,
      "theoretical_loss": 3.338630429993915,
      "tokens_seen": 2799108096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015446464646464648,
      "loss": 2.4567,
      "theoretical_loss": 3.3386243350322395,
      "tokens_seen": 2799173632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015444444444444444,
      "loss": 2.3606,
      "theoretical_loss": 3.3386182402532167,
      "tokens_seen": 2799239168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015442424242424245,
      "loss": 2.6451,
      "theoretical_loss": 3.338612145656836,
      "tokens_seen": 2799304704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001544040404040404,
      "loss": 2.7099,
      "theoretical_loss": 3.3386060512430875,
      "tokens_seen": 2799370240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001543838383838384,
      "loss": 2.5981,
      "theoretical_loss": 3.338599957011962,
      "tokens_seen": 2799435776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015436363636363637,
      "loss": 2.6463,
      "theoretical_loss": 3.3385938629634495,
      "tokens_seen": 2799501312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015434343434343433,
      "loss": 2.4691,
      "theoretical_loss": 3.3385877690975403,
      "tokens_seen": 2799566848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015432323232323234,
      "loss": 2.4918,
      "theoretical_loss": 3.3385816754142246,
      "tokens_seen": 2799632384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001543030303030303,
      "loss": 2.4971,
      "theoretical_loss": 3.3385755819134926,
      "tokens_seen": 2799697920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015428282828282828,
      "loss": 2.3587,
      "theoretical_loss": 3.338569488595335,
      "tokens_seen": 2799763456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015426262626262627,
      "loss": 2.6274,
      "theoretical_loss": 3.3385633954597416,
      "tokens_seen": 2799828992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015424242424242425,
      "loss": 2.6756,
      "theoretical_loss": 3.3385573025067026,
      "tokens_seen": 2799894528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015422222222222223,
      "loss": 2.6012,
      "theoretical_loss": 3.3385512097362087,
      "tokens_seen": 2799960064
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015420202020202022,
      "loss": 2.4313,
      "theoretical_loss": 3.3385451171482496,
      "tokens_seen": 2800025600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015418181818181817,
      "loss": 2.5133,
      "theoretical_loss": 3.338539024742816,
      "tokens_seen": 2800091136
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015416161616161616,
      "loss": 2.3476,
      "theoretical_loss": 3.338532932519898,
      "tokens_seen": 2800156672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015414141414141414,
      "loss": 2.5692,
      "theoretical_loss": 3.338526840479486,
      "tokens_seen": 2800222208
    },
    {
      "epoch": 0.69,
      "objective/train/docs_used": 1535268,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.515688419342041,
      "objective/train/theoretical_loss": 3.33852074862157,
      "objective/train/tokens_used": 1159146976,
      "theoretical_loss": 3.33852074862157,
      "tokens_seen": 2800287744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015412121212121213,
      "loss": 2.4144,
      "theoretical_loss": 3.33852074862157,
      "tokens_seen": 2800287744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001541010101010101,
      "loss": 2.7267,
      "theoretical_loss": 3.3385146569461406,
      "tokens_seen": 2800353280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015408080808080807,
      "loss": 2.6239,
      "theoretical_loss": 3.338508565453188,
      "tokens_seen": 2800418816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015406060606060608,
      "loss": 2.6105,
      "theoretical_loss": 3.3385024741427016,
      "tokens_seen": 2800484352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015404040404040403,
      "loss": 2.5347,
      "theoretical_loss": 3.338496383014673,
      "tokens_seen": 2800549888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015402020202020205,
      "loss": 2.5925,
      "theoretical_loss": 3.3384902920690918,
      "tokens_seen": 2800615424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000154,
      "loss": 2.5823,
      "theoretical_loss": 3.338484201305948,
      "tokens_seen": 2800680960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015397979797979796,
      "loss": 2.5441,
      "theoretical_loss": 3.3384781107252324,
      "tokens_seen": 2800746496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015395959595959597,
      "loss": 2.5492,
      "theoretical_loss": 3.338472020326935,
      "tokens_seen": 2800812032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015393939393939393,
      "loss": 2.554,
      "theoretical_loss": 3.338465930111046,
      "tokens_seen": 2800877568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015391919191919194,
      "loss": 2.6539,
      "theoretical_loss": 3.3384598400775563,
      "tokens_seen": 2800943104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001538989898989899,
      "loss": 2.3872,
      "theoretical_loss": 3.338453750226455,
      "tokens_seen": 2801008640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001538787878787879,
      "loss": 2.3886,
      "theoretical_loss": 3.3384476605577333,
      "tokens_seen": 2801074176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015385858585858586,
      "loss": 2.5951,
      "theoretical_loss": 3.3384415710713817,
      "tokens_seen": 2801139712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015383838383838382,
      "loss": 2.5277,
      "theoretical_loss": 3.338435481767389,
      "tokens_seen": 2801205248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015381818181818183,
      "loss": 2.4389,
      "theoretical_loss": 3.3384293926457476,
      "tokens_seen": 2801270784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001537979797979798,
      "loss": 2.5033,
      "theoretical_loss": 3.3384233037064455,
      "tokens_seen": 2801336320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001537777777777778,
      "loss": 2.4545,
      "theoretical_loss": 3.338417214949475,
      "tokens_seen": 2801401856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015375757575757576,
      "loss": 2.4545,
      "theoretical_loss": 3.338411126374825,
      "tokens_seen": 2801467392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015373737373737374,
      "loss": 2.5227,
      "theoretical_loss": 3.3384050379824863,
      "tokens_seen": 2801532928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015371717171717172,
      "loss": 2.7292,
      "theoretical_loss": 3.3383989497724493,
      "tokens_seen": 2801598464
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001536969696969697,
      "loss": 2.3962,
      "theoretical_loss": 3.338392861744704,
      "tokens_seen": 2801664000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001536767676767677,
      "loss": 2.3383,
      "theoretical_loss": 3.3383867738992405,
      "tokens_seen": 2801729536
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015365656565656565,
      "loss": 2.6677,
      "theoretical_loss": 3.3383806862360497,
      "tokens_seen": 2801795072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015363636363636363,
      "loss": 2.5149,
      "theoretical_loss": 3.3383745987551214,
      "tokens_seen": 2801860608
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1536573,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4081475734710693,
      "objective/train/theoretical_loss": 3.338368511456446,
      "objective/train/tokens_used": 1160785376,
      "theoretical_loss": 3.338368511456446,
      "tokens_seen": 2801926144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015361616161616162,
      "loss": 2.5285,
      "theoretical_loss": 3.338368511456446,
      "tokens_seen": 2801926144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001535959595959596,
      "loss": 2.3222,
      "theoretical_loss": 3.338362424340014,
      "tokens_seen": 2801991680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015357575757575758,
      "loss": 2.6336,
      "theoretical_loss": 3.338356337405815,
      "tokens_seen": 2802057216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015355555555555557,
      "loss": 2.6195,
      "theoretical_loss": 3.33835025065384,
      "tokens_seen": 2802122752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015353535353535353,
      "loss": 2.627,
      "theoretical_loss": 3.3383441640840794,
      "tokens_seen": 2802188288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015351515151515154,
      "loss": 2.4161,
      "theoretical_loss": 3.338338077696523,
      "tokens_seen": 2802253824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001534949494949495,
