{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.19859395480001588,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-06,
      "loss": 4.3952,
      "theoretical_loss": 17.566201104328645,
      "tokens_seen": 131072
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-06,
      "loss": 4.3745,
      "theoretical_loss": 14.89231675598857,
      "tokens_seen": 262144
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.9523809523809525e-06,
      "loss": 4.1183,
      "theoretical_loss": 13.552561472550224,
      "tokens_seen": 393216
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507936e-06,
      "loss": 3.9369,
      "theoretical_loss": 12.690129625483323,
      "tokens_seen": 524288
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-06,
      "loss": 4.0965,
      "theoretical_loss": 12.067412607035077,
      "tokens_seen": 655360
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.1904761904761905e-05,
      "loss": 3.5987,
      "theoretical_loss": 11.586719208706729,
      "tokens_seen": 786432
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.3888888888888888e-05,
      "loss": 3.6811,
      "theoretical_loss": 11.199011702111871,
      "tokens_seen": 917504
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.5873015873015872e-05,
      "loss": 3.7779,
      "theoretical_loss": 10.87642808645695,
      "tokens_seen": 1048576
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.7857142857142855e-05,
      "loss": 3.8919,
      "theoretical_loss": 10.60172987623028,
      "tokens_seen": 1179648
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-05,
      "loss": 4.0009,
      "theoretical_loss": 10.36356394376333,
      "tokens_seen": 1310720
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.1825396825396824e-05,
      "loss": 3.75,
      "theoretical_loss": 10.15408655327002,
      "tokens_seen": 1441792
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.380952380952381e-05,
      "loss": 3.8353,
      "theoretical_loss": 9.967669178840278,
      "tokens_seen": 1572864
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.36052054166793823,
      "objective/train/avg_token_score": 0.01564830169081688,
      "objective/train/docs_used": 13112,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 3.4145798683166504,
      "objective/train/original_loss": 9.366209983825684,
      "objective/train/theoretical_loss": 9.881763126393109,
      "objective/train/tokens_used": 1638400,
      "objective/train/unlikelihood_loss": 0.004730077460408211,
      "theoretical_loss": 9.881763126393109,
      "tokens_seen": 1638400
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.5793650793650793e-05,
      "loss": 3.9383,
      "theoretical_loss": 9.80014659154056,
      "tokens_seen": 1703936
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.7777777777777776e-05,
      "loss": 3.6506,
      "theoretical_loss": 9.648356759081546,
      "tokens_seen": 1835008
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.9761904761904762e-05,
      "loss": 3.5887,
      "theoretical_loss": 9.509847046764852,
      "tokens_seen": 1966080
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.1746031746031745e-05,
      "loss": 3.7839,
      "theoretical_loss": 9.382679790910457,
      "tokens_seen": 2097152
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.373015873015873e-05,
      "loss": 3.3506,
      "theoretical_loss": 9.265299666660276,
      "tokens_seen": 2228224
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.571428571428571e-05,
      "loss": 3.6369,
      "theoretical_loss": 9.156440812508292,
      "tokens_seen": 2359296
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.76984126984127e-05,
      "loss": 3.1405,
      "theoretical_loss": 9.055060296533734,
      "tokens_seen": 2490368
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-05,
      "loss": 3.5961,
      "theoretical_loss": 8.960289489909357,
      "tokens_seen": 2621440
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.1666666666666665e-05,
      "loss": 3.8762,
      "theoretical_loss": 8.871397892478225,
      "tokens_seen": 2752512
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.365079365079365e-05,
      "loss": 3.1495,
      "theoretical_loss": 8.787765792778412,
      "tokens_seen": 2883584
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.563492063492063e-05,
      "loss": 3.2901,
      "theoretical_loss": 8.708863309520833,
      "tokens_seen": 3014656
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.761904761904762e-05,
      "loss": 3.7481,
      "theoretical_loss": 8.634234117735474,
      "tokens_seen": 3145728
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.3303396999835968,
      "objective/train/avg_token_score": 0.04211259260773659,
      "objective/train/docs_used": 13755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.995987892150879,
      "objective/train/original_loss": 9.143326759338379,
      "objective/train/theoretical_loss": 8.563482664611069,
      "objective/train/tokens_used": 3276800,
      "objective/train/unlikelihood_loss": 0.01486299093812704,
      "theoretical_loss": 8.563482664611069,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.96031746031746e-05,
      "loss": 3.5744,
      "theoretical_loss": 8.563482664611069,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.1587301587301586e-05,
      "loss": 3.5767,
      "theoretical_loss": 8.496264019646002,
      "tokens_seen": 3407872
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.357142857142857e-05,
      "loss": 3.0622,
      "theoretical_loss": 8.432275737672779,
      "tokens_seen": 3538944
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.555555555555555e-05,
      "loss": 3.7336,
      "theoretical_loss": 8.371251277120209,
      "tokens_seen": 3670016
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.753968253968254e-05,
      "loss": 3.411,
      "theoretical_loss": 8.31295463228533,
      "tokens_seen": 3801088
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.9523809523809524e-05,
      "loss": 3.38,
      "theoretical_loss": 8.257175922251864,
      "tokens_seen": 3932160
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.15079365079365e-05,
      "loss": 3.1689,
      "theoretical_loss": 8.20372774027797,
      "tokens_seen": 4063232
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.349206349206349e-05,
      "loss": 3.3772,
      "theoretical_loss": 8.152442112639616,
      "tokens_seen": 4194304
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.547619047619048e-05,
      "loss": 2.919,
      "theoretical_loss": 8.10316794961571,
      "tokens_seen": 4325376
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.746031746031745e-05,
      "loss": 3.1994,
      "theoretical_loss": 8.055768896701416,
      "tokens_seen": 4456448
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.944444444444444e-05,
      "loss": 3.6138,
      "theoretical_loss": 8.010121513461836,
      "tokens_seen": 4587520
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.142857142857142e-05,
      "loss": 3.0735,
      "theoretical_loss": 7.966113722271801,
      "tokens_seen": 4718592
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.341269841269842e-05,
      "loss": 3.063,
      "theoretical_loss": 7.9236434806675184,
      "tokens_seen": 4849664
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.3146994113922119,
      "objective/train/avg_token_score": 0.016941465437412262,
      "objective/train/docs_used": 15099,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.699511766433716,
      "objective/train/original_loss": 8.369515419006348,
      "objective/train/theoretical_loss": 7.902955512484067,
      "objective/train/tokens_used": 4915200,
      "objective/train/unlikelihood_loss": 0.032859548926353455,
      "theoretical_loss": 7.902955512484067,
      "tokens_seen": 4915200
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.53968253968254e-05,
      "loss": 3.196,
      "theoretical_loss": 7.882617639989203,
      "tokens_seen": 4980736
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.738095238095239e-05,
      "loss": 2.9935,
      "theoretical_loss": 7.842950960027937,
      "tokens_seen": 5111808
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507937e-05,
      "loss": 3.4802,
      "theoretical_loss": 7.804565254954165,
      "tokens_seen": 5242880
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.134920634920635e-05,
      "loss": 3.2085,
      "theoretical_loss": 7.767388650235364,
      "tokens_seen": 5373952
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.333333333333333e-05,
      "loss": 3.3755,
      "theoretical_loss": 7.731354933799318,
      "tokens_seen": 5505024
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.531746031746032e-05,
      "loss": 3.6461,
      "theoretical_loss": 7.696402987558934,
      "tokens_seen": 5636096
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.73015873015873e-05,
      "loss": 3.0517,
      "theoretical_loss": 7.662476287731328,
      "tokens_seen": 5767168
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.928571428571429e-05,
      "loss": 3.0863,
      "theoretical_loss": 7.629522464270861,
      "tokens_seen": 5898240
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.126984126984126e-05,
      "loss": 2.7058,
      "theoretical_loss": 7.59749291128028,
      "tokens_seen": 6029312
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.325396825396825e-05,
      "loss": 3.3358,
      "theoretical_loss": 7.5663424415343705,
      "tokens_seen": 6160384
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.523809523809524e-05,
      "loss": 3.0732,
      "theoretical_loss": 7.536028979299919,
      "tokens_seen": 6291456
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.722222222222223e-05,
      "loss": 3.0706,
      "theoretical_loss": 7.506513286506497,
      "tokens_seen": 6422528
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.41287878155708313,
      "objective/train/avg_token_score": 0.03952885419130325,
      "objective/train/docs_used": 15777,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 3.0034613609313965,
      "objective/train/original_loss": 7.385845184326172,
      "objective/train/theoretical_loss": 7.4777587180480305,
      "objective/train/tokens_used": 6553600,
      "objective/train/unlikelihood_loss": 0.09759358316659927,
      "theoretical_loss": 7.4777587180480305,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-05,
      "loss": 3.0284,
      "theoretical_loss": 7.4777587180480305,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001011904761904762,
      "loss": 2.8636,
      "theoretical_loss": 7.449731002601916,
      "tokens_seen": 6684672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010317460317460317,
      "loss": 2.724,
      "theoretical_loss": 7.422398045861905,
      "tokens_seen": 6815744
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010515873015873016,
      "loss": 2.9955,
      "theoretical_loss": 7.395729753510345,
      "tokens_seen": 6946816
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010714285714285714,
      "loss": 2.967,
      "theoretical_loss": 7.369697871618373,
      "tokens_seen": 7077888
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010912698412698413,
      "loss": 3.2671,
      "theoretical_loss": 7.3442758424706875,
      "tokens_seen": 7208960
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001111111111111111,
      "loss": 3.0036,
      "theoretical_loss": 7.319438674073677,
      "tokens_seen": 7340032
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011309523809523809,
      "loss": 3.0592,
      "theoretical_loss": 7.295162821829564,
      "tokens_seen": 7471104
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011507936507936508,
      "loss": 3.2368,
      "theoretical_loss": 7.271426081050832,
      "tokens_seen": 7602176
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011706349206349207,
      "loss": 3.1647,
      "theoretical_loss": 7.24820748915387,
      "tokens_seen": 7733248
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011904761904761905,
      "loss": 2.9356,
      "theoretical_loss": 7.225487236512497,
      "tokens_seen": 7864320
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012103174603174604,
      "loss": 2.8582,
      "theoretical_loss": 7.2032465850744005,
      "tokens_seen": 7995392
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000123015873015873,
      "loss": 2.7811,
      "theoretical_loss": 7.1814677939495155,
      "tokens_seen": 8126464
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.4550953209400177,
      "objective/train/avg_token_score": 0.02127336896955967,
      "objective/train/docs_used": 17166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 3.3246047496795654,
      "objective/train/original_loss": 7.3253278732299805,
      "objective/train/theoretical_loss": 7.1707463077509646,
      "objective/train/tokens_used": 8192000,
      "objective/train/unlikelihood_loss": 0.044195178896188736,
      "theoretical_loss": 7.1707463077509646,
      "tokens_seen": 8192000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000125,
      "loss": 3.3126,
      "theoretical_loss": 7.160134051271272,
      "tokens_seen": 8257536
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012698412698412698,
      "loss": 3.0628,
      "theoretical_loss": 7.139229411711638,
      "tokens_seen": 8388608
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012896825396825398,
      "loss": 2.7911,
      "theoretical_loss": 7.118738739100616,
      "tokens_seen": 8519680
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013095238095238096,
      "loss": 2.8125,
      "theoretical_loss": 7.09864765366177,
      "tokens_seen": 8650752
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013293650793650793,
      "loss": 2.6533,
      "theoretical_loss": 7.078942483428749,
      "tokens_seen": 8781824
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001349206349206349,
      "loss": 2.9263,
      "theoretical_loss": 7.059610219454568,
      "tokens_seen": 8912896
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001369047619047619,
      "loss": 2.8425,
      "theoretical_loss": 7.040638474466625,
      "tokens_seen": 9043968
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001388888888888889,
      "loss": 3.1806,
      "theoretical_loss": 7.022015444656678,
      "tokens_seen": 9175040
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014087301587301586,
      "loss": 3.0648,
      "theoretical_loss": 7.003729874327071,
      "tokens_seen": 9306112
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014285714285714284,
      "loss": 3.0663,
      "theoretical_loss": 6.98577102314278,
      "tokens_seen": 9437184
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014484126984126984,
      "loss": 2.7552,
      "theoretical_loss": 6.968128635764015,
      "tokens_seen": 9568256
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014682539682539685,
      "loss": 3.0583,
      "theoretical_loss": 6.950792913656309,
      "tokens_seen": 9699328
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.36840176582336426,
      "objective/train/avg_token_score": 0.020729999989271164,
      "objective/train/docs_used": 17861,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.664196014404297,
      "objective/train/original_loss": 7.552342891693115,
      "objective/train/theoretical_loss": 6.9337544888949,
      "objective/train/tokens_used": 9830400,
      "objective/train/unlikelihood_loss": 0.06143795698881149,
      "theoretical_loss": 6.9337544888949,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014880952380952382,
      "loss": 2.7889,
      "theoretical_loss": 6.9337544888949,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001507936507936508,
      "loss": 2.387,
      "theoretical_loss": 6.917004399797798,
      "tokens_seen": 9961472
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001527777777777778,
      "loss": 2.846,
      "theoretical_loss": 6.900534068237688,
      "tokens_seen": 10092544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015476190476190478,
      "loss": 2.8487,