      "loss": 2.6368,
      "theoretical_loss": 3.338331991491161,
      "tokens_seen": 2802319360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015347474747474748,
      "loss": 2.5341,
      "theoretical_loss": 3.338325905467984,
      "tokens_seen": 2802384896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015345454545454546,
      "loss": 2.5303,
      "theoretical_loss": 3.338319819626982,
      "tokens_seen": 2802450432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015343434343434342,
      "loss": 2.579,
      "theoretical_loss": 3.3383137339681457,
      "tokens_seen": 2802515968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015341414141414143,
      "loss": 2.6865,
      "theoretical_loss": 3.338307648491465,
      "tokens_seen": 2802581504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015339393939393939,
      "loss": 2.5492,
      "theoretical_loss": 3.3383015631969304,
      "tokens_seen": 2802647040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001533737373737374,
      "loss": 2.6026,
      "theoretical_loss": 3.338295478084532,
      "tokens_seen": 2802712576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015335353535353535,
      "loss": 2.618,
      "theoretical_loss": 3.3382893931542603,
      "tokens_seen": 2802778112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015333333333333334,
      "loss": 2.6311,
      "theoretical_loss": 3.3382833084061057,
      "tokens_seen": 2802843648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015331313131313132,
      "loss": 2.4723,
      "theoretical_loss": 3.338277223840058,
      "tokens_seen": 2802909184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015329292929292928,
      "loss": 2.6152,
      "theoretical_loss": 3.338271139456108,
      "tokens_seen": 2802974720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001532727272727273,
      "loss": 2.5175,
      "theoretical_loss": 3.3382650552542463,
      "tokens_seen": 2803040256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015325252525252525,
      "loss": 2.5136,
      "theoretical_loss": 3.338258971234462,
      "tokens_seen": 2803105792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015323232323232323,
      "loss": 2.4572,
      "theoretical_loss": 3.338252887396746,
      "tokens_seen": 2803171328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015321212121212121,
      "loss": 2.7119,
      "theoretical_loss": 3.3382468037410895,
      "tokens_seen": 2803236864
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001531919191919192,
      "loss": 2.3428,
      "theoretical_loss": 3.338240720267481,
      "tokens_seen": 2803302400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015317171717171718,
      "loss": 2.4571,
      "theoretical_loss": 3.3382346369759124,
      "tokens_seen": 2803367936
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015315151515151514,
      "loss": 2.6215,
      "theoretical_loss": 3.338228553866373,
      "tokens_seen": 2803433472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015313131313131312,
      "loss": 2.3747,
      "theoretical_loss": 3.3382224709388537,
      "tokens_seen": 2803499008
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1537291,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.69579815864563,
      "objective/train/theoretical_loss": 3.3382163881933447,
      "objective/train/tokens_used": 1162423776,
      "theoretical_loss": 3.3382163881933447,
      "tokens_seen": 2803564544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001531111111111111,
      "loss": 2.5256,
      "theoretical_loss": 3.3382163881933447,
      "tokens_seen": 2803564544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001530909090909091,
      "loss": 2.414,
      "theoretical_loss": 3.338210305629836,
      "tokens_seen": 2803630080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015307070707070708,
      "loss": 2.3806,
      "theoretical_loss": 3.338204223248318,
      "tokens_seen": 2803695616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015305050505050506,
      "loss": 2.6628,
      "theoretical_loss": 3.338198141048781,
      "tokens_seen": 2803761152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015303030303030302,
      "loss": 2.2473,
      "theoretical_loss": 3.3381920590312153,
      "tokens_seen": 2803826688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015301010101010103,
      "loss": 2.4544,
      "theoretical_loss": 3.3381859771956113,
      "tokens_seen": 2803892224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015298989898989898,
      "loss": 2.5157,
      "theoretical_loss": 3.3381798955419595,
      "tokens_seen": 2803957760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015296969696969697,
      "loss": 2.2781,
      "theoretical_loss": 3.3381738140702497,
      "tokens_seen": 2804023296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015294949494949495,
      "loss": 2.5778,
      "theoretical_loss": 3.3381677327804726,
      "tokens_seen": 2804088832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015292929292929294,
      "loss": 2.5348,
      "theoretical_loss": 3.3381616516726185,
      "tokens_seen": 2804154368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015290909090909092,
      "loss": 2.4686,
      "theoretical_loss": 3.3381555707466775,
      "tokens_seen": 2804219904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015288888888888888,
      "loss": 2.725,
      "theoretical_loss": 3.33814949000264,
      "tokens_seen": 2804285440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001528686868686869,
      "loss": 2.4045,
      "theoretical_loss": 3.3381434094404963,
      "tokens_seen": 2804350976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015284848484848484,
      "loss": 2.6955,
      "theoretical_loss": 3.338137329060237,
      "tokens_seen": 2804416512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015282828282828286,
      "loss": 2.335,
      "theoretical_loss": 3.3381312488618518,
      "tokens_seen": 2804482048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001528080808080808,
      "loss": 2.606,
      "theoretical_loss": 3.3381251688453313,
      "tokens_seen": 2804547584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015278787878787877,
      "loss": 2.3848,
      "theoretical_loss": 3.338119089010666,
      "tokens_seen": 2804613120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015276767676767678,
      "loss": 2.5601,
      "theoretical_loss": 3.338113009357846,
      "tokens_seen": 2804678656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015274747474747474,
      "loss": 2.8272,
      "theoretical_loss": 3.3381069298868615,
      "tokens_seen": 2804744192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015272727272727275,
      "loss": 2.4516,
      "theoretical_loss": 3.3381008505977032,
      "tokens_seen": 2804809728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001527070707070707,
      "loss": 2.6024,
      "theoretical_loss": 3.3380947714903613,
      "tokens_seen": 2804875264
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001526868686868687,
      "loss": 2.5488,
      "theoretical_loss": 3.338088692564826,
      "tokens_seen": 2804940800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015266666666666667,
      "loss": 2.4285,
      "theoretical_loss": 3.3380826138210873,
      "tokens_seen": 2805006336
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015264646464646466,
      "loss": 2.5808,
      "theoretical_loss": 3.3380765352591357,
      "tokens_seen": 2805071872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015262626262626264,
      "loss": 2.5374,
      "theoretical_loss": 3.3380704568789623,
      "tokens_seen": 2805137408
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1538491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.382876396179199,
      "objective/train/theoretical_loss": 3.338064378680556,
      "objective/train/tokens_used": 1164062176,
      "theoretical_loss": 3.338064378680556,
      "tokens_seen": 2805202944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001526060606060606,
      "loss": 2.4924,
      "theoretical_loss": 3.338064378680556,
      "tokens_seen": 2805202944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015258585858585858,
      "loss": 2.5221,
      "theoretical_loss": 3.3380583006639086,
      "tokens_seen": 2805268480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015256565656565657,
      "loss": 2.4082,
      "theoretical_loss": 3.3380522228290093,
      "tokens_seen": 2805334016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015254545454545455,
      "loss": 2.2068,
      "theoretical_loss": 3.338046145175849,
      "tokens_seen": 2805399552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015252525252525253,
      "loss": 2.5177,
      "theoretical_loss": 3.3380400677044175,