      "theoretical_loss": 6.884335278496871,
      "tokens_seen": 10223616
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015674603174603175,
      "loss": 2.8823,
      "theoretical_loss": 6.868400157541997,
      "tokens_seen": 10354688
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015873015873015873,
      "loss": 2.799,
      "theoretical_loss": 6.85272115660663,
      "tokens_seen": 10485760
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016071428571428573,
      "loss": 2.9633,
      "theoretical_loss": 6.8372910339797945,
      "tokens_seen": 10616832
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001626984126984127,
      "loss": 2.5692,
      "theoretical_loss": 6.8221028389077185,
      "tokens_seen": 10747904
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016468253968253969,
      "loss": 2.8166,
      "theoretical_loss": 6.807149896524181,
      "tokens_seen": 10878976
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016666666666666666,
      "loss": 3.0421,
      "theoretical_loss": 6.7924257937322245,
      "tokens_seen": 11010048
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016865079365079366,
      "loss": 2.6027,
      "theoretical_loss": 6.777924365966638,
      "tokens_seen": 11141120
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017063492063492064,
      "loss": 2.9229,
      "theoretical_loss": 6.763639684772625,
      "tokens_seen": 11272192
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017261904761904762,
      "loss": 2.8946,
      "theoretical_loss": 6.749566046141486,
      "tokens_seen": 11403264
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.4332233667373657,
      "objective/train/avg_token_score": 0.019017228856682777,
      "objective/train/docs_used": 18493,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.7806520462036133,
      "objective/train/original_loss": 7.158781051635742,
      "objective/train/theoretical_loss": 6.7426066439450905,
      "objective/train/tokens_used": 11468800,
      "objective/train/unlikelihood_loss": 0.05772477760910988,
      "theoretical_loss": 6.7426066439450905,
      "tokens_seen": 11468800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001746031746031746,
      "loss": 2.8532,
      "theoretical_loss": 6.735697959549075,
      "tokens_seen": 11534336
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001765873015873016,
      "loss": 2.5473,
      "theoretical_loss": 6.722030137647226,
      "tokens_seen": 11665408
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017857142857142857,
      "loss": 2.4778,
      "theoretical_loss": 6.7085574865624125,
      "tokens_seen": 11796480
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018055555555555555,
      "loss": 2.9024,
      "theoretical_loss": 6.695275096759559,
      "tokens_seen": 11927552
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018253968253968252,
      "loss": 2.7962,
      "theoretical_loss": 6.682178234432274,
      "tokens_seen": 12058624
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018452380952380953,
      "loss": 2.6879,
      "theoretical_loss": 6.669262333383815,
      "tokens_seen": 12189696
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001865079365079365,
      "loss": 2.9128,
      "theoretical_loss": 6.656522987365879,
      "tokens_seen": 12320768
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001884920634920635,
      "loss": 2.4778,
      "theoretical_loss": 6.643955942844831,
      "tokens_seen": 12451840
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019047619047619048,
      "loss": 2.5956,
      "theoretical_loss": 6.631557092167304,
      "tokens_seen": 12582912
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019246031746031748,
      "loss": 2.8293,
      "theoretical_loss": 6.619322467099223,
      "tokens_seen": 12713984
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019444444444444446,
      "loss": 3.043,
      "theoretical_loss": 6.607248232714213,
      "tokens_seen": 12845056
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019642857142857144,
      "loss": 2.7412,
      "theoretical_loss": 6.59533068160918,
      "tokens_seen": 12976128
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.4409824013710022,
      "objective/train/avg_token_score": 0.007921183481812477,
      "objective/train/docs_used": 19731,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.8607468605041504,
      "objective/train/original_loss": 5.793218612670898,
      "objective/train/theoretical_loss": 6.583566228426414,
      "objective/train/tokens_used": 13107200,
      "objective/train/unlikelihood_loss": 0.14077132940292358,
      "theoretical_loss": 6.583566228426414,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001984126984126984,
      "loss": 2.5021,
      "theoretical_loss": 6.583566228426414,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020039682539682542,
      "loss": 2.5258,
      "theoretical_loss": 6.571951404663098,
      "tokens_seen": 13238272
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002023809523809524,
      "loss": 2.8582,
      "theoretical_loss": 6.560482853750463,
      "tokens_seen": 13369344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020436507936507937,
      "loss": 2.828,
      "theoretical_loss": 6.549157326386091,
      "tokens_seen": 13500416
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020634920634920634,
      "loss": 2.8331,
      "theoretical_loss": 6.537971676104026,
      "tokens_seen": 13631488
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020833333333333335,
      "loss": 2.0467,
      "theoretical_loss": 6.5269228550684195,
      "tokens_seen": 13762560
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021031746031746032,
      "loss": 2.6811,
      "theoretical_loss": 6.516007910077416,
      "tokens_seen": 13893632
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002123015873015873,
      "loss": 2.7855,
      "theoretical_loss": 6.50522397876491,
      "tokens_seen": 14024704
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021428571428571427,
      "loss": 2.7877,
      "theoretical_loss": 6.494568285988618,
      "tokens_seen": 14155776
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021626984126984128,
      "loss": 2.624,
      "theoretical_loss": 6.484038140393699,
      "tokens_seen": 14286848
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021825396825396825,
      "loss": 2.7998,
      "theoretical_loss": 6.473630931141869,
      "tokens_seen": 14417920
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022023809523809523,
      "loss": 2.71,
      "theoretical_loss": 6.463344124796616,
      "tokens_seen": 14548992
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002222222222222222,
      "loss": 2.5073,
      "theoretical_loss": 6.45317526235573,
      "tokens_seen": 14680064
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.49395161867141724,
      "objective/train/avg_token_score": 0.028753770515322685,
      "objective/train/docs_used": 20428,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 3.2851853370666504,
      "objective/train/original_loss": 6.822603225708008,
      "objective/train/theoretical_loss": 6.448134311315593,
      "objective/train/tokens_used": 14745600,
      "objective/train/unlikelihood_loss": 0.016451282426714897,
      "theoretical_loss": 6.448134311315593,
      "tokens_seen": 14745600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002242063492063492,
      "loss": 2.5962,
      "theoretical_loss": 6.443121956422939,
      "tokens_seen": 14811136
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022619047619047618,
      "loss": 2.6017,
      "theoretical_loss": 6.433181888510964,
      "tokens_seen": 14942208
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022817460317460316,
      "loss": 2.7709,
      "theoretical_loss": 6.4233528064687855,
      "tokens_seen": 15073280
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023015873015873016,
      "loss": 2.5309,
      "theoretical_loss": 6.413632522026391,
      "tokens_seen": 15204352
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023214285714285717,
      "loss": 2.5149,
      "theoretical_loss": 6.404018908450656,
      "tokens_seen": 15335424
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023412698412698414,
      "loss": 2.5856,
      "theoretical_loss": 6.394509898306452,
      "tokens_seen": 15466496
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023611111111111112,
      "loss": 2.5106,
      "theoretical_loss": 6.385103481317387,
      "tokens_seen": 15597568
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002380952380952381,
      "loss": 2.559,
      "theoretical_loss": 6.375797702320966,
      "tokens_seen": 15728640
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002400793650793651,
      "loss": 2.2329,
      "theoretical_loss": 6.366590659313248,
      "tokens_seen": 15859712
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024206349206349207,
      "loss": 2.6909,
      "theoretical_loss": 6.357480501578371,
      "tokens_seen": 15990784
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024404761904761905,
      "loss": 2.6294,
      "theoretical_loss": 6.348465427898629,
      "tokens_seen": 16121856
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000246031746031746,
      "loss": 2.464,
      "theoretical_loss": 6.33954368484097,
      "tokens_seen": 16252928
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.39424487948417664,
      "objective/train/avg_token_score": 0.013139091432094574,
      "objective/train/docs_used": 21641,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.531733751296997,
      "objective/train/original_loss": 6.893299102783203,
      "objective/train/theoretical_loss": 6.330713565116083,
      "objective/train/tokens_used": 16384000,
      "objective/train/unlikelihood_loss": 0.03236621245741844,
      "theoretical_loss": 6.330713565116083,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000248015873015873,
      "loss": 2.6212,
      "theoretical_loss": 6.330713565116083,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025,
      "loss": 2.7361,
      "theoretical_loss": 6.32197340600647,
      "tokens_seen": 16515072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000251984126984127,
      "loss": 2.4975,
      "theoretical_loss": 6.313321587860021,
      "tokens_seen": 16646144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025396825396825396,
      "loss": 2.75,
      "theoretical_loss": 6.304756532645939,
      "tokens_seen": 16777216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025595238095238096,
      "loss": 2.38,
      "theoretical_loss": 6.296276702569918,
      "tokens_seen": 16908288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025793650793650796,
      "loss": 2.4812,
      "theoretical_loss": 6.28788059874573,
      "tokens_seen": 17039360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025992063492063497,
      "loss": 2.5431,
      "theoretical_loss": 6.279566759920507,
      "tokens_seen": 17170432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002619047619047619,
      "loss": 2.9019,
      "theoretical_loss": 6.271333761251142,
      "tokens_seen": 17301504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002638888888888889,
      "loss": 2.454,
      "theoretical_loss": 6.2631802131294085,
      "tokens_seen": 17432576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026587301587301587,
      "loss": 2.6175,
      "theoretical_loss": 6.255104760053497,
      "tokens_seen": 17563648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026785714285714287,
      "loss": 2.6316,
      "theoretical_loss": 6.247106079543801,
      "tokens_seen": 17694720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002698412698412698,
      "loss": 2.4119,
      "theoretical_loss": 6.239182881100916,
      "tokens_seen": 17825792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002718253968253968,
      "loss": 2.7574,
      "theoretical_loss": 6.231333905203899,
      "tokens_seen": 17956864
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.3819648027420044,
      "objective/train/avg_token_score": 0.008278077468276024,
      "objective/train/docs_used": 22253,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.420483112335205,
      "objective/train/original_loss": 6.370981693267822,
      "objective/train/theoretical_loss": 6.227436865342643,
      "objective/train/tokens_used": 18022400,
      "objective/train/unlikelihood_loss": 0.04339008778333664,
      "theoretical_loss": 6.227436865342643,
      "tokens_seen": 18022400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002738095238095238,
      "loss": 2.6268,
      "theoretical_loss": 6.223557922346955,
      "tokens_seen": 18087936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027579365079365083,
      "loss": 2.5336,
      "theoretical_loss": 6.215853732112821,
      "tokens_seen": 18219008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002777777777777778,
      "loss": 2.3994,
      "theoretical_loss": 6.208220162281178,
      "tokens_seen": 18350080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002797619047619048,
      "loss": 2.3323,
      "theoretical_loss": 6.20065606797053,
      "tokens_seen": 18481152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028174603174603173,
      "loss": 2.5059,
      "theoretical_loss": 6.1931603308120975,
      "tokens_seen": 18612224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028373015873015873,
      "loss": 2.28,
      "theoretical_loss": 6.185731858154261,
      "tokens_seen": 18743296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002857142857142857,
      "loss": 2.4878,
      "theoretical_loss": 6.17836958229627,
      "tokens_seen": 18874368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002876984126984127,
      "loss": 2.4956,
      "theoretical_loss": 6.171072459749913,
      "tokens_seen": 19005440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002896825396825397,
      "loss": 2.2678,
      "theoretical_loss": 6.163839470527964,
      "tokens_seen": 19136512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002916666666666667,
      "loss": 2.1494,
      "theoretical_loss": 6.156669617458243,
      "tokens_seen": 19267584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002936507936507937,
      "loss": 2.3674,
      "theoretical_loss": 6.149561925522211,
      "tokens_seen": 19398656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029563492063492064,
      "loss": 2.3958,
      "theoretical_loss": 6.142515441217064,
      "tokens_seen": 19529728
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.4558895528316498,
      "objective/train/avg_token_score": 0.004920742008835077,
      "objective/train/docs_used": 23612,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.9035089015960693,
      "objective/train/original_loss": 6.259584426879883,
      "objective/train/theoretical_loss": 6.135529231940326,
      "objective/train/tokens_used": 19660800,
      "objective/train/unlikelihood_loss": 0.0633547455072403,
      "theoretical_loss": 6.135529231940326,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029761904761904765,
      "loss": 2.5385,
      "theoretical_loss": 6.135529231940326,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002996031746031746,
      "loss": 2.3807,
      "theoretical_loss": 6.128602385396022,
      "tokens_seen": 19791872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003015873015873016,
      "loss": 2.0421,
      "theoretical_loss": 6.121734009021521,
      "tokens_seen": 19922944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030357142857142855,
      "loss": 2.5316,
      "theoretical_loss": 6.114923229434213,
      "tokens_seen": 20054016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003055555555555556,
      "loss": 2.4607,
      "theoretical_loss": 6.108169191897195,
      "tokens_seen": 20185088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030753968253968255,
      "loss": 2.5377,
      "theoretical_loss": 6.101471059803204,
      "tokens_seen": 20316160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030952380952380956,