      "tokens_seen": 2805465088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015250505050505052,
      "loss": 2.7238,
      "theoretical_loss": 3.338033990414706,
      "tokens_seen": 2805530624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015248484848484847,
      "loss": 2.5127,
      "theoretical_loss": 3.3380279133067043,
      "tokens_seen": 2805596160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015246464646464649,
      "loss": 2.3417,
      "theoretical_loss": 3.338021836380402,
      "tokens_seen": 2805661696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015244444444444444,
      "loss": 2.3691,
      "theoretical_loss": 3.338015759635791,
      "tokens_seen": 2805727232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015242424242424243,
      "loss": 2.6362,
      "theoretical_loss": 3.33800968307286,
      "tokens_seen": 2805792768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001524040404040404,
      "loss": 2.5807,
      "theoretical_loss": 3.3380036066916006,
      "tokens_seen": 2805858304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015238383838383837,
      "loss": 2.4407,
      "theoretical_loss": 3.337997530492003,
      "tokens_seen": 2805923840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015236363636363638,
      "loss": 2.5439,
      "theoretical_loss": 3.3379914544740563,
      "tokens_seen": 2805989376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015234343434343434,
      "loss": 2.594,
      "theoretical_loss": 3.337985378637752,
      "tokens_seen": 2806054912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015232323232323235,
      "loss": 2.724,
      "theoretical_loss": 3.33797930298308,
      "tokens_seen": 2806120448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001523030303030303,
      "loss": 2.4991,
      "theoretical_loss": 3.337973227510031,
      "tokens_seen": 2806185984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015228282828282826,
      "loss": 2.5294,
      "theoretical_loss": 3.337967152218595,
      "tokens_seen": 2806251520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015226262626262627,
      "loss": 2.2893,
      "theoretical_loss": 3.337961077108762,
      "tokens_seen": 2806317056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015224242424242423,
      "loss": 2.2974,
      "theoretical_loss": 3.337955002180523,
      "tokens_seen": 2806382592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015222222222222224,
      "loss": 2.5174,
      "theoretical_loss": 3.337948927433868,
      "tokens_seen": 2806448128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001522020202020202,
      "loss": 2.6392,
      "theoretical_loss": 3.3379428528687876,
      "tokens_seen": 2806513664
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015218181818181818,
      "loss": 2.4954,
      "theoretical_loss": 3.3379367784852714,
      "tokens_seen": 2806579200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015216161616161616,
      "loss": 2.4398,
      "theoretical_loss": 3.337930704283311,
      "tokens_seen": 2806644736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015214141414141415,
      "loss": 2.5839,
      "theoretical_loss": 3.3379246302628953,
      "tokens_seen": 2806710272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015212121212121213,
      "loss": 2.4893,
      "theoretical_loss": 3.3379185564240155,
      "tokens_seen": 2806775808
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1539064,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.295426845550537,
      "objective/train/theoretical_loss": 3.3379124827666615,
      "objective/train/tokens_used": 1165700576,
      "theoretical_loss": 3.3379124827666615,
      "tokens_seen": 2806841344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001521010101010101,
      "loss": 2.3355,
      "theoretical_loss": 3.3379124827666615,
      "tokens_seen": 2806841344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001520808080808081,
      "loss": 2.3514,
      "theoretical_loss": 3.337906409290824,
      "tokens_seen": 2806906880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015206060606060606,
      "loss": 2.4208,
      "theoretical_loss": 3.3379003359964936,
      "tokens_seen": 2806972416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015204040404040404,
      "loss": 2.525,
      "theoretical_loss": 3.33789426288366,
      "tokens_seen": 2807037952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015202020202020202,
      "loss": 2.7146,
      "theoretical_loss": 3.337888189952314,
      "tokens_seen": 2807103488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000152,
      "loss": 2.4362,
      "theoretical_loss": 3.3378821172024455,
      "tokens_seen": 2807169024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000151979797979798,
      "loss": 2.4429,
      "theoretical_loss": 3.337876044634045,
      "tokens_seen": 2807234560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015195959595959598,
      "loss": 2.5753,
      "theoretical_loss": 3.3378699722471032,
      "tokens_seen": 2807300096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015193939393939393,
      "loss": 2.5767,
      "theoretical_loss": 3.33786390004161,
      "tokens_seen": 2807365632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015191919191919192,
      "loss": 2.6383,
      "theoretical_loss": 3.337857828017556,
      "tokens_seen": 2807431168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001518989898989899,
      "loss": 2.5414,
      "theoretical_loss": 3.3378517561749312,
      "tokens_seen": 2807496704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015187878787878789,
      "loss": 2.3931,
      "theoretical_loss": 3.3378456845137263,
      "tokens_seen": 2807562240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015185858585858587,
      "loss": 2.5273,
      "theoretical_loss": 3.337839613033932,
      "tokens_seen": 2807627776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015183838383838383,
      "loss": 2.4399,
      "theoretical_loss": 3.3378335417355376,
      "tokens_seen": 2807693312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015181818181818184,
      "loss": 2.5798,
      "theoretical_loss": 3.3378274706185342,
      "tokens_seen": 2807758848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001517979797979798,
      "loss": 2.5198,
      "theoretical_loss": 3.337821399682912,
      "tokens_seen": 2807824384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001517777777777778,
      "loss": 2.539,
      "theoretical_loss": 3.3378153289286616,
      "tokens_seen": 2807889920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015175757575757576,
      "loss": 2.4885,
      "theoretical_loss": 3.337809258355773,
      "tokens_seen": 2807955456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015173737373737372,
      "loss": 2.4359,
      "theoretical_loss": 3.3378031879642363,
      "tokens_seen": 2808020992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015171717171717173,
      "loss": 2.3851,
      "theoretical_loss": 3.337797117754042,
      "tokens_seen": 2808086528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001516969696969697,
      "loss": 2.3875,
      "theoretical_loss": 3.337791047725181,
      "tokens_seen": 2808152064
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001516767676767677,
      "loss": 2.6053,
      "theoretical_loss": 3.337784977877643,
      "tokens_seen": 2808217600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015165656565656565,
      "loss": 2.5705,
      "theoretical_loss": 3.337778908211419,
      "tokens_seen": 2808283136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015163636363636364,
      "loss": 2.3152,
      "theoretical_loss": 3.3377728387264987,
      "tokens_seen": 2808348672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015161616161616162,
      "loss": 2.4822,
      "theoretical_loss": 3.3377667694228728,
      "tokens_seen": 2808414208
    },
    {
      "debugging/Self-BLEU-5": 0.4089085537148741,
      "debugging/distinct-1-grams": 0.7963924931509507,
      "debugging/distinct-2-grams": 0.9664141286579346,
      "debugging/entropy-1-grams": 5.66123923722335,
      "debugging/entropy-2-grams": 6.274232850805275,
      "debugging/length": 520.625,
      "debugging/num_segments": 8,
      "epoch": 0.7,
      "objective/train/docs_used": 1540274,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8786096572875977,
      "objective/train/theoretical_loss": 3.3377607003005316,
      "objective/train/tokens_used": 1167338976,
      "theoretical_loss": 3.3377607003005316,
      "tokens_seen": 2808479744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001515959595959596,
      "loss": 2.6427,
      "theoretical_loss": 3.3377607003005316,
      "tokens_seen": 2808479744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001515757575757576,
      "loss": 2.2885,
      "theoretical_loss": 3.3377546313594655,
      "tokens_seen": 2808545280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015155555555555555,
      "loss": 2.5785,
      "theoretical_loss": 3.3377485625996646,
      "tokens_seen": 2808610816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015153535353535353,
      "loss": 2.5783,
      "theoretical_loss": 3.3377424940211196,
      "tokens_seen": 2808676352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015151515151515152,
      "loss": 2.3863,
      "theoretical_loss": 3.337736425623821,
      "tokens_seen": 2808741888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001514949494949495,
      "loss": 2.6048,
      "theoretical_loss": 3.337730357407758,
      "tokens_seen": 2808807424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015147474747474748,
      "loss": 2.3878,
      "theoretical_loss": 3.3377242893729226,
      "tokens_seen": 2808872960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015145454545454547,
      "loss": 2.7406,
      "theoretical_loss": 3.3377182215193044,
      "tokens_seen": 2808938496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015143434343434342,
      "loss": 2.5364,
      "theoretical_loss": 3.3377121538468932,
      "tokens_seen": 2809004032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001514141414141414,
      "loss": 2.6531,
      "theoretical_loss": 3.3377060863556807,
      "tokens_seen": 2809069568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001513939393939394,
      "loss": 2.5639,
      "theoretical_loss": 3.3377000190456556,
      "tokens_seen": 2809135104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015137373737373738,
      "loss": 2.6856,
      "theoretical_loss": 3.3376939519168096,
      "tokens_seen": 2809200640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015135353535353536,
      "loss": 2.28,
      "theoretical_loss": 3.3376878849691325,
      "tokens_seen": 2809266176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015133333333333332,
      "loss": 2.6922,
      "theoretical_loss": 3.337681818202615,
      "tokens_seen": 2809331712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015131313131313133,
      "loss": 2.6139,
      "theoretical_loss": 3.337675751617247,
      "tokens_seen": 2809397248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015129292929292928,
      "loss": 2.586,
      "theoretical_loss": 3.337669685213019,
      "tokens_seen": 2809462784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001512727272727273,
      "loss": 2.35,
      "theoretical_loss": 3.3376636189899216,
      "tokens_seen": 2809528320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015125252525252525,
      "loss": 2.6281,
      "theoretical_loss": 3.3376575529479453,
      "tokens_seen": 2809593856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015123232323232324,
      "loss": 2.4651,
      "theoretical_loss": 3.3376514870870797,
      "tokens_seen": 2809659392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015121212121212122,
      "loss": 2.7367,
      "theoretical_loss": 3.3376454214073155,
      "tokens_seen": 2809724928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015119191919191918,
      "loss": 2.2225,
      "theoretical_loss": 3.3376393559086437,
      "tokens_seen": 2809790464
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001511717171717172,
      "loss": 2.6578,
      "theoretical_loss": 3.337633290591054,
      "tokens_seen": 2809856000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015115151515151515,
      "loss": 2.4453,
      "theoretical_loss": 3.337627225454537,
      "tokens_seen": 2809921536
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015113131313131316,
      "loss": 2.6169,
      "theoretical_loss": 3.3376211604990833,
      "tokens_seen": 2809987072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001511111111111111,
      "loss": 2.4065,
      "theoretical_loss": 3.337615095724683,
      "tokens_seen": 2810052608
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1540907,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3177528381347656,
      "objective/train/theoretical_loss": 3.337609031131326,
      "objective/train/tokens_used": 1168977376,
      "theoretical_loss": 3.337609031131326,
      "tokens_seen": 2810118144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001510909090909091,
      "loss": 2.373,
      "theoretical_loss": 3.337609031131326,
      "tokens_seen": 2810118144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015107070707070708,
      "loss": 2.4572,
      "theoretical_loss": 3.3376029667190035,
      "tokens_seen": 2810183680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015105050505050504,
      "loss": 2.4451,
      "theoretical_loss": 3.3375969024877055,
      "tokens_seen": 2810249216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015103030303030305,
      "loss": 2.3305,
      "theoretical_loss": 3.337590838437422,
      "tokens_seen": 2810314752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000151010101010101,
      "loss": 2.4307,
      "theoretical_loss": 3.3375847745681444,
      "tokens_seen": 2810380288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000150989898989899,
      "loss": 2.3383,
      "theoretical_loss": 3.3375787108798622,
      "tokens_seen": 2810445824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015096969696969697,
      "loss": 2.2622,
      "theoretical_loss": 3.337572647372566,
      "tokens_seen": 2810511360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015094949494949496,
      "loss": 2.6768,
      "theoretical_loss": 3.337566584046246,
      "tokens_seen": 2810576896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015092929292929294,
      "loss": 2.6204,
      "theoretical_loss": 3.337560520900893,
      "tokens_seen": 2810642432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015090909090909093,
      "loss": 2.3181,
      "theoretical_loss": 3.3375544579364975,
      "tokens_seen": 2810707968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015088888888888888,
      "loss": 2.2021,
      "theoretical_loss": 3.337548395153049,
      "tokens_seen": 2810773504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015086868686868687,
      "loss": 2.452,
      "theoretical_loss": 3.3375423325505387,
      "tokens_seen": 2810839040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015084848484848485,
      "loss": 2.446,
      "theoretical_loss": 3.3375362701289566,
      "tokens_seen": 2810904576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015082828282828283,
      "loss": 2.4306,
      "theoretical_loss": 3.3375302078882934,
      "tokens_seen": 2810970112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015080808080808082,
      "loss": 2.7435,
      "theoretical_loss": 3.337524145828539,
      "tokens_seen": 2811035648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015078787878787877,
      "loss": 2.3333,
      "theoretical_loss": 3.3375180839496847,
      "tokens_seen": 2811101184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015076767676767679,
      "loss": 2.5902,
      "theoretical_loss": 3.33751202225172,
      "tokens_seen": 2811166720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015074747474747474,
      "loss": 2.5236,
      "theoretical_loss": 3.337505960734635,
      "tokens_seen": 2811232256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015072727272727275,
      "loss": 2.4051,
      "theoretical_loss": 3.337499899398421,
      "tokens_seen": 2811297792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001507070707070707,
      "loss": 2.5784,
      "theoretical_loss": 3.337493838243068,
      "tokens_seen": 2811363328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015068686868686867,
      "loss": 2.574,
      "theoretical_loss": 3.3374877772685663,
      "tokens_seen": 2811428864
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015066666666666668,
      "loss": 2.6748,
      "theoretical_loss": 3.3374817164749064,
      "tokens_seen": 2811494400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015064646464646464,
      "loss": 2.3427,
      "theoretical_loss": 3.337475655862079,
      "tokens_seen": 2811559936
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015062626262626265,
      "loss": 2.6201,
      "theoretical_loss": 3.3374695954300737,
      "tokens_seen": 2811625472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001506060606060606,
      "loss": 2.5978,
      "theoretical_loss": 3.3374635351788813,
      "tokens_seen": 2811691008
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1542005,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.391789436340332,
      "objective/train/theoretical_loss": 3.3374574751084927,
      "objective/train/tokens_used": 1170615776,
      "theoretical_loss": 3.3374574751084927,
      "tokens_seen": 2811756544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001505858585858586,