      "loss": 2.7593,
      "theoretical_loss": 6.094828014176053,
      "tokens_seen": 20447232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003115079365079365,
      "loss": 2.6748,
      "theoretical_loss": 6.088239253188885,
      "tokens_seen": 20578304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003134920634920635,
      "loss": 2.387,
      "theoretical_loss": 6.0817039916985465,
      "tokens_seen": 20709376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031547619047619046,
      "loss": 2.3745,
      "theoretical_loss": 6.075221460795472,
      "tokens_seen": 20840448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031746031746031746,
      "loss": 2.3548,
      "theoretical_loss": 6.068790907368448,
      "tokens_seen": 20971520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003194444444444444,
      "loss": 2.5693,
      "theoretical_loss": 6.062411593683687,
      "tokens_seen": 21102592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032142857142857147,
      "loss": 2.754,
      "theoretical_loss": 6.056082796977648,
      "tokens_seen": 21233664
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.5116080045700073,
      "objective/train/avg_token_score": 0.0047623817808926105,
      "objective/train/docs_used": 24330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 3.2419803142547607,
      "objective/train/original_loss": 5.9219136238098145,
      "objective/train/theoretical_loss": 6.052937120624258,
      "objective/train/tokens_used": 21299200,
      "objective/train/unlikelihood_loss": 0.10302968323230743,
      "theoretical_loss": 6.052937120624258,
      "tokens_seen": 21299200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003234126984126984,
      "loss": 2.6761,
      "theoretical_loss": 6.049803809063083,
      "tokens_seen": 21364736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003253968253968254,
      "loss": 2.3248,
      "theoretical_loss": 6.04357393594778,
      "tokens_seen": 21495808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032738095238095237,
      "loss": 2.5682,
      "theoretical_loss": 6.037392497465552,
      "tokens_seen": 21626880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032936507936507937,
      "loss": 2.2946,
      "theoretical_loss": 6.031258826918979,
      "tokens_seen": 21757952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003313492063492063,
      "loss": 2.6749,
      "theoretical_loss": 6.025172270733464,
      "tokens_seen": 21889024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003333333333333333,
      "loss": 2.6852,
      "theoretical_loss": 6.0191321881221995,
      "tokens_seen": 22020096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003353174603174603,
      "loss": 2.5813,
      "theoretical_loss": 6.01313795076161,
      "tokens_seen": 22151168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033730158730158733,
      "loss": 2.6132,
      "theoretical_loss": 6.007188942476907,
      "tokens_seen": 22282240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033928571428571433,
      "loss": 2.1199,
      "theoretical_loss": 6.001284558937368,
      "tokens_seen": 22413312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003412698412698413,
      "loss": 2.5638,
      "theoretical_loss": 5.995424207360987,
      "tokens_seen": 22544384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003432539682539683,
      "loss": 2.4583,
      "theoretical_loss": 5.989607306228168,
      "tokens_seen": 22675456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034523809523809523,
      "loss": 2.6098,
      "theoretical_loss": 5.983833285004112,
      "tokens_seen": 22806528
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.374450147151947,
      "objective/train/avg_token_score": 0.010239812545478344,
      "objective/train/docs_used": 24983,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.322624921798706,
      "objective/train/original_loss": 6.397256374359131,
      "objective/train/theoretical_loss": 5.978101583869607,
      "objective/train/tokens_used": 22937600,
      "objective/train/unlikelihood_loss": 0.04547956585884094,
      "theoretical_loss": 5.978101583869607,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034722222222222224,
      "loss": 2.6236,
      "theoretical_loss": 5.978101583869607,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003492063492063492,
      "loss": 2.2842,
      "theoretical_loss": 5.972411653459913,
      "tokens_seen": 23068672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003511904761904762,
      "loss": 2.4053,
      "theoretical_loss": 5.966762954611432,
      "tokens_seen": 23199744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003531746031746032,
      "loss": 2.4468,
      "theoretical_loss": 5.961154958115937,
      "tokens_seen": 23330816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003551587301587302,
      "loss": 2.7427,
      "theoretical_loss": 5.955587144482044,
      "tokens_seen": 23461888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035714285714285714,
      "loss": 2.6201,
      "theoretical_loss": 5.950059003703704,
      "tokens_seen": 23592960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035912698412698415,
      "loss": 2.6046,
      "theoretical_loss": 5.944570035035458,
      "tokens_seen": 23724032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003611111111111111,
      "loss": 2.4813,
      "theoretical_loss": 5.939119746774228,
      "tokens_seen": 23855104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003630952380952381,
      "loss": 2.464,
      "theoretical_loss": 5.933707656047414,
      "tokens_seen": 23986176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036507936507936505,
      "loss": 2.5273,
      "theoretical_loss": 5.928333288607086,
      "tokens_seen": 24117248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036706349206349205,
      "loss": 2.6318,
      "theoretical_loss": 5.92299617863006,
      "tokens_seen": 24248320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036904761904761905,
      "loss": 2.5566,
      "theoretical_loss": 5.9176958685236585,
      "tokens_seen": 24379392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037103174603174606,
      "loss": 2.3378,
      "theoretical_loss": 5.912431908736972,
      "tokens_seen": 24510464
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.3531891405582428,
      "objective/train/avg_token_score": 0.034930985420942307,
      "objective/train/docs_used": 26439,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0727763175964355,
      "objective/train/original_loss": 6.117815971374512,
      "objective/train/theoretical_loss": 5.909813421882534,
      "objective/train/tokens_used": 24576000,
      "objective/train/unlikelihood_loss": 0.0909704715013504,
      "theoretical_loss": 5.909813421882534,
      "tokens_seen": 24576000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000373015873015873,
      "loss": 2.2126,
      "theoretical_loss": 5.907203857577422,
      "tokens_seen": 24641536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000375,
      "loss": 2.7691,
      "theoretical_loss": 5.902011281032472,
      "tokens_seen": 24772608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000376984126984127,
      "loss": 2.3763,
      "theoretical_loss": 5.896853752596286,
      "tokens_seen": 24903680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037896825396825396,
      "loss": 2.1913,
      "theoretical_loss": 5.891730853101199,
      "tokens_seen": 25034752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038095238095238096,
      "loss": 2.6541,
      "theoretical_loss": 5.8866421705538325,
      "tokens_seen": 25165824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038293650793650797,
      "loss": 2.3709,
      "theoretical_loss": 5.881587299975694,
      "tokens_seen": 25296896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038492063492063497,
      "loss": 2.0019,
      "theoretical_loss": 5.876565843248124,
      "tokens_seen": 25427968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003869047619047619,
      "loss": 2.3081,
      "theoretical_loss": 5.871577408961457,
      "tokens_seen": 25559040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003888888888888889,
      "loss": 2.1361,
      "theoretical_loss": 5.866621612268246,
      "tokens_seen": 25690112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039087301587301587,
      "loss": 2.4905,
      "theoretical_loss": 5.8616980747404295,
      "tokens_seen": 25821184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003928571428571429,
      "loss": 2.2697,
      "theoretical_loss": 5.856806424230314,
      "tokens_seen": 25952256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003948412698412698,
      "loss": 2.3797,
      "theoretical_loss": 5.851946294735258,
      "tokens_seen": 26083328
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.48717010021209717,
      "objective/train/avg_token_score": 0.020001767203211784,
      "objective/train/docs_used": 27060,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.8721370697021484,
      "objective/train/original_loss": 5.758991718292236,
      "objective/train/theoretical_loss": 5.8471173262659235,
      "objective/train/tokens_used": 26214400,
      "objective/train/unlikelihood_loss": 0.08089856803417206,
      "theoretical_loss": 5.8471173262659235,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003968253968253968,
      "loss": 2.4683,
      "theoretical_loss": 5.8471173262659235,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039880952380952383,
      "loss": 2.3909,
      "theoretical_loss": 5.842319164718004,
      "tokens_seen": 26345472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040079365079365083,
      "loss": 2.3413,
      "theoretical_loss": 5.83755146174731,
      "tokens_seen": 26476544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004027777777777778,
      "loss": 2.6428,
      "theoretical_loss": 5.832813874648102,
      "tokens_seen": 26607616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004047619047619048,
      "loss": 2.5088,
      "theoretical_loss": 5.828106066234588,
      "tokens_seen": 26738688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040674603174603173,
      "loss": 2.5131,
      "theoretical_loss": 5.823427704725473,
      "tokens_seen": 26869760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040873015873015874,
      "loss": 2.8614,
      "theoretical_loss": 5.818778463631473,
      "tokens_seen": 27000832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004107142857142857,
      "loss": 2.3157,
      "theoretical_loss": 5.8141580216457065,
      "tokens_seen": 27131904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004126984126984127,
      "loss": 2.4281,
      "theoretical_loss": 5.809566062536868,
      "tokens_seen": 27262976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004146825396825397,
      "loss": 2.6006,
      "theoretical_loss": 5.805002275045111,
      "tokens_seen": 27394048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004166666666666667,
      "loss": 2.2208,
      "theoretical_loss": 5.800466352780546,
      "tokens_seen": 27525120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004186507936507937,
      "loss": 2.4363,
      "theoretical_loss": 5.795957994124291,
      "tokens_seen": 27656192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042063492063492065,
      "loss": 2.5718,
      "theoretical_loss": 5.791476902131985,
      "tokens_seen": 27787264
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.39797165989875793,
      "objective/train/avg_token_score": 0.010648269206285477,
      "objective/train/docs_used": 28453,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.4116973876953125,
      "objective/train/original_loss": 6.296491622924805,
      "objective/train/theoretical_loss": 5.789246489594688,
      "objective/train/tokens_used": 27852800,
      "objective/train/unlikelihood_loss": 0.0855846107006073,
      "theoretical_loss": 5.789246489594688,
      "tokens_seen": 27852800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042261904761904765,
      "loss": 2.3402,
      "theoretical_loss": 5.787022784439701,
      "tokens_seen": 27918336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004246031746031746,
      "loss": 2.5929,
      "theoretical_loss": 5.782595353172176,
      "tokens_seen": 28049408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004265873015873016,
      "loss": 2.671,
      "theoretical_loss": 5.778194324853311,
      "tokens_seen": 28180480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042857142857142855,
      "loss": 2.5018,
      "theoretical_loss": 5.773819420318858,
      "tokens_seen": 28311552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004305555555555556,
      "loss": 2.2753,
      "theoretical_loss": 5.769470364631225,
      "tokens_seen": 28442624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043253968253968256,
      "loss": 2.0779,
      "theoretical_loss": 5.765146886996363,
      "tokens_seen": 28573696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043452380952380956,
      "loss": 2.1053,
      "theoretical_loss": 5.760848720682651,
      "tokens_seen": 28704768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004365079365079365,
      "loss": 2.3278,
      "theoretical_loss": 5.756575602941732,
      "tokens_seen": 28835840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004384920634920635,
      "loss": 2.4976,
      "theoretical_loss": 5.752327274931249,
      "tokens_seen": 28966912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044047619047619046,
      "loss": 2.2979,
      "theoretical_loss": 5.7481034816394105,
      "tokens_seen": 29097984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044246031746031746,
      "loss": 2.4162,
      "theoretical_loss": 5.74390397181136,
      "tokens_seen": 29229056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004444444444444444,
      "loss": 2.4311,
      "theoretical_loss": 5.739728497877267,
      "tokens_seen": 29360128
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.28812316060066223,
      "objective/train/avg_token_score": 0.02659354917705059,
      "objective/train/docs_used": 29115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.703152060508728,
      "objective/train/original_loss": 6.603550434112549,
      "objective/train/theoretical_loss": 5.7355768158821245,
      "objective/train/tokens_used": 29491200,
      "objective/train/unlikelihood_loss": 0.04069405049085617,
      "theoretical_loss": 5.7355768158821245,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044642857142857147,
      "loss": 2.4414,
      "theoretical_loss": 5.7355768158821245,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004484126984126984,
      "loss": 2.4985,
      "theoretical_loss": 5.731448685417178,
      "tokens_seen": 29622272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004503968253968254,
      "loss": 2.2532,
      "theoretical_loss": 5.7273438695529535,
      "tokens_seen": 29753344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045238095238095237,
      "loss": 2.4096,
      "theoretical_loss": 5.7232621347738455,
      "tokens_seen": 29884416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045436507936507937,
      "loss": 2.1835,
      "theoretical_loss": 5.719203250914208,
      "tokens_seen": 30015488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004563492063492063,
      "loss": 2.4457,
      "theoretical_loss": 5.715166991095922,
      "tokens_seen": 30146560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004583333333333333,
      "loss": 2.598,
      "theoretical_loss": 5.71115313166738,
      "tokens_seen": 30277632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046031746031746033,
      "loss": 2.5442,
      "theoretical_loss": 5.707161452143879,
      "tokens_seen": 30408704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046230158730158733,
      "loss": 2.2213,
      "theoretical_loss": 5.7031917351493515,
      "tokens_seen": 30539776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046428571428571433,
      "loss": 2.178,
      "theoretical_loss": 5.699243766359421,
      "tokens_seen": 30670848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004662698412698413,
      "loss": 2.4453,
      "theoretical_loss": 5.695317334445736,
      "tokens_seen": 30801920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004682539682539683,
      "loss": 2.4695,
      "theoretical_loss": 5.691412231021549,
      "tokens_seen": 30932992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047023809523809523,
      "loss": 2.4932,
      "theoretical_loss": 5.687528250588518,
      "tokens_seen": 31064064
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.21346530318260193,
      "objective/train/avg_token_score": 0.01769358292222023,
      "objective/train/docs_used": 30152,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2651164531707764,
      "objective/train/original_loss": 6.490805149078369,
      "objective/train/theoretical_loss": 5.6855941180524265,
      "objective/train/tokens_used": 31129600,
      "objective/train/unlikelihood_loss": 0.0511309988796711,
      "theoretical_loss": 5.6855941180524265,
      "tokens_seen": 31129600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047222222222222224,
      "loss": 2.1794,
      "theoretical_loss": 5.683665190484683,
      "tokens_seen": 31195136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004742063492063492,
      "loss": 2.251,
      "theoretical_loss": 5.679822850833591,
      "tokens_seen": 31326208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004761904761904762,
      "loss": 2.3855,
      "theoretical_loss": 5.676001034494554,
      "tokens_seen": 31457280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004781746031746032,
      "loss": 2.352,
      "theoretical_loss": 5.672199547013983,
      "tokens_seen": 31588352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004801587301587302,
      "loss": 2.5004,
      "theoretical_loss": 5.6684181965778,
      "tokens_seen": 31719424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048214285714285715,
      "loss": 2.3015,
      "theoretical_loss": 5.6646567939648715,
      "tokens_seen": 31850496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048412698412698415,
      "loss": 2.358,
      "theoretical_loss": 5.660915152501465,
      "tokens_seen": 31981568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004861111111111111,
      "loss": 2.496,
      "theoretical_loss": 5.657193088016677,
      "tokens_seen": 32112640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004880952380952381,
      "loss": 2.3284,
      "theoretical_loss": 5.653490418798825,
      "tokens_seen": 32243712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004900793650793651,
      "loss": 2.2022,
      "theoretical_loss": 5.649806965552774,
      "tokens_seen": 32374784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000492063492063492,
      "loss": 2.1381,
      "theoretical_loss": 5.6461425513581665,
      "tokens_seen": 32505856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004940476190476191,
      "loss": 2.6351,
      "theoretical_loss": 5.6424970016285485,
      "tokens_seen": 32636928
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.4376832842826843,
      "objective/train/avg_token_score": 0.009709020145237446,
      "objective/train/docs_used": 30758,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.5283288955688477,
      "objective/train/original_loss": 5.68566370010376,
      "objective/train/theoretical_loss": 5.638870144071353,
      "objective/train/tokens_used": 32768000,
      "objective/train/unlikelihood_loss": 0.07930290699005127,
      "theoretical_loss": 5.638870144071353,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000496031746031746,
      "loss": 2.3134,
      "theoretical_loss": 5.638870144071353,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498015873015873,
      "loss": 2.3275,
      "theoretical_loss": 5.635261808648728,
      "tokens_seen": 32899072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005,
      "loss": 2.2282,
      "theoretical_loss": 5.631671827539186,
      "tokens_seen": 33030144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999799398194584,
      "loss": 2.256,
      "theoretical_loss": 5.628100035100061,
      "tokens_seen": 33161216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999598796389167,
      "loss": 2.2104,
      "theoretical_loss": 5.6245462678307385,
      "tokens_seen": 33292288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999398194583751,
      "loss": 2.2591,
      "theoretical_loss": 5.621010364336651,
      "tokens_seen": 33423360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999197592778335,
      "loss": 2.4118,
      "theoretical_loss": 5.61749216529402,
      "tokens_seen": 33554432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998996990972919,
      "loss": 2.2046,
      "theoretical_loss": 5.61399151341532,
      "tokens_seen": 33685504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998796389167503,
      "loss": 2.3628,
      "theoretical_loss": 5.610508253415453,
      "tokens_seen": 33816576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998595787362087,
      "loss": 2.3158,
      "theoretical_loss": 5.6070422319786095,
      "tokens_seen": 33947648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499839518555667,
      "loss": 2.2646,
      "theoretical_loss": 5.603593297725807,
      "tokens_seen": 34078720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998194583751254,
      "loss": 2.4531,
      "theoretical_loss": 5.600161301183084,
      "tokens_seen": 34209792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997993981945837,
      "loss": 2.3033,
      "theoretical_loss": 5.596746094750342,
      "tokens_seen": 34340864
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.31781524419784546,
      "objective/train/avg_token_score": 0.06681742519140244,
      "objective/train/docs_used": 31355,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7797188758850098,
      "objective/train/original_loss": 5.599770545959473,
      "objective/train/theoretical_loss": 5.595044742205037,
      "objective/train/tokens_used": 34406400,
      "objective/train/unlikelihood_loss": 0.1539888083934784,
      "theoretical_loss": 5.595044742205037,
      "tokens_seen": 34406400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997793380140421,
      "loss": 2.0253,
      "theoretical_loss": 5.5933475326707995,
      "tokens_seen": 34471936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997592778335005,
      "loss": 2.3513,
      "theoretical_loss": 5.589965471001077,
      "tokens_seen": 34603008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997392176529588,
      "loss": 2.385,
      "theoretical_loss": 5.586599767581859,
      "tokens_seen": 34734080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997191574724173,
      "loss": 2.4408,
      "theoretical_loss": 5.583250282009159,
      "tokens_seen": 34865152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996990972918757,
      "loss": 2.2335,
      "theoretical_loss": 5.579916875606134,
      "tokens_seen": 34996224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499679037111334,
      "loss": 2.2096,
      "theoretical_loss": 5.576599411395472,
      "tokens_seen": 35127296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996589769307924,
      "loss": 2.5149,
      "theoretical_loss": 5.5732977540723105,
      "tokens_seen": 35258368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996389167502507,
      "loss": 2.1646,
      "theoretical_loss": 5.570011769977693,
      "tokens_seen": 35389440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996188565697091,
      "loss": 2.2061,
      "theoretical_loss": 5.566741327072535,
      "tokens_seen": 35520512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995987963891675,
      "loss": 2.462,
      "theoretical_loss": 5.563486294912105,
      "tokens_seen": 35651584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995787362086258,
      "loss": 2.1712,
      "theoretical_loss": 5.560246544620993,
      "tokens_seen": 35782656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995586760280842,
      "loss": 2.4396,
      "theoretical_loss": 5.557021948868571,
      "tokens_seen": 35913728
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.37939882278442383,
      "objective/train/avg_token_score": 0.02247542515397072,
      "objective/train/docs_used": 32385,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.2949864864349365,
      "objective/train/original_loss": 5.736522674560547,
      "objective/train/theoretical_loss": 5.553812381844907,
      "objective/train/tokens_used": 36044800,
      "objective/train/unlikelihood_loss": 0.08487144112586975,
      "theoretical_loss": 5.553812381844907,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995386158475427,
      "loss": 2.2289,
      "theoretical_loss": 5.553812381844907,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499518555667001,
      "loss": 2.3559,
      "theoretical_loss": 5.550617719237167,
      "tokens_seen": 36175872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994984954864594,
      "loss": 2.1262,
      "theoretical_loss": 5.547437838206435,
      "tokens_seen": 36306944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994784353059178,
      "loss": 2.2879,
      "theoretical_loss": 5.544272617365014,
      "tokens_seen": 36438016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994583751253761,
      "loss": 2.2899,
      "theoretical_loss": 5.541121936754111,
      "tokens_seen": 36569088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994383149448345,
      "loss": 2.1249,
      "theoretical_loss": 5.537985677821986,
      "tokens_seen": 36700160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994182547642928,
      "loss": 2.5925,
      "theoretical_loss": 5.5348637234024824,
      "tokens_seen": 36831232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993981945837512,
      "loss": 2.3988,
      "theoretical_loss": 5.5317559576939725,
      "tokens_seen": 36962304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993781344032096,
      "loss": 2.1127,
      "theoretical_loss": 5.5286622662386975,
      "tokens_seen": 37093376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499358074222668,
      "loss": 2.2598,
      "theoretical_loss": 5.525582535902489,
      "tokens_seen": 37224448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993380140421264,
      "loss": 2.2746,
      "theoretical_loss": 5.52251665485486,
      "tokens_seen": 37355520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993179538615848,
      "loss": 2.0758,
      "theoretical_loss": 5.519464512549478,
      "tokens_seen": 37486592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992978936810431,
      "loss": 2.0689,
      "theoretical_loss": 5.516425999704987,
      "tokens_seen": 37617664
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.4496578574180603,
      "objective/train/avg_token_score": 0.006494433153420687,
      "objective/train/docs_used": 32883,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.570333480834961,
      "objective/train/original_loss": 5.788158416748047,
      "objective/train/theoretical_loss": 5.514911820521407,
      "objective/train/tokens_used": 37683200,
      "objective/train/unlikelihood_loss": 0.06332194805145264,
      "theoretical_loss": 5.514911820521407,
      "tokens_seen": 37683200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992778335005015,
      "loss": 2.2661,
      "theoretical_loss": 5.5134010082861895,
      "tokens_seen": 37748736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992577733199598,
      "loss": 2.4036,
      "theoretical_loss": 5.51038943148556,
      "tokens_seen": 37879808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992377131394183,
      "loss": 2.6818,
      "theoretical_loss": 5.50739116370511,
      "tokens_seen": 38010880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992176529588767,
      "loss": 2.1309,
      "theoretical_loss": 5.5044061005385725,
      "tokens_seen": 38141952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499197592778335,
      "loss": 2.3905,
      "theoretical_loss": 5.501434138753918,
      "tokens_seen": 38273024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991775325977934,
      "loss": 2.5159,
      "theoretical_loss": 5.498475176276176,
      "tokens_seen": 38404096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991574724172518,
      "loss": 2.6962,
      "theoretical_loss": 5.495529112170568,
      "tokens_seen": 38535168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991374122367101,
      "loss": 2.1093,
      "theoretical_loss": 5.492595846625951,
      "tokens_seen": 38666240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991173520561685,
      "loss": 2.4076,
      "theoretical_loss": 5.489675280938547,
      "tokens_seen": 38797312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499097291875627,
      "loss": 2.3561,
      "theoretical_loss": 5.486767317495966,
      "tokens_seen": 38928384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990772316950853,
      "loss": 2.4408,
      "theoretical_loss": 5.483871859761511,
      "tokens_seen": 39059456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990571715145437,
      "loss": 2.1575,
      "theoretical_loss": 5.480988812258763,
      "tokens_seen": 39190528
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.41135141253471375,
      "objective/train/avg_token_score": 0.026532353833317757,
      "objective/train/docs_used": 34166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.235038995742798,
      "objective/train/original_loss": 5.851961612701416,
      "objective/train/theoretical_loss": 5.478118080556438,
      "objective/train/tokens_used": 39321600,
      "objective/train/unlikelihood_loss": 0.04781878739595413,
      "theoretical_loss": 5.478118080556438,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499037111334002,
      "loss": 2.1814,
      "theoretical_loss": 5.478118080556438,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990170511534604,
      "loss": 2.2678,
      "theoretical_loss": 5.475259571253502,
      "tokens_seen": 39452672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989969909729188,
      "loss": 2.4555,
      "theoretical_loss": 5.4724131919645576,
      "tokens_seen": 39583744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989769307923771,
      "loss": 2.2421,
      "theoretical_loss": 5.4695788513054815,
      "tokens_seen": 39714816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989568706118355,
      "loss": 2.2966,
      "theoretical_loss": 5.466756458879306,
      "tokens_seen": 39845888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989368104312939,
      "loss": 2.578,
      "theoretical_loss": 5.463945925262355,
      "tokens_seen": 39976960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989167502507523,
      "loss": 1.9973,
      "theoretical_loss": 5.461147161990611,
      "tokens_seen": 40108032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988966900702107,
      "loss": 2.4388,
      "theoretical_loss": 5.458360081546321,
      "tokens_seen": 40239104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988766298896691,
      "loss": 2.4184,
      "theoretical_loss": 5.455584597344835,
      "tokens_seen": 40370176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988565697091274,
      "loss": 2.1967,
      "theoretical_loss": 5.452820623721662,
      "tokens_seen": 40501248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988365095285858,
      "loss": 2.3451,
      "theoretical_loss": 5.450068075919752,
      "tokens_seen": 40632320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988164493480441,
      "loss": 2.0896,
      "theoretical_loss": 5.447326870076996,
      "tokens_seen": 40763392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987963891675025,
      "loss": 2.1861,
      "theoretical_loss": 5.444596923213931,
      "tokens_seen": 40894464