      "loss": 2.4917,
      "theoretical_loss": 3.3374574751084927,
      "tokens_seen": 2811756544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015056565656565657,
      "loss": 2.3723,
      "theoretical_loss": 3.337451415218897,
      "tokens_seen": 2811822080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015054545454545453,
      "loss": 2.4739,
      "theoretical_loss": 3.3374453555100865,
      "tokens_seen": 2811887616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015052525252525254,
      "loss": 2.6674,
      "theoretical_loss": 3.33743929598205,
      "tokens_seen": 2811953152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001505050505050505,
      "loss": 2.5012,
      "theoretical_loss": 3.3374332366347783,
      "tokens_seen": 2812018688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015048484848484848,
      "loss": 2.3637,
      "theoretical_loss": 3.3374271774682622,
      "tokens_seen": 2812084224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015046464646464646,
      "loss": 2.6741,
      "theoretical_loss": 3.337421118482492,
      "tokens_seen": 2812149760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015044444444444445,
      "loss": 2.6175,
      "theoretical_loss": 3.337415059677457,
      "tokens_seen": 2812215296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015042424242424243,
      "loss": 2.3771,
      "theoretical_loss": 3.337409001053149,
      "tokens_seen": 2812280832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015040404040404042,
      "loss": 2.5683,
      "theoretical_loss": 3.3374029426095584,
      "tokens_seen": 2812346368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001503838383838384,
      "loss": 2.5182,
      "theoretical_loss": 3.3373968843466746,
      "tokens_seen": 2812411904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015036363636363636,
      "loss": 2.6133,
      "theoretical_loss": 3.3373908262644885,
      "tokens_seen": 2812477440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015034343434343434,
      "loss": 2.2606,
      "theoretical_loss": 3.3373847683629903,
      "tokens_seen": 2812542976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015032323232323232,
      "loss": 2.5747,
      "theoretical_loss": 3.337378710642171,
      "tokens_seen": 2812608512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001503030303030303,
      "loss": 2.5645,
      "theoretical_loss": 3.3373726531020207,
      "tokens_seen": 2812674048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001502828282828283,
      "loss": 2.5088,
      "theoretical_loss": 3.3373665957425294,
      "tokens_seen": 2812739584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015026262626262628,
      "loss": 2.6207,
      "theoretical_loss": 3.3373605385636878,
      "tokens_seen": 2812805120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015024242424242423,
      "loss": 2.6206,
      "theoretical_loss": 3.337354481565486,
      "tokens_seen": 2812870656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015022222222222224,
      "loss": 2.6419,
      "theoretical_loss": 3.3373484247479155,
      "tokens_seen": 2812936192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001502020202020202,
      "loss": 2.3964,
      "theoretical_loss": 3.3373423681109653,
      "tokens_seen": 2813001728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015018181818181819,
      "loss": 2.5811,
      "theoretical_loss": 3.3373363116546266,
      "tokens_seen": 2813067264
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015016161616161617,
      "loss": 2.6422,
      "theoretical_loss": 3.3373302553788897,
      "tokens_seen": 2813132800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015014141414141413,
      "loss": 2.439,
      "theoretical_loss": 3.337324199283745,
      "tokens_seen": 2813198336
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015012121212121214,
      "loss": 2.6012,
      "theoretical_loss": 3.3373181433691825,
      "tokens_seen": 2813263872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001501010101010101,
      "loss": 2.5056,
      "theoretical_loss": 3.337312087635193,
      "tokens_seen": 2813329408
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1543381,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4076125621795654,
      "objective/train/theoretical_loss": 3.337306032081767,
      "objective/train/tokens_used": 1172254176,
      "theoretical_loss": 3.337306032081767,
      "tokens_seen": 2813394944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001500808080808081,
      "loss": 2.4433,
      "theoretical_loss": 3.337306032081767,
      "tokens_seen": 2813394944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015006060606060606,
      "loss": 2.5988,
      "theoretical_loss": 3.337299976708895,
      "tokens_seen": 2813460480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015004040404040405,
      "loss": 2.6104,
      "theoretical_loss": 3.337293921516567,
      "tokens_seen": 2813526016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015002020202020203,
      "loss": 2.6658,
      "theoretical_loss": 3.337287866504773,
      "tokens_seen": 2813591552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015,
      "loss": 2.4219,
      "theoretical_loss": 3.337281811673505,
      "tokens_seen": 2813657088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000149979797979798,
      "loss": 2.6402,
      "theoretical_loss": 3.337275757022752,
      "tokens_seen": 2813722624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014995959595959595,
      "loss": 2.3286,
      "theoretical_loss": 3.3372697025525047,
      "tokens_seen": 2813788160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014993939393939394,
      "loss": 2.5936,
      "theoretical_loss": 3.337263648262754,
      "tokens_seen": 2813853696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014991919191919192,
      "loss": 2.3746,
      "theoretical_loss": 3.337257594153489,
      "tokens_seen": 2813919232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001498989898989899,
      "loss": 2.3566,
      "theoretical_loss": 3.3372515402247025,
      "tokens_seen": 2813984768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001498787878787879,
      "loss": 2.3936,
      "theoretical_loss": 3.3372454864763825,
      "tokens_seen": 2814050304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014985858585858585,
      "loss": 2.5389,
      "theoretical_loss": 3.337239432908521,
      "tokens_seen": 2814115840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014983838383838383,
      "loss": 2.5794,
      "theoretical_loss": 3.3372333795211073,
      "tokens_seen": 2814181376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014981818181818182,
      "loss": 2.7035,
      "theoretical_loss": 3.3372273263141325,
      "tokens_seen": 2814246912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001497979797979798,
      "loss": 2.5967,
      "theoretical_loss": 3.337221273287587,
      "tokens_seen": 2814312448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014977777777777778,
      "loss": 2.4073,
      "theoretical_loss": 3.337215220441461,
      "tokens_seen": 2814377984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014975757575757577,
      "loss": 2.2951,
      "theoretical_loss": 3.337209167775745,
      "tokens_seen": 2814443520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014973737373737372,
      "loss": 2.4842,
      "theoretical_loss": 3.33720311529043,
      "tokens_seen": 2814509056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014971717171717174,
      "loss": 2.3849,
      "theoretical_loss": 3.337197062985505,
      "tokens_seen": 2814574592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001496969696969697,
      "loss": 2.5229,
      "theoretical_loss": 3.3371910108609617,
      "tokens_seen": 2814640128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014967676767676768,
      "loss": 2.4882,
      "theoretical_loss": 3.33718495891679,
      "tokens_seen": 2814705664
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014965656565656566,
      "loss": 2.5955,
      "theoretical_loss": 3.3371789071529805,
      "tokens_seen": 2814771200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014963636363636362,
      "loss": 2.4903,
      "theoretical_loss": 3.3371728555695235,
      "tokens_seen": 2814836736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014961616161616163,
      "loss": 2.4997,
      "theoretical_loss": 3.3371668041664098,
      "tokens_seen": 2814902272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014959595959595958,
      "loss": 2.5782,
      "theoretical_loss": 3.337160752943629,
      "tokens_seen": 2814967808
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1543761,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.297109603881836,
      "objective/train/theoretical_loss": 3.3371547019011722,