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.33492180705070496,
      "objective/train/avg_token_score": 0.03136708214879036,
      "objective/train/docs_used": 34866,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9246761798858643,
      "objective/train/original_loss": 5.993686199188232,
      "objective/train/theoretical_loss": 5.443236146207074,
      "objective/train/tokens_used": 40960000,
      "objective/train/unlikelihood_loss": 0.09010782837867737,
      "theoretical_loss": 5.443236146207074,
      "tokens_seen": 40960000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987763289869609,
      "loss": 2.2846,
      "theoretical_loss": 5.441878153221662,
      "tokens_seen": 41025536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987562688064192,
      "loss": 2.4934,
      "theoretical_loss": 5.439170478849976,
      "tokens_seen": 41156608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987362086258777,
      "loss": 1.8671,
      "theoretical_loss": 5.4364738196956655,
      "tokens_seen": 41287680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987161484453361,
      "loss": 2.2663,
      "theoretical_loss": 5.433788096191039,
      "tokens_seen": 41418752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986960882647944,
      "loss": 2.3139,
      "theoretical_loss": 5.4311132295926345,
      "tokens_seen": 41549824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986760280842528,
      "loss": 2.5888,
      "theoretical_loss": 5.428449141970107,
      "tokens_seen": 41680896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986559679037111,
      "loss": 2.1469,
      "theoretical_loss": 5.42579575619531,
      "tokens_seen": 41811968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986359077231695,
      "loss": 2.3618,
      "theoretical_loss": 5.423152995931552,
      "tokens_seen": 41943040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986158475426279,
      "loss": 2.3092,
      "theoretical_loss": 5.420520785623031,
      "tokens_seen": 42074112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985957873620862,
      "loss": 2.3179,
      "theoretical_loss": 5.417899050484451,
      "tokens_seen": 42205184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985757271815446,
      "loss": 2.4495,
      "theoretical_loss": 5.415287716490787,
      "tokens_seen": 42336256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985556670010031,
      "loss": 2.2482,
      "theoretical_loss": 5.412686710367245,
      "tokens_seen": 42467328
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.3299120366573334,
      "objective/train/avg_token_score": 0.04832848906517029,
      "objective/train/docs_used": 35868,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8417558670043945,
      "objective/train/original_loss": 5.99447774887085,
      "objective/train/theoretical_loss": 5.410095959579362,
      "objective/train/tokens_used": 42598400,
      "objective/train/unlikelihood_loss": 0.052462801337242126,
      "theoretical_loss": 5.410095959579362,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985356068204614,
      "loss": 1.9885,
      "theoretical_loss": 5.410095959579362,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985155466399198,
      "loss": 2.4452,
      "theoretical_loss": 5.407515392323276,
      "tokens_seen": 42729472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984954864593782,
      "loss": 2.3031,
      "theoretical_loss": 5.404944937516161,
      "tokens_seen": 42860544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984754262788365,
      "loss": 2.2327,
      "theoretical_loss": 5.402384524786797,
      "tokens_seen": 42991616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984553660982949,
      "loss": 1.8473,
      "theoretical_loss": 5.399834084466306,
      "tokens_seen": 43122688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984353059177532,
      "loss": 2.2829,
      "theoretical_loss": 5.397293547579041,
      "tokens_seen": 43253760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984152457372116,
      "loss": 2.2311,
      "theoretical_loss": 5.394762845833601,
      "tokens_seen": 43384832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049839518555667,
      "loss": 2.3256,
      "theoretical_loss": 5.392241911614005,
      "tokens_seen": 43515904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983751253761284,
      "loss": 2.2566,
      "theoretical_loss": 5.389730677971002,
      "tokens_seen": 43646976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983550651955868,
      "loss": 2.5088,
      "theoretical_loss": 5.387229078613521,
      "tokens_seen": 43778048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983350050150452,
      "loss": 2.1468,
      "theoretical_loss": 5.384737047900243,
      "tokens_seen": 43909120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983149448345035,
      "loss": 2.3067,
      "theoretical_loss": 5.382254520831328,
      "tokens_seen": 44040192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982948846539619,
      "loss": 1.9734,
      "theoretical_loss": 5.379781433040252,
      "tokens_seen": 44171264
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.345246821641922,
      "objective/train/avg_token_score": 0.005966670345515013,
      "objective/train/docs_used": 36544,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9924170970916748,
      "objective/train/original_loss": 5.951955795288086,
      "objective/train/theoretical_loss": 5.378548408930558,
      "objective/train/tokens_used": 44236800,
      "objective/train/unlikelihood_loss": 0.06286505609750748,
      "theoretical_loss": 5.378548408930558,
      "tokens_seen": 44236800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982748244734202,
      "loss": 2.169,
      "theoretical_loss": 5.377317720785777,
      "tokens_seen": 44302336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982547642928786,
      "loss": 2.5078,
      "theoretical_loss": 5.374863320944057,
      "tokens_seen": 44433408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498234704112337,
      "loss": 2.2855,
      "theoretical_loss": 5.372418171000847,
      "tokens_seen": 44564480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982146439317953,
      "loss": 2.2421,
      "theoretical_loss": 5.369982209043851,
      "tokens_seen": 44695552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981945837512538,
      "loss": 2.4581,
      "theoretical_loss": 5.367555373755179,
      "tokens_seen": 44826624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981745235707122,
      "loss": 2.3194,
      "theoretical_loss": 5.365137604403923,
      "tokens_seen": 44957696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981544633901705,
      "loss": 2.1544,
      "theoretical_loss": 5.362728840838843,
      "tokens_seen": 45088768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981344032096289,
      "loss": 2.0627,
      "theoretical_loss": 5.360329023481169,
      "tokens_seen": 45219840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981143430290873,
      "loss": 2.2282,
      "theoretical_loss": 5.357938093317518,
      "tokens_seen": 45350912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980942828485456,
      "loss": 2.1852,
      "theoretical_loss": 5.355555991892905,
      "tokens_seen": 45481984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498074222668004,
      "loss": 2.147,
      "theoretical_loss": 5.353182661303873,
      "tokens_seen": 45613056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980541624874623,
      "loss": 2.3338,
      "theoretical_loss": 5.350818044191721,
      "tokens_seen": 45744128
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.4638318717479706,
      "objective/train/avg_token_score": 0.017905525863170624,
      "objective/train/docs_used": 37915,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.485879421234131,
      "objective/train/original_loss": 5.818847179412842,
      "objective/train/theoretical_loss": 5.348462083735834,
      "objective/train/tokens_used": 45875200,
      "objective/train/unlikelihood_loss": 0.05798850208520889,
      "theoretical_loss": 5.348462083735834,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980341023069207,
      "loss": 2.2988,
      "theoretical_loss": 5.348462083735834,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980140421263792,
      "loss": 2.3518,
      "theoretical_loss": 5.346114723647119,
      "tokens_seen": 46006272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979939819458375,
      "loss": 2.4436,
      "theoretical_loss": 5.343775908161532,
      "tokens_seen": 46137344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979739217652959,
      "loss": 2.5148,
      "theoretical_loss": 5.341445582033705,
      "tokens_seen": 46268416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979538615847543,
      "loss": 2.2889,
      "theoretical_loss": 5.339123690530673,
      "tokens_seen": 46399488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979338014042126,
      "loss": 2.4053,
      "theoretical_loss": 5.336810179425685,
      "tokens_seen": 46530560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497913741223671,
      "loss": 2.5862,
      "theoretical_loss": 5.334504994992115,
      "tokens_seen": 46661632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978936810431293,
      "loss": 2.2837,
      "theoretical_loss": 5.332208083997459,
      "tokens_seen": 46792704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978736208625877,
      "loss": 2.2526,
      "theoretical_loss": 5.329919393697422,
      "tokens_seen": 46923776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978535606820461,
      "loss": 2.2839,
      "theoretical_loss": 5.327638871830089,
      "tokens_seen": 47054848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978335005015045,
      "loss": 2.1244,
      "theoretical_loss": 5.32536646661019,
      "tokens_seen": 47185920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978134403209629,
      "loss": 2.3265,
      "theoretical_loss": 5.323102126723439,
      "tokens_seen": 47316992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977933801404213,
      "loss": 2.1693,
      "theoretical_loss": 5.320845801320959,
      "tokens_seen": 47448064
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.5400782227516174,
      "objective/train/avg_token_score": 0.00798075646162033,
      "objective/train/docs_used": 38651,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.944997549057007,
      "objective/train/original_loss": 5.1642985343933105,
      "objective/train/theoretical_loss": 5.319720628286955,
      "objective/train/tokens_used": 47513600,
      "objective/train/unlikelihood_loss": 0.09301882237195969,
      "theoretical_loss": 5.319720628286955,
      "tokens_seen": 47513600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977733199598796,
      "loss": 2.4382,
      "theoretical_loss": 5.318597440013795,
      "tokens_seen": 47579136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497753259779338,
      "loss": 2.4145,
      "theoretical_loss": 5.316356992867491,
      "tokens_seen": 47710208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977331995987965,
      "loss": 2.3468,
      "theoretical_loss": 5.314124410396767,
      "tokens_seen": 47841280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977131394182548,
      "loss": 1.8894,
      "theoretical_loss": 5.311899643560251,
      "tokens_seen": 47972352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976930792377132,
      "loss": 2.3309,
      "theoretical_loss": 5.3096826437553,
      "tokens_seen": 48103424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976730190571715,
      "loss": 2.1528,
      "theoretical_loss": 5.3074733628129005,
      "tokens_seen": 48234496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976529588766299,
      "loss": 2.1967,
      "theoretical_loss": 5.305271752992619,
      "tokens_seen": 48365568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976328986960883,
      "loss": 2.3363,
      "theoretical_loss": 5.303077766977653,
      "tokens_seen": 48496640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976128385155466,
      "loss": 2.1614,
      "theoretical_loss": 5.300891357869929,
      "tokens_seen": 48627712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497592778335005,
      "loss": 2.4155,
      "theoretical_loss": 5.298712479185288,
      "tokens_seen": 48758784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975727181544635,
      "loss": 2.1652,
      "theoretical_loss": 5.296541084848727,
      "tokens_seen": 48889856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975526579739218,
      "loss": 2.2516,
      "theoretical_loss": 5.294377129189715,
      "tokens_seen": 49020928
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.4233260154724121,
      "objective/train/avg_token_score": 0.01209855917841196,
      "objective/train/docs_used": 39958,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.3168957233428955,
      "objective/train/original_loss": 5.3540496826171875,
      "objective/train/theoretical_loss": 5.292220566937567,
      "objective/train/tokens_used": 49152000,
      "objective/train/unlikelihood_loss": 0.07595264166593552,
      "theoretical_loss": 5.292220566937567,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975325977933802,
      "loss": 2.3305,
      "theoretical_loss": 5.292220566937567,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975125376128386,
      "loss": 2.3771,
      "theoretical_loss": 5.290071353216895,
      "tokens_seen": 49283072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974924774322969,
      "loss": 2.1448,
      "theoretical_loss": 5.28792944354311,
      "tokens_seen": 49414144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974724172517553,
      "loss": 2.3501,
      "theoretical_loss": 5.285794793817999,
      "tokens_seen": 49545216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974523570712136,
      "loss": 2.0919,
      "theoretical_loss": 5.283667360325351,
      "tokens_seen": 49676288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497432296890672,
      "loss": 2.1672,
      "theoretical_loss": 5.281547099726654,
      "tokens_seen": 49807360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974122367101304,
      "loss": 2.2725,
      "theoretical_loss": 5.279433969056848,
      "tokens_seen": 49938432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973921765295888,
      "loss": 2.3125,
      "theoretical_loss": 5.277327925720137,
      "tokens_seen": 50069504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973721163490472,
      "loss": 2.1681,
      "theoretical_loss": 5.275228927485855,
      "tokens_seen": 50200576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973520561685056,
      "loss": 2.1653,
      "theoretical_loss": 5.273136932484399,
      "tokens_seen": 50331648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973319959879639,
      "loss": 2.4031,
      "theoretical_loss": 5.271051899203207,
      "tokens_seen": 50462720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973119358074223,
      "loss": 2.0327,
      "theoretical_loss": 5.268973786482794,
      "tokens_seen": 50593792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972918756268806,
      "loss": 2.1619,
      "theoretical_loss": 5.266902553512847,
      "tokens_seen": 50724864
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.25061094760894775,
      "objective/train/avg_token_score": 0.02978723682463169,
      "objective/train/docs_used": 40584,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3529692888259888,
      "objective/train/original_loss": 4.955077648162842,
      "objective/train/theoretical_loss": 5.265869504275602,
      "objective/train/tokens_used": 50790400,
      "objective/train/unlikelihood_loss": 0.16319940984249115,
      "theoretical_loss": 5.265869504275602,
      "tokens_seen": 50790400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497271815446339,
      "loss": 2.1136,
      "theoretical_loss": 5.264838159828369,