      "objective/train/tokens_used": 1173892576,
      "theoretical_loss": 3.3371547019011722,
      "tokens_seen": 2815033344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001495757575757576,
      "loss": 2.5068,
      "theoretical_loss": 3.3371547019011722,
      "tokens_seen": 2815033344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014955555555555555,
      "loss": 2.6634,
      "theoretical_loss": 3.33714865103903,
      "tokens_seen": 2815098880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014953535353535356,
      "loss": 2.3954,
      "theoretical_loss": 3.337142600357192,
      "tokens_seen": 2815164416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014951515151515152,
      "loss": 2.5692,
      "theoretical_loss": 3.337136549855649,
      "tokens_seen": 2815229952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014949494949494948,
      "loss": 2.332,
      "theoretical_loss": 3.3371304995343922,
      "tokens_seen": 2815295488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001494747474747475,
      "loss": 2.5497,
      "theoretical_loss": 3.337124449393411,
      "tokens_seen": 2815361024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014945454545454545,
      "loss": 2.5498,
      "theoretical_loss": 3.3371183994326965,
      "tokens_seen": 2815426560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014943434343434346,
      "loss": 2.3472,
      "theoretical_loss": 3.3371123496522386,
      "tokens_seen": 2815492096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001494141414141414,
      "loss": 2.5596,
      "theoretical_loss": 3.337106300052028,
      "tokens_seen": 2815557632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001493939393939394,
      "loss": 2.4906,
      "theoretical_loss": 3.337100250632055,
      "tokens_seen": 2815623168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014937373737373738,
      "loss": 2.4941,
      "theoretical_loss": 3.3370942013923104,
      "tokens_seen": 2815688704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014935353535353537,
      "loss": 2.505,
      "theoretical_loss": 3.337088152332784,
      "tokens_seen": 2815754240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014933333333333335,
      "loss": 2.7451,
      "theoretical_loss": 3.3370821034534672,
      "tokens_seen": 2815819776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001493131313131313,
      "loss": 2.4954,
      "theoretical_loss": 3.3370760547543497,
      "tokens_seen": 2815885312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001492929292929293,
      "loss": 2.5513,
      "theoretical_loss": 3.337070006235422,
      "tokens_seen": 2815950848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014927272727272727,
      "loss": 2.3103,
      "theoretical_loss": 3.3370639578966745,
      "tokens_seen": 2816016384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014925252525252526,
      "loss": 2.507,
      "theoretical_loss": 3.337057909738098,
      "tokens_seen": 2816081920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014923232323232324,
      "loss": 2.6062,
      "theoretical_loss": 3.337051861759683,
      "tokens_seen": 2816147456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014921212121212123,
      "loss": 2.5351,
      "theoretical_loss": 3.337045813961419,
      "tokens_seen": 2816212992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014919191919191918,
      "loss": 2.4436,
      "theoretical_loss": 3.3370397663432976,
      "tokens_seen": 2816278528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001491717171717172,
      "loss": 2.5447,
      "theoretical_loss": 3.3370337189053085,
      "tokens_seen": 2816344064
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014915151515151515,
      "loss": 2.1893,
      "theoretical_loss": 3.3370276716474425,
      "tokens_seen": 2816409600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014913131313131313,
      "loss": 2.4392,
      "theoretical_loss": 3.3370216245696898,
      "tokens_seen": 2816475136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014911111111111112,
      "loss": 2.5965,
      "theoretical_loss": 3.337015577672041,
      "tokens_seen": 2816540672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014909090909090908,
      "loss": 2.3938,
      "theoretical_loss": 3.3370095309544867,
      "tokens_seen": 2816606208
    },
    {
      "epoch": 0.7,
      "objective/train/docs_used": 1544909,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0402820110321045,
      "objective/train/theoretical_loss": 3.337003484417017,
      "objective/train/tokens_used": 1175530976,
      "theoretical_loss": 3.337003484417017,
      "tokens_seen": 2816671744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001490707070707071,
      "loss": 2.4325,
      "theoretical_loss": 3.337003484417017,
      "tokens_seen": 2816671744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014905050505050504,
      "loss": 2.6141,
      "theoretical_loss": 3.3369974380596226,
      "tokens_seen": 2816737280
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014903030303030305,
      "loss": 2.5239,
      "theoretical_loss": 3.336991391882294,
      "tokens_seen": 2816802816
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000149010101010101,
      "loss": 2.5857,
      "theoretical_loss": 3.336985345885021,
      "tokens_seen": 2816868352
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014898989898989897,
      "loss": 2.3817,
      "theoretical_loss": 3.336979300067795,
      "tokens_seen": 2816933888
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014896969696969698,
      "loss": 2.3698,
      "theoretical_loss": 3.336973254430606,
      "tokens_seen": 2816999424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014894949494949494,
      "loss": 2.6252,
      "theoretical_loss": 3.3369672089734443,
      "tokens_seen": 2817064960
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014892929292929295,
      "loss": 2.6003,
      "theoretical_loss": 3.3369611636963006,
      "tokens_seen": 2817130496
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001489090909090909,
      "loss": 2.3109,
      "theoretical_loss": 3.3369551185991653,
      "tokens_seen": 2817196032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001488888888888889,
      "loss": 2.5003,
      "theoretical_loss": 3.3369490736820286,
      "tokens_seen": 2817261568
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014886868686868687,
      "loss": 2.5494,
      "theoretical_loss": 3.3369430289448814,
      "tokens_seen": 2817327104
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014884848484848486,
      "loss": 2.548,
      "theoretical_loss": 3.336936984387714,
      "tokens_seen": 2817392640
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014882828282828284,
      "loss": 2.5223,
      "theoretical_loss": 3.3369309400105163,
      "tokens_seen": 2817458176
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001488080808080808,
      "loss": 2.5994,
      "theoretical_loss": 3.33692489581328,
      "tokens_seen": 2817523712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014878787878787878,
      "loss": 2.4584,
      "theoretical_loss": 3.3369188517959945,
      "tokens_seen": 2817589248
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014876767676767676,
      "loss": 2.3667,
      "theoretical_loss": 3.33691280795865,
      "tokens_seen": 2817654784
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014874747474747475,
      "loss": 2.5087,
      "theoretical_loss": 3.3369067643012382,
      "tokens_seen": 2817720320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014872727272727273,
      "loss": 2.5249,
      "theoretical_loss": 3.3369007208237482,
      "tokens_seen": 2817785856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014870707070707072,
      "loss": 2.5113,
      "theoretical_loss": 3.3368946775261716,
      "tokens_seen": 2817851392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001486868686868687,
      "loss": 2.6231,
      "theoretical_loss": 3.3368886344084983,
      "tokens_seen": 2817916928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014866666666666668,
      "loss": 2.4657,
      "theoretical_loss": 3.3368825914707188,
      "tokens_seen": 2817982464
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014864646464646464,
      "loss": 2.452,
      "theoretical_loss": 3.336876548712824,
      "tokens_seen": 2818048000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014862626262626263,
      "loss": 2.5397,
      "theoretical_loss": 3.3368705061348036,
      "tokens_seen": 2818113536
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001486060606060606,
      "loss": 2.6348,
      "theoretical_loss": 3.336864463736648,
      "tokens_seen": 2818179072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001485858585858586,
      "loss": 2.6421,