      "tokens_seen": 50855936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972517552657974,
      "loss": 2.3096,
      "theoretical_loss": 5.262780565305875,
      "tokens_seen": 50987008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972316950852557,
      "loss": 2.3798,
      "theoretical_loss": 5.260729730159641,
      "tokens_seen": 51118080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972116349047142,
      "loss": 2.3046,
      "theoretical_loss": 5.2586856149380035,
      "tokens_seen": 51249152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971915747241726,
      "loss": 2.3706,
      "theoretical_loss": 5.256648180519708,
      "tokens_seen": 51380224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971715145436309,
      "loss": 2.0973,
      "theoretical_loss": 5.25461738811031,
      "tokens_seen": 51511296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971514543630893,
      "loss": 2.6331,
      "theoretical_loss": 5.252593199238619,
      "tokens_seen": 51642368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971313941825477,
      "loss": 2.2023,
      "theoretical_loss": 5.2505755757531904,
      "tokens_seen": 51773440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497111334002006,
      "loss": 2.1155,
      "theoretical_loss": 5.248564479818876,
      "tokens_seen": 51904512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970912738214644,
      "loss": 2.3576,
      "theoretical_loss": 5.246559873913396,
      "tokens_seen": 52035584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970712136409227,
      "loss": 2.0642,
      "theoretical_loss": 5.24456172082399,
      "tokens_seen": 52166656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970511534603811,
      "loss": 2.393,
      "theoretical_loss": 5.242569983644074,
      "tokens_seen": 52297728
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.42399805784225464,
      "objective/train/avg_token_score": 0.009449027478694916,
      "objective/train/docs_used": 41941,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1671478748321533,
      "objective/train/original_loss": 5.030263423919678,
      "objective/train/theoretical_loss": 5.240584625769978,
      "objective/train/tokens_used": 52428800,
      "objective/train/unlikelihood_loss": 0.07760672271251678,
      "theoretical_loss": 5.240584625769978,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970310932798396,
      "loss": 2.2038,
      "theoretical_loss": 5.240584625769978,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970110330992979,
      "loss": 2.1783,
      "theoretical_loss": 5.238605610897698,
      "tokens_seen": 52559872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969909729187563,
      "loss": 2.1103,
      "theoretical_loss": 5.2366329030197125,
      "tokens_seen": 52690944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969709127382147,
      "loss": 2.1368,
      "theoretical_loss": 5.2346664664218245,
      "tokens_seen": 52822016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496950852557673,
      "loss": 2.2045,
      "theoretical_loss": 5.232706265680049,
      "tokens_seen": 52953088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969307923771314,
      "loss": 1.911,
      "theoretical_loss": 5.230752265657554,
      "tokens_seen": 53084160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969107321965897,
      "loss": 2.103,
      "theoretical_loss": 5.228804431501619,
      "tokens_seen": 53215232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968906720160481,
      "loss": 2.4849,
      "theoretical_loss": 5.226862728640651,
      "tokens_seen": 53346304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968706118355065,
      "loss": 2.005,
      "theoretical_loss": 5.2249271227812315,
      "tokens_seen": 53477376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968505516549649,
      "loss": 2.2629,
      "theoretical_loss": 5.222997579905204,
      "tokens_seen": 53608448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968304914744233,
      "loss": 2.2283,
      "theoretical_loss": 5.2210740662667945,
      "tokens_seen": 53739520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968104312938817,
      "loss": 2.399,
      "theoretical_loss": 5.219156548389775,
      "tokens_seen": 53870592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049679037111334,
      "loss": 2.3043,
      "theoretical_loss": 5.217244993064664,
      "tokens_seen": 54001664
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.42821359634399414,
      "objective/train/avg_token_score": 0.012333965860307217,
      "objective/train/docs_used": 42708,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.223114490509033,
      "objective/train/original_loss": 5.237239837646484,
      "objective/train/theoretical_loss": 5.216291441053366,
      "objective/train/tokens_used": 54067200,
      "objective/train/unlikelihood_loss": 0.06622930616140366,
      "theoretical_loss": 5.216291441053366,
      "tokens_seen": 54067200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967703109327984,
      "loss": 2.5708,
      "theoretical_loss": 5.215339367345955,
      "tokens_seen": 54132736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967502507522568,
      "loss": 2.3715,
      "theoretical_loss": 5.2134396385493815,
      "tokens_seen": 54263808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967301905717151,
      "loss": 2.1358,
      "theoretical_loss": 5.211545774249233,
      "tokens_seen": 54394880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967101303911735,
      "loss": 2.4727,
      "theoretical_loss": 5.209657742275683,
      "tokens_seen": 54525952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496690070210632,
      "loss": 2.102,
      "theoretical_loss": 5.207775510712159,
      "tokens_seen": 54657024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966700100300903,
      "loss": 2.2568,
      "theoretical_loss": 5.205899047892753,
      "tokens_seen": 54788096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966499498495487,
      "loss": 1.8779,
      "theoretical_loss": 5.204028322399658,
      "tokens_seen": 54919168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496629889669007,
      "loss": 2.4808,
      "theoretical_loss": 5.202163303060633,
      "tokens_seen": 55050240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966098294884654,
      "loss": 2.1792,
      "theoretical_loss": 5.20030395894651,
      "tokens_seen": 55181312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965897693079238,
      "loss": 2.3935,
      "theoretical_loss": 5.198450259368721,
      "tokens_seen": 55312384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965697091273821,
      "loss": 2.2388,
      "theoretical_loss": 5.196602173876867,
      "tokens_seen": 55443456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965496489468405,
      "loss": 2.209,
      "theoretical_loss": 5.194759672256309,
      "tokens_seen": 55574528
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.34781280159950256,
      "objective/train/avg_token_score": 0.039099015295505524,
      "objective/train/docs_used": 44116,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9161739349365234,
      "objective/train/original_loss": 5.530422687530518,
      "objective/train/theoretical_loss": 5.192922724525789,
      "objective/train/tokens_used": 55705600,
      "objective/train/unlikelihood_loss": 0.06317880749702454,
      "theoretical_loss": 5.192922724525789,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965295887662988,
      "loss": 2.2517,
      "theoretical_loss": 5.192922724525789,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965095285857573,
      "loss": 2.2552,
      "theoretical_loss": 5.19109130093509,
      "tokens_seen": 55836672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964894684052157,
      "loss": 2.5353,
      "theoretical_loss": 5.189265371962712,
      "tokens_seen": 55967744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496469408224674,
      "loss": 2.2548,
      "theoretical_loss": 5.187444908313586,
      "tokens_seen": 56098816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964493480441324,
      "loss": 2.3618,
      "theoretical_loss": 5.185629880916814,
      "tokens_seen": 56229888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964292878635908,
      "loss": 1.9889,
      "theoretical_loss": 5.18382026092343,
      "tokens_seen": 56360960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964092276830491,
      "loss": 2.6223,
      "theoretical_loss": 5.182016019704204,
      "tokens_seen": 56492032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963891675025075,
      "loss": 2.2145,
      "theoretical_loss": 5.180217128847451,
      "tokens_seen": 56623104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963691073219659,
      "loss": 2.0782,
      "theoretical_loss": 5.178423560156894,
      "tokens_seen": 56754176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963490471414242,
      "loss": 2.1989,
      "theoretical_loss": 5.176635285649521,
      "tokens_seen": 56885248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963289869608827,
      "loss": 2.174,
      "theoretical_loss": 5.174852277553498,
      "tokens_seen": 57016320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496308926780341,
      "loss": 2.2559,
      "theoretical_loss": 5.17307450830609,
      "tokens_seen": 57147392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962888665997994,
      "loss": 2.221,
      "theoretical_loss": 5.1713019505516105,
      "tokens_seen": 57278464
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.39821603894233704,
      "objective/train/avg_token_score": 0.009783880785107613,
      "objective/train/docs_used": 44698,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.974114179611206,
      "objective/train/original_loss": 5.170925617218018,
      "objective/train/theoretical_loss": 5.170417617491682,
      "objective/train/tokens_used": 57344000,
      "objective/train/unlikelihood_loss": 0.07565154135227203,
      "theoretical_loss": 5.170417617491682,
      "tokens_seen": 57344000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962688064192578,
      "loss": 2.1205,
      "theoretical_loss": 5.169534577139395,
      "tokens_seen": 57409536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962487462387161,
      "loss": 1.8749,
      "theoretical_loss": 5.167772361121805,
      "tokens_seen": 57540608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962286860581746,
      "loss": 2.1519,
      "theoretical_loss": 5.1660152757522475,
      "tokens_seen": 57671680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496208625877633,
      "loss": 2.2882,
      "theoretical_loss": 5.164263294483226,
      "tokens_seen": 57802752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961885656970913,
      "loss": 2.175,
      "theoretical_loss": 5.162516390964408,
      "tokens_seen": 57933824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961685055165497,
      "loss": 1.834,
      "theoretical_loss": 5.160774539040716,
      "tokens_seen": 58064896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961484453360081,
      "loss": 2.0703,
      "theoretical_loss": 5.159037712750455,
      "tokens_seen": 58195968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961283851554664,
      "loss": 2.0747,
      "theoretical_loss": 5.157305886323435,
      "tokens_seen": 58327040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961083249749248,
      "loss": 2.2992,
      "theoretical_loss": 5.155579034179144,
      "tokens_seen": 58458112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960882647943831,
      "loss": 2.192,
      "theoretical_loss": 5.153857130924929,
      "tokens_seen": 58589184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960682046138415,
      "loss": 2.4796,
      "theoretical_loss": 5.152140151354191,
      "tokens_seen": 58720256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960481444333,
      "loss": 2.053,
      "theoretical_loss": 5.150428070444621,
      "tokens_seen": 58851328
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.42919111251831055,
      "objective/train/avg_token_score": 0.023179521784186363,
      "objective/train/docs_used": 46033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9980182647705078,
      "objective/train/original_loss": 5.299570560455322,
      "objective/train/theoretical_loss": 5.1487208633564405,
      "objective/train/tokens_used": 58982400,
      "objective/train/unlikelihood_loss": 0.05876965448260307,
      "theoretical_loss": 5.1487208633564405,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960280842527583,
      "loss": 2.0719,
      "theoretical_loss": 5.1487208633564405,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960080240722167,
      "loss": 2.1807,
      "theoretical_loss": 5.147018505430666,
      "tokens_seen": 59113472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959879638916751,
      "loss": 1.8661,
      "theoretical_loss": 5.145320972187402,
      "tokens_seen": 59244544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959679037111334,
      "loss": 2.313,
      "theoretical_loss": 5.143628239324139,
      "tokens_seen": 59375616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959478435305918,
      "loss": 2.3328,
      "theoretical_loss": 5.1419402827140885,
      "tokens_seen": 59506688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959277833500501,
      "loss": 2.192,
      "theoretical_loss": 5.140257078404524,
      "tokens_seen": 59637760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959077231695085,
      "loss": 2.0898,
      "theoretical_loss": 5.138578602615146,
      "tokens_seen": 59768832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958876629889669,
      "loss": 2.1709,
      "theoretical_loss": 5.1369048317364685,
      "tokens_seen": 59899904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958676028084253,
      "loss": 2.0217,
      "theoretical_loss": 5.135235742328217,
      "tokens_seen": 60030976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958475426278837,
      "loss": 2.1712,
      "theoretical_loss": 5.133571311117755,
      "tokens_seen": 60162048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958274824473421,
      "loss": 2.1534,
      "theoretical_loss": 5.131911514998518,
      "tokens_seen": 60293120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958074222668004,
      "loss": 2.0424,
      "theoretical_loss": 5.130256331028474,
      "tokens_seen": 60424192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957873620862588,
      "loss": 2.0909,
      "theoretical_loss": 5.128605736428597,
      "tokens_seen": 60555264
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.31408846378326416,
      "objective/train/avg_token_score": 0.03549068421125412,
      "objective/train/docs_used": 46630,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6117610931396484,
      "objective/train/original_loss": 5.108785152435303,
      "objective/train/theoretical_loss": 5.127782153068225,
      "objective/train/tokens_used": 60620800,
      "objective/train/unlikelihood_loss": 0.07724561542272568,
      "theoretical_loss": 5.127782153068225,
      "tokens_seen": 60620800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957673019057172,
      "loss": 2.0965,
      "theoretical_loss": 5.126959708581356,
      "tokens_seen": 60686336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957472417251755,
      "loss": 2.104,
      "theoretical_loss": 5.125318225029231,
      "tokens_seen": 60817408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957271815446339,
      "loss": 2.107,
      "theoretical_loss": 5.12368126347323,
      "tokens_seen": 60948480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957071213640923,
      "loss": 2.6384,
      "theoretical_loss": 5.122048801771443,
      "tokens_seen": 61079552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956870611835507,
      "loss": 2.4012,
      "theoretical_loss": 5.120420817937591,
      "tokens_seen": 61210624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956670010030091,
      "loss": 2.2013,
      "theoretical_loss": 5.118797290139605,
      "tokens_seen": 61341696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956469408224674,
      "loss": 2.0302,
      "theoretical_loss": 5.1171781966982195,
      "tokens_seen": 61472768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956268806419258,
      "loss": 2.0847,
      "theoretical_loss": 5.11556351608558,
      "tokens_seen": 61603840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956068204613842,
      "loss": 2.2934,
      "theoretical_loss": 5.113953226923864,
      "tokens_seen": 61734912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955867602808425,
      "loss": 2.5557,
      "theoretical_loss": 5.112347307983919,
      "tokens_seen": 61865984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955667001003009,
      "loss": 2.1684,
      "theoretical_loss": 5.110745738183919,
      "tokens_seen": 61997056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955466399197592,
      "loss": 2.2634,
      "theoretical_loss": 5.109148496588032,
      "tokens_seen": 62128128
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.44507575035095215,
      "objective/train/avg_token_score": 0.011363647878170013,
      "objective/train/docs_used": 47896,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.073281764984131,
      "objective/train/original_loss": 4.909066200256348,
      "objective/train/theoretical_loss": 5.107555562405102,
      "objective/train/tokens_used": 62259200,
      "objective/train/unlikelihood_loss": 0.08025174587965012,
      "theoretical_loss": 5.107555562405102,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955265797392177,
      "loss": 2.0933,
      "theoretical_loss": 5.107555562405102,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955065195586761,
      "loss": 2.0567,
      "theoretical_loss": 5.105966914987349,
      "tokens_seen": 62390272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954864593781344,
      "loss": 2.1778,
      "theoretical_loss": 5.10438253382908,
      "tokens_seen": 62521344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954663991975928,
      "loss": 2.406,
      "theoretical_loss": 5.102802398565418,
      "tokens_seen": 62652416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954463390170512,
      "loss": 2.2442,
      "theoretical_loss": 5.101226488971042,
      "tokens_seen": 62783488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954262788365095,
      "loss": 1.9896,
      "theoretical_loss": 5.09965478495894,
      "tokens_seen": 62914560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954062186559679,
      "loss": 1.906,
      "theoretical_loss": 5.09808726657918,
      "tokens_seen": 63045632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953861584754263,
      "loss": 2.3592,
      "theoretical_loss": 5.096523914017688,
      "tokens_seen": 63176704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953660982948846,
      "loss": 2.2084,
      "theoretical_loss": 5.094964707595047,
      "tokens_seen": 63307776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953460381143431,
      "loss": 2.1077,
      "theoretical_loss": 5.093409627765306,
      "tokens_seen": 63438848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953259779338014,
      "loss": 2.1148,
      "theoretical_loss": 5.091858655114796,
      "tokens_seen": 63569920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953059177532598,
      "loss": 2.2873,
      "theoretical_loss": 5.090311770360971,
      "tokens_seen": 63700992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952858575727182,
      "loss": 2.1432,
      "theoretical_loss": 5.088768954351249,
      "tokens_seen": 63832064
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.5087976455688477,
      "objective/train/avg_token_score": 0.013270500116050243,
      "objective/train/docs_used": 48551,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.6218936443328857,
      "objective/train/original_loss": 5.463879108428955,
      "objective/train/theoretical_loss": 5.087999066175502,
      "objective/train/tokens_used": 63897600,
      "objective/train/unlikelihood_loss": 0.0694500133395195,
      "theoretical_loss": 5.087999066175502,
      "tokens_seen": 63897600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952657973921765,
      "loss": 2.1419,
      "theoretical_loss": 5.0872301880618735,
      "tokens_seen": 63963136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952457372116349,
      "loss": 1.9254,
      "theoretical_loss": 5.085695452596788,
      "tokens_seen": 64094208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952256770310933,
      "loss": 1.9951,
      "theoretical_loss": 5.0841647291865115,
      "tokens_seen": 64225280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952056168505516,
      "loss": 2.2865,
      "theoretical_loss": 5.082637999187046,
      "tokens_seen": 64356352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049518555667001,
      "loss": 2.0893,
      "theoretical_loss": 5.0811152440787755,
      "tokens_seen": 64487424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951654964894685,
      "loss": 2.2674,
      "theoretical_loss": 5.079596445465386,
      "tokens_seen": 64618496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951454363089268,
      "loss": 2.0007,
      "theoretical_loss": 5.078081585072802,
      "tokens_seen": 64749568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951253761283852,
      "loss": 2.0058,
      "theoretical_loss": 5.076570644748123,
      "tokens_seen": 64880640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951053159478435,
      "loss": 1.895,
      "theoretical_loss": 5.075063606458576,
      "tokens_seen": 65011712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950852557673019,
      "loss": 2.2423,
      "theoretical_loss": 5.07356045229049,
      "tokens_seen": 65142784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950651955867603,
      "loss": 2.0065,
      "theoretical_loss": 5.072061164448261,
      "tokens_seen": 65273856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950451354062186,
      "loss": 2.1188,
      "theoretical_loss": 5.070565725253344,
      "tokens_seen": 65404928
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.34683528542518616,
      "objective/train/avg_token_score": 0.024793358519673347,
      "objective/train/docs_used": 49128,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.808795690536499,
      "objective/train/original_loss": 5.487620830535889,
      "objective/train/theoretical_loss": 5.069074117143246,
      "objective/train/tokens_used": 65536000,
      "objective/train/unlikelihood_loss": 0.0768253281712532,
      "theoretical_loss": 5.069074117143246,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495025075225677,
      "loss": 1.8184,
      "theoretical_loss": 5.069074117143246,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950050150451354,
      "loss": 1.9084,
      "theoretical_loss": 5.067586322670541,
      "tokens_seen": 65667072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949849548645938,
      "loss": 2.3579,
      "theoretical_loss": 5.066102324501883,
      "tokens_seen": 65798144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949648946840522,
      "loss": 2.1562,
      "theoretical_loss": 5.064622105417033,
      "tokens_seen": 65929216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949448345035105,
      "loss": 2.0672,
      "theoretical_loss": 5.063145648307904,
      "tokens_seen": 66060288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949247743229689,
      "loss": 1.9244,
      "theoretical_loss": 5.061672936177604,
      "tokens_seen": 66191360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949047141424273,
      "loss": 1.8588,
      "theoretical_loss": 5.060203952139497,
      "tokens_seen": 66322432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948846539618856,
      "loss": 1.9714,
      "theoretical_loss": 5.0587386794162725,
      "tokens_seen": 66453504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494864593781344,
      "loss": 2.1933,
      "theoretical_loss": 5.057277101339023,
      "tokens_seen": 66584576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948445336008024,
      "loss": 1.9907,
      "theoretical_loss": 5.055819201346331,
      "tokens_seen": 66715648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948244734202607,
      "loss": 2.2394,
      "theoretical_loss": 5.054364962983367,
      "tokens_seen": 66846720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948044132397192,
      "loss": 2.1289,
      "theoretical_loss": 5.052914369900997,
      "tokens_seen": 66977792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947843530591776,
      "loss": 2.362,
      "theoretical_loss": 5.051467405854897,
      "tokens_seen": 67108864
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.3599095940589905,
      "objective/train/avg_token_score": 0.01841607131063938,
      "objective/train/docs_used": 50658,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9209699630737305,
      "objective/train/original_loss": 5.435420036315918,
      "objective/train/theoretical_loss": 5.05074527967254,
      "objective/train/tokens_used": 67174400,
      "objective/train/unlikelihood_loss": 0.08842965960502625,
      "theoretical_loss": 5.05074527967254,
      "tokens_seen": 67174400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947642928786359,
      "loss": 2.086,
      "theoretical_loss": 5.050024054704677,
      "tokens_seen": 67239936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947442326980943,
      "loss": 1.9711,
      "theoretical_loss": 5.048584300413019,
      "tokens_seen": 67371008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947241725175527,
      "loss": 2.2735,
      "theoretical_loss": 5.0471481270448155,
      "tokens_seen": 67502080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947041123370111,
      "loss": 2.0652,
      "theoretical_loss": 5.045715518766322,
      "tokens_seen": 67633152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946840521564695,
      "loss": 1.9284,
      "theoretical_loss": 5.044286459844319,
      "tokens_seen": 67764224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946639919759278,
      "loss": 2.1492,
      "theoretical_loss": 5.0428609346452795,
      "tokens_seen": 67895296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946439317953862,
      "loss": 2.3224,
      "theoretical_loss": 5.041438927634549,
      "tokens_seen": 68026368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946238716148446,
      "loss": 2.1179,
      "theoretical_loss": 5.040020423375525,
      "tokens_seen": 68157440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946038114343029,
      "loss": 2.1086,
      "theoretical_loss": 5.038605406528857,
      "tokens_seen": 68288512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945837512537613,
      "loss": 2.042,
      "theoretical_loss": 5.037193861851646,
      "tokens_seen": 68419584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945636910732196,
      "loss": 2.3074,
      "theoretical_loss": 5.035785774196654,
      "tokens_seen": 68550656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945436308926781,
      "loss": 2.2202,
      "theoretical_loss": 5.034381128511525,
      "tokens_seen": 68681728
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.3374877870082855,
      "objective/train/avg_token_score": 0.023196162655949593,
      "objective/train/docs_used": 51428,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4241688251495361,
      "objective/train/original_loss": 4.484879493713379,
      "objective/train/theoretical_loss": 5.032979909838007,
      "objective/train/tokens_used": 68812800,
      "objective/train/unlikelihood_loss": 0.11043024063110352,
      "theoretical_loss": 5.032979909838007,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945235707121365,
      "loss": 1.958,
      "theoretical_loss": 5.032979909838007,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945035105315948,
      "loss": 2.023,
      "theoretical_loss": 5.031582103311187,
      "tokens_seen": 68943872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944834503510532,
      "loss": 1.984,
      "theoretical_loss": 5.030187694158739,
      "tokens_seen": 69074944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944633901705116,
      "loss": 2.2235,
      "theoretical_loss": 5.028796667700159,
      "tokens_seen": 69206016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944433299899699,
      "loss": 2.0468,
      "theoretical_loss": 5.0274090093460355,
      "tokens_seen": 69337088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944232698094283,
      "loss": 2.2206,
      "theoretical_loss": 5.0260247045973045,
      "tokens_seen": 69468160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944032096288867,
      "loss": 2.1891,
      "theoretical_loss": 5.024643739044526,
      "tokens_seen": 69599232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494383149448345,
      "loss": 2.2133,
      "theoretical_loss": 5.023266098367161,
      "tokens_seen": 69730304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943630892678035,
      "loss": 2.1522,
      "theoretical_loss": 5.0218917683328534,
      "tokens_seen": 69861376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943430290872618,
      "loss": 1.9963,
      "theoretical_loss": 5.020520734796728,
      "tokens_seen": 69992448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943229689067202,
      "loss": 2.0446,
      "theoretical_loss": 5.019152983700687,
      "tokens_seen": 70123520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943029087261786,
      "loss": 2.1678,
      "theoretical_loss": 5.01778850107272,
      "tokens_seen": 70254592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942828485456369,
      "loss": 2.3371,
      "theoretical_loss": 5.016427273026212,
      "tokens_seen": 70385664
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.4471529722213745,
      "objective/train/avg_token_score": 0.013587483204901218,
      "objective/train/docs_used": 52769,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.070784568786621,
      "objective/train/original_loss": 5.132371425628662,
      "objective/train/theoretical_loss": 5.015747875154622,
      "objective/train/tokens_used": 70451200,
      "objective/train/unlikelihood_loss": 0.07507549971342087,
      "theoretical_loss": 5.015747875154622,
      "tokens_seen": 70451200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942627883650953,
      "loss": 1.8887,
      "theoretical_loss": 5.015069285759269,
      "tokens_seen": 70516736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942427281845537,
      "loss": 2.0888,
      "theoretical_loss": 5.0137145255540405,
      "tokens_seen": 70647808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494222668004012,
      "loss": 2.0747,
      "theoretical_loss": 5.012362978776057,
      "tokens_seen": 70778880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942026078234704,
      "loss": 1.994,
      "theoretical_loss": 5.011014631873566,
      "tokens_seen": 70909952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941825476429289,
      "loss": 2.2445,
      "theoretical_loss": 5.009669471376882,
      "tokens_seen": 71041024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941624874623872,
      "loss": 2.2045,
      "theoretical_loss": 5.008327483897736,
      "tokens_seen": 71172096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941424272818456,
      "loss": 2.0795,
      "theoretical_loss": 5.006988656128635,
      "tokens_seen": 71303168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941223671013039,
      "loss": 2.2727,
      "theoretical_loss": 5.00565297484223,
      "tokens_seen": 71434240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941023069207623,
      "loss": 2.0863,
      "theoretical_loss": 5.004320426890686,
      "tokens_seen": 71565312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940822467402207,