      "theoretical_loss": 3.3368584215183485,
      "tokens_seen": 2818244608
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1545329,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4824910163879395,
      "objective/train/theoretical_loss": 3.3368523794798954,
      "objective/train/tokens_used": 1177169376,
      "theoretical_loss": 3.3368523794798954,
      "tokens_seen": 2818310144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014856565656565658,
      "loss": 2.4511,
      "theoretical_loss": 3.3368523794798954,
      "tokens_seen": 2818310144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014854545454545453,
      "loss": 2.5684,
      "theoretical_loss": 3.336846337621279,
      "tokens_seen": 2818375680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014852525252525254,
      "loss": 2.6883,
      "theoretical_loss": 3.3368402959424888,
      "tokens_seen": 2818441216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001485050505050505,
      "loss": 2.5521,
      "theoretical_loss": 3.336834254443517,
      "tokens_seen": 2818506752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001484848484848485,
      "loss": 2.7295,
      "theoretical_loss": 3.336828213124353,
      "tokens_seen": 2818572288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014846464646464647,
      "loss": 2.3323,
      "theoretical_loss": 3.3368221719849873,
      "tokens_seen": 2818637824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014844444444444443,
      "loss": 2.6035,
      "theoretical_loss": 3.3368161310254107,
      "tokens_seen": 2818703360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014842424242424244,
      "loss": 2.4436,
      "theoretical_loss": 3.3368100902456135,
      "tokens_seen": 2818768896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001484040404040404,
      "loss": 2.4129,
      "theoretical_loss": 3.3368040496455866,
      "tokens_seen": 2818834432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001483838383838384,
      "loss": 2.5415,
      "theoretical_loss": 3.3367980092253196,
      "tokens_seen": 2818899968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014836363636363636,
      "loss": 2.2387,
      "theoretical_loss": 3.3367919689848033,
      "tokens_seen": 2818965504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014834343434343435,
      "loss": 2.6121,
      "theoretical_loss": 3.3367859289240287,
      "tokens_seen": 2819031040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014832323232323233,
      "loss": 2.6852,
      "theoretical_loss": 3.336779889042986,
      "tokens_seen": 2819096576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001483030303030303,
      "loss": 2.681,
      "theoretical_loss": 3.3367738493416654,
      "tokens_seen": 2819162112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001482828282828283,
      "loss": 2.3518,
      "theoretical_loss": 3.3367678098200573,
      "tokens_seen": 2819227648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014826262626262626,
      "loss": 2.4424,
      "theoretical_loss": 3.3367617704781525,
      "tokens_seen": 2819293184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014824242424242424,
      "loss": 2.6784,
      "theoretical_loss": 3.3367557313159413,
      "tokens_seen": 2819358720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014822222222222222,
      "loss": 2.4099,
      "theoretical_loss": 3.336749692333415,
      "tokens_seen": 2819424256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001482020202020202,
      "loss": 2.6596,
      "theoretical_loss": 3.3367436535305623,
      "tokens_seen": 2819489792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001481818181818182,
      "loss": 2.4806,
      "theoretical_loss": 3.336737614907375,
      "tokens_seen": 2819555328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014816161616161617,
      "loss": 2.7067,
      "theoretical_loss": 3.3367315764638437,
      "tokens_seen": 2819620864
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014814141414141413,
      "loss": 2.4387,
      "theoretical_loss": 3.336725538199958,
      "tokens_seen": 2819686400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014812121212121212,
      "loss": 2.3936,
      "theoretical_loss": 3.3367195001157093,
      "tokens_seen": 2819751936
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001481010101010101,
      "loss": 2.6026,
      "theoretical_loss": 3.3367134622110877,
      "tokens_seen": 2819817472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014808080808080808,
      "loss": 2.331,
      "theoretical_loss": 3.3367074244860833,
      "tokens_seen": 2819883008
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1546518,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9841865301132202,
      "objective/train/theoretical_loss": 3.3367013869406867,
      "objective/train/tokens_used": 1178807776,
      "theoretical_loss": 3.3367013869406867,
      "tokens_seen": 2819948544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014806060606060607,
      "loss": 2.4183,
      "theoretical_loss": 3.3367013869406867,
      "tokens_seen": 2819948544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014804040404040402,
      "loss": 2.4696,
      "theoretical_loss": 3.336695349574889,
      "tokens_seen": 2820014080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014802020202020204,
      "loss": 2.5295,
      "theoretical_loss": 3.33668931238868,
      "tokens_seen": 2820079616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000148,
      "loss": 2.4702,
      "theoretical_loss": 3.3366832753820503,
      "tokens_seen": 2820145152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000147979797979798,
      "loss": 2.3708,
      "theoretical_loss": 3.3366772385549908,
      "tokens_seen": 2820210688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014795959595959596,
      "loss": 2.46,
      "theoretical_loss": 3.336671201907492,
      "tokens_seen": 2820276224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014793939393939392,
      "loss": 2.492,
      "theoretical_loss": 3.336665165439544,
      "tokens_seen": 2820341760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014791919191919193,
      "loss": 2.5065,
      "theoretical_loss": 3.336659129151137,
      "tokens_seen": 2820407296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014789898989898989,
      "loss": 2.718,
      "theoretical_loss": 3.336653093042262,
      "tokens_seen": 2820472832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001478787878787879,
      "loss": 2.4638,
      "theoretical_loss": 3.3366470571129097,
      "tokens_seen": 2820538368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014785858585858585,
      "loss": 2.3556,
      "theoretical_loss": 3.3366410213630697,
      "tokens_seen": 2820603904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014783838383838386,
      "loss": 2.583,
      "theoretical_loss": 3.3366349857927333,
      "tokens_seen": 2820669440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014781818181818182,
      "loss": 2.4604,
      "theoretical_loss": 3.336628950401891,
      "tokens_seen": 2820734976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001477979797979798,
      "loss": 2.3676,
      "theoretical_loss": 3.336622915190533,
      "tokens_seen": 2820800512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001477777777777778,
      "loss": 2.3986,
      "theoretical_loss": 3.3366168801586498,
      "tokens_seen": 2820866048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014775757575757575,
      "loss": 2.561,
      "theoretical_loss": 3.3366108453062315,
      "tokens_seen": 2820931584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014773737373737376,
      "loss": 2.5619,
      "theoretical_loss": 3.3366048106332697,
      "tokens_seen": 2820997120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014771717171717171,
      "loss": 2.4221,
      "theoretical_loss": 3.3365987761397538,
      "tokens_seen": 2821062656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001476969696969697,
      "loss": 2.4276,
      "theoretical_loss": 3.3365927418256747,
      "tokens_seen": 2821128192
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014767676767676768,
      "loss": 2.5575,
      "theoretical_loss": 3.336586707691023,
      "tokens_seen": 2821193728
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014765656565656567,
      "loss": 2.6101,
      "theoretical_loss": 3.336580673735789,
      "tokens_seen": 2821259264
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014763636363636365,
      "loss": 2.2437,
      "theoretical_loss": 3.3365746399599634,
      "tokens_seen": 2821324800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014761616161616163,
      "loss": 2.4829,
      "theoretical_loss": 3.336568606363537,
      "tokens_seen": 2821390336
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001475959595959596,
      "loss": 2.4354,
      "theoretical_loss": 3.3365625729464994,
      "tokens_seen": 2821455872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014757575757575757,
      "loss": 2.5635,
      "theoretical_loss": 3.3365565397088415,
      "tokens_seen": 2821521408
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1547222,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7162699699401855,
      "objective/train/theoretical_loss": 3.336550506650554,
      "objective/train/tokens_used": 1180446176,
      "theoretical_loss": 3.336550506650554,
      "tokens_seen": 2821586944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014755555555555556,
      "loss": 2.6892,
      "theoretical_loss": 3.336550506650554,
      "tokens_seen": 2821586944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014753535353535354,
      "loss": 2.4847,
      "theoretical_loss": 3.336544473771627,
      "tokens_seen": 2821652480
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014751515151515153,
      "loss": 2.577,
      "theoretical_loss": 3.336538441072052,
      "tokens_seen": 2821718016
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014749494949494948,
      "loss": 2.6168,
      "theoretical_loss": 3.3365324085518187,
      "tokens_seen": 2821783552
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001474747474747475,
      "loss": 2.5693,
      "theoretical_loss": 3.3365263762109176,
      "tokens_seen": 2821849088
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014745454545454545,
      "loss": 2.4566,
      "theoretical_loss": 3.336520344049339,
      "tokens_seen": 2821914624
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014743434343434346,
      "loss": 2.4969,
      "theoretical_loss": 3.3365143120670737,
      "tokens_seen": 2821980160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014741414141414142,
      "loss": 2.471,
      "theoretical_loss": 3.3365082802641126,
      "tokens_seen": 2822045696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014739393939393938,
      "loss": 2.349,
      "theoretical_loss": 3.3365022486404454,
      "tokens_seen": 2822111232
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001473737373737374,
      "loss": 2.4419,
      "theoretical_loss": 3.3364962171960633,
      "tokens_seen": 2822176768
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014735353535353534,
      "loss": 2.6028,
      "theoretical_loss": 3.336490185930957,
      "tokens_seen": 2822242304
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014733333333333335,
      "loss": 2.4462,
      "theoretical_loss": 3.336484154845116,
      "tokens_seen": 2822307840
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001473131313131313,
      "loss": 2.3973,
      "theoretical_loss": 3.336478123938531,
      "tokens_seen": 2822373376
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001472929292929293,
      "loss": 2.2181,
      "theoretical_loss": 3.3364720932111935,
      "tokens_seen": 2822438912
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014727272727272728,
      "loss": 2.4296,
      "theoretical_loss": 3.336466062663093,
      "tokens_seen": 2822504448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014725252525252524,
      "loss": 2.5094,
      "theoretical_loss": 3.336460032294221,
      "tokens_seen": 2822569984
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014723232323232325,
      "loss": 2.5197,
      "theoretical_loss": 3.336454002104567,
      "tokens_seen": 2822635520
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001472121212121212,
      "loss": 2.3919,
      "theoretical_loss": 3.336447972094122,
      "tokens_seen": 2822701056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001471919191919192,
      "loss": 2.4516,
      "theoretical_loss": 3.336441942262876,
      "tokens_seen": 2822766592
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014717171717171717,
      "loss": 2.5135,
      "theoretical_loss": 3.3364359126108205,
      "tokens_seen": 2822832128
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014715151515151516,
      "loss": 2.459,
      "theoretical_loss": 3.3364298831379453,
      "tokens_seen": 2822897664
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014713131313131314,
      "loss": 2.2841,
      "theoretical_loss": 3.336423853844241,
      "tokens_seen": 2822963200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014711111111111112,
      "loss": 2.418,
      "theoretical_loss": 3.3364178247296983,
      "tokens_seen": 2823028736
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014709090909090908,
      "loss": 2.448,
      "theoretical_loss": 3.3364117957943074,
      "tokens_seen": 2823094272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014707070707070706,
      "loss": 2.5341,
      "theoretical_loss": 3.336405767038059,
      "tokens_seen": 2823159808
    },
    {
      "epoch": 0.71,
      "objective/train/docs_used": 1548263,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.856104612350464,
      "objective/train/theoretical_loss": 3.336399738460944,
      "objective/train/tokens_used": 1182084576,
      "theoretical_loss": 3.336399738460944,
      "tokens_seen": 2823225344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014705050505050505,
      "loss": 2.5434,
      "theoretical_loss": 3.336399738460944,
      "tokens_seen": 2823225344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014703030303030303,
      "loss": 2.5174,
      "theoretical_loss": 3.336393710062952,
      "tokens_seen": 2823290880
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014701010101010102,
      "loss": 2.4975,
      "theoretical_loss": 3.3363876818440747,
      "tokens_seen": 2823356416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014698989898989897,
      "loss": 2.649,
      "theoretical_loss": 3.3363816538043016,
      "tokens_seen": 2823421952
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014696969696969698,
      "loss": 2.6029,
      "theoretical_loss": 3.336375625943624,
      "tokens_seen": 2823487488
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014694949494949494,
      "loss": 2.4688,
      "theoretical_loss": 3.336369598262032,
      "tokens_seen": 2823553024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014692929292929295,
      "loss": 2.4812,
      "theoretical_loss": 3.3363635707595156,
      "tokens_seen": 2823618560
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001469090909090909,
      "loss": 2.4266,
      "theoretical_loss": 3.3363575434360664,
      "tokens_seen": 2823684096
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001468888888888889,
      "loss": 2.5619,
      "theoretical_loss": 3.3363515162916744,
      "tokens_seen": 2823749632
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014686868686868688,
      "loss": 2.3878,
      "theoretical_loss": 3.3363454893263302,
      "tokens_seen": 2823815168
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014684848484848483,
      "loss": 2.4137,
      "theoretical_loss": 3.3363394625400242,
      "tokens_seen": 2823880704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014682828282828285,
      "loss": 2.51,
      "theoretical_loss": 3.3363334359327466,
      "tokens_seen": 2823946240
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001468080808080808,
      "loss": 2.3203,
      "theoretical_loss": 3.336327409504489,
      "tokens_seen": 2824011776
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001467878787878788,
      "loss": 2.5331,
      "theoretical_loss": 3.3363213832552407,
      "tokens_seen": 2824077312
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014676767676767677,
      "loss": 2.484,
      "theoretical_loss": 3.336315357184993,
      "tokens_seen": 2824142848
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014674747474747475,
      "loss": 2.5866,
      "theoretical_loss": 3.3363093312937364,
      "tokens_seen": 2824208384
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014672727272727274,
      "loss": 2.4145,
      "theoretical_loss": 3.336303305581461,
      "tokens_seen": 2824273920
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001467070707070707,
      "loss": 2.4557,
      "theoretical_loss": 3.3362972800481576,
      "tokens_seen": 2824339456
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001466868686868687,
      "loss": 2.5438,
      "theoretical_loss": 3.336291254693817,
      "tokens_seen": 2824404992
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014666666666666666,
      "loss": 2.5279,
      "theoretical_loss": 3.336285229518429,
      "tokens_seen": 2824470528
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014664646464646465,
      "loss": 2.4785,
      "theoretical_loss": 3.3362792045219845,
      "tokens_seen": 2824536064
    }
  ],
  "max_steps": 25000,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 5.9348826980352e+17,
  "trial_name": null,
  "trial_params": null
}