      "loss": 2.2777,
      "theoretical_loss": 5.002990999205057,
      "tokens_seen": 71696384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494062186559679,
      "loss": 2.0031,
      "theoretical_loss": 5.001664678794671,
      "tokens_seen": 71827456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940421263791374,
      "loss": 2.0639,
      "theoretical_loss": 5.0003414527465235,
      "tokens_seen": 71958528
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.4258919954299927,
      "objective/train/avg_token_score": 0.020259426906704903,
      "objective/train/docs_used": 53229,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.168090343475342,
      "objective/train/original_loss": 5.140798091888428,
      "objective/train/theoretical_loss": 4.999021308224664,
      "objective/train/tokens_used": 72089600,
      "objective/train/unlikelihood_loss": 0.11368393152952194,
      "theoretical_loss": 4.999021308224664,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940220661985958,
      "loss": 2.0618,
      "theoretical_loss": 4.999021308224664,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940020060180542,
      "loss": 1.9749,
      "theoretical_loss": 4.997704232469606,
      "tokens_seen": 72220672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939819458375126,
      "loss": 2.1991,
      "theoretical_loss": 4.996390212797728,
      "tokens_seen": 72351744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493961885656971,
      "loss": 2.288,
      "theoretical_loss": 4.995079236600686,
      "tokens_seen": 72482816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939418254764293,
      "loss": 1.9311,
      "theoretical_loss": 4.993771291344839,
      "tokens_seen": 72613888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939217652958877,
      "loss": 1.9251,
      "theoretical_loss": 4.992466364570659,
      "tokens_seen": 72744960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493901705115346,
      "loss": 1.8568,
      "theoretical_loss": 4.991164443892175,
      "tokens_seen": 72876032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938816449348044,
      "loss": 2.2996,
      "theoretical_loss": 4.989865516996396,
      "tokens_seen": 73007104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938615847542628,
      "loss": 2.1793,
      "theoretical_loss": 4.988569571642756,
      "tokens_seen": 73138176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938415245737211,
      "loss": 2.0888,
      "theoretical_loss": 4.9872765956625615,
      "tokens_seen": 73269248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938214643931796,
      "loss": 2.0231,
      "theoretical_loss": 4.9859865769584335,
      "tokens_seen": 73400320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493801404212638,
      "loss": 2.1265,
      "theoretical_loss": 4.984699503503771,
      "tokens_seen": 73531392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937813440320963,
      "loss": 2.1259,
      "theoretical_loss": 4.9834153633422105,
      "tokens_seen": 73662464
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.43444526195526123,
      "objective/train/avg_token_score": 0.0062065767124295235,
      "objective/train/docs_used": 53865,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.2150537967681885,
      "objective/train/original_loss": 5.1299591064453125,
      "objective/train/theoretical_loss": 4.982774389529053,
      "objective/train/tokens_used": 73728000,
      "objective/train/unlikelihood_loss": 0.08083882927894592,
      "theoretical_loss": 4.982774389529053,
      "tokens_seen": 73728000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937612838515547,
      "loss": 1.9495,
      "theoretical_loss": 4.9821341445870875,
      "tokens_seen": 73793536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493741223671013,
      "loss": 1.9215,
      "theoretical_loss": 4.98085583542091,
      "tokens_seen": 73924608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937211634904714,
      "loss": 2.021,
      "theoretical_loss": 4.979580424094836,
      "tokens_seen": 74055680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937011033099298,
      "loss": 1.8973,
      "theoretical_loss": 4.978307898928149,
      "tokens_seen": 74186752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936810431293881,
      "loss": 2.1481,
      "theoretical_loss": 4.97703824830775,
      "tokens_seen": 74317824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936609829488465,
      "loss": 1.8232,
      "theoretical_loss": 4.975771460687641,
      "tokens_seen": 74448896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493640922768305,
      "loss": 1.8557,
      "theoretical_loss": 4.974507524588424,
      "tokens_seen": 74579968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936208625877633,
      "loss": 1.8265,
      "theoretical_loss": 4.973246428596802,
      "tokens_seen": 74711040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936008024072217,
      "loss": 2.192,
      "theoretical_loss": 4.971988161365077,
      "tokens_seen": 74842112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049358074222668,
      "loss": 1.9407,
      "theoretical_loss": 4.970732711610667,
      "tokens_seen": 74973184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935606820461384,
      "loss": 2.2415,
      "theoretical_loss": 4.96948006811561,
      "tokens_seen": 75104256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935406218655968,
      "loss": 2.0764,
      "theoretical_loss": 4.968230219726093,
      "tokens_seen": 75235328
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.41452834010124207,
      "objective/train/avg_token_score": 0.014995137229561806,
      "objective/train/docs_used": 54881,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9440230131149292,
      "objective/train/original_loss": 4.934903144836426,
      "objective/train/theoretical_loss": 4.966983155351962,
      "objective/train/tokens_used": 75366400,
      "objective/train/unlikelihood_loss": 0.07033995538949966,
      "theoretical_loss": 4.966983155351962,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935205616850551,
      "loss": 2.1972,
      "theoretical_loss": 4.966983155351962,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935005015045135,
      "loss": 2.145,
      "theoretical_loss": 4.96573886396626,
      "tokens_seen": 75497472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934804413239719,
      "loss": 1.9567,
      "theoretical_loss": 4.964497334604748,
      "tokens_seen": 75628544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934603811434303,
      "loss": 1.9129,
      "theoretical_loss": 4.963258556365449,
      "tokens_seen": 75759616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934403209628887,
      "loss": 2.0695,
      "theoretical_loss": 4.962022518408183,
      "tokens_seen": 75890688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493420260782347,
      "loss": 2.2744,
      "theoretical_loss": 4.9607892099541075,
      "tokens_seen": 76021760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934002006018054,
      "loss": 2.2838,
      "theoretical_loss": 4.959558620285274,
      "tokens_seen": 76152832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933801404212638,
      "loss": 1.9697,
      "theoretical_loss": 4.958330738744172,
      "tokens_seen": 76283904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933600802407221,
      "loss": 2.2717,
      "theoretical_loss": 4.957105554733289,
      "tokens_seen": 76414976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933400200601805,
      "loss": 2.0457,
      "theoretical_loss": 4.955883057714669,
      "tokens_seen": 76546048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933199598796389,
      "loss": 2.0651,
      "theoretical_loss": 4.954663237209477,
      "tokens_seen": 76677120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932998996990972,
      "loss": 2.0885,
      "theoretical_loss": 4.9534460827975675,
      "tokens_seen": 76808192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932798395185557,
      "loss": 2.2084,
      "theoretical_loss": 4.952231584117056,
      "tokens_seen": 76939264
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.3369990289211273,
      "objective/train/avg_token_score": 0.013007127679884434,
      "objective/train/docs_used": 55424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6116856336593628,
      "objective/train/original_loss": 4.758896350860596,
      "objective/train/theoretical_loss": 4.951625327453812,
      "objective/train/tokens_used": 77004800,
      "objective/train/unlikelihood_loss": 0.08679111301898956,
      "theoretical_loss": 4.951625327453812,
      "tokens_seen": 77004800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932597793380141,
      "loss": 1.8865,
      "theoretical_loss": 4.951019730863894,
      "tokens_seen": 77070336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932397191574724,
      "loss": 2.1724,
      "theoretical_loss": 4.94981051279145,
      "tokens_seen": 77201408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932196589769308,
      "loss": 2.1451,
      "theoretical_loss": 4.948603919710088,
      "tokens_seen": 77332480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931995987963893,
      "loss": 1.9967,
      "theoretical_loss": 4.947399941486762,
      "tokens_seen": 77463552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931795386158476,
      "loss": 2.005,
      "theoretical_loss": 4.946198568044602,
      "tokens_seen": 77594624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493159478435306,
      "loss": 2.2193,
      "theoretical_loss": 4.944999789362508,
      "tokens_seen": 77725696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931394182547643,
      "loss": 1.9879,
      "theoretical_loss": 4.9438035954747495,
      "tokens_seen": 77856768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931193580742227,
      "loss": 2.0917,
      "theoretical_loss": 4.942609976470566,
      "tokens_seen": 77987840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930992978936811,
      "loss": 2.0912,
      "theoretical_loss": 4.941418922493774,
      "tokens_seen": 78118912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930792377131394,
      "loss": 2.1054,
      "theoretical_loss": 4.940230423742372,
      "tokens_seen": 78249984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930591775325978,
      "loss": 2.302,
      "theoretical_loss": 4.939044470468156,
      "tokens_seen": 78381056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930391173520562,
      "loss": 2.1726,
      "theoretical_loss": 4.937861052976332,
      "tokens_seen": 78512128
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.4615102708339691,
      "objective/train/avg_token_score": 0.007493212353438139,
      "objective/train/docs_used": 56825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.250938892364502,
      "objective/train/original_loss": 4.587989330291748,
      "objective/train/theoretical_loss": 4.9366801616251355,
      "objective/train/tokens_used": 78643200,
      "objective/train/unlikelihood_loss": 0.1308807134628296,
      "theoretical_loss": 4.9366801616251355,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930190571715146,
      "loss": 2.0669,
      "theoretical_loss": 4.9366801616251355,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492998996990973,
      "loss": 1.8013,
      "theoretical_loss": 4.935501786825457,
      "tokens_seen": 78774272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929789368104313,
      "loss": 1.9622,
      "theoretical_loss": 4.934325919040461,
      "tokens_seen": 78905344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929588766298897,
      "loss": 1.9327,
      "theoretical_loss": 4.933152548785222,
      "tokens_seen": 79036416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929388164493481,
      "loss": 2.1009,
      "theoretical_loss": 4.931981666626351,
      "tokens_seen": 79167488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929187562688064,
      "loss": 2.2199,
      "theoretical_loss": 4.930813263181631,
      "tokens_seen": 79298560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928986960882648,
      "loss": 1.9262,
      "theoretical_loss": 4.929647329119659,
      "tokens_seen": 79429632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928786359077232,
      "loss": 2.1295,
      "theoretical_loss": 4.928483855159485,
      "tokens_seen": 79560704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928585757271815,
      "loss": 2.0885,
      "theoretical_loss": 4.9273228320702565,
      "tokens_seen": 79691776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049283851554664,
      "loss": 2.2262,
      "theoretical_loss": 4.926164250670868,
      "tokens_seen": 79822848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928184553660984,
      "loss": 2.1799,
      "theoretical_loss": 4.925008101829608,
      "tokens_seen": 79953920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927983951855567,
      "loss": 2.0411,
      "theoretical_loss": 4.923854376463816,
      "tokens_seen": 80084992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927783350050151,
      "loss": 2.0662,
      "theoretical_loss": 4.92270306553954,
      "tokens_seen": 80216064
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.46413734555244446,
      "objective/train/avg_token_score": 0.03281039744615555,
      "objective/train/docs_used": 57296,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.3605399131774902,
      "objective/train/original_loss": 5.119383811950684,
      "objective/train/theoretical_loss": 4.922128312683096,
      "objective/train/tokens_used": 80281600,
      "objective/train/unlikelihood_loss": 0.08986959606409073,
      "theoretical_loss": 4.922128312683096,
      "tokens_seen": 80281600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927582748244734,
      "loss": 1.9299,
      "theoretical_loss": 4.921554160071194,
      "tokens_seen": 80347136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927382146439318,
      "loss": 2.0743,
      "theoretical_loss": 4.920407651121222,
      "tokens_seen": 80478208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927181544633902,
      "loss": 2.0166,
      "theoretical_loss": 4.919263529799759,
      "tokens_seen": 80609280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926980942828485,
      "loss": 1.9821,
      "theoretical_loss": 4.91812178726431,
      "tokens_seen": 80740352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926780341023069,
      "loss": 1.962,
      "theoretical_loss": 4.916982414719408,
      "tokens_seen": 80871424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926579739217654,
      "loss": 1.8903,
      "theoretical_loss": 4.915845403416299,
      "tokens_seen": 81002496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926379137412237,
      "loss": 2.0539,
      "theoretical_loss": 4.914710744652614,
      "tokens_seen": 81133568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926178535606821,
      "loss": 2.0817,
      "theoretical_loss": 4.913578429772047,
      "tokens_seen": 81264640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925977933801404,
      "loss": 2.0058,
      "theoretical_loss": 4.912448450164041,
      "tokens_seen": 81395712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925777331995988,
      "loss": 1.741,
      "theoretical_loss": 4.911320797263471,
      "tokens_seen": 81526784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925576730190572,
      "loss": 2.1737,
      "theoretical_loss": 4.910195462550334,
      "tokens_seen": 81657856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925376128385155,
      "loss": 2.246,
      "theoretical_loss": 4.909072437549434,
      "tokens_seen": 81788928
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.46144917607307434,
      "objective/train/avg_token_score": 0.00516983587294817,
      "objective/train/docs_used": 58605,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.2914438247680664,
      "objective/train/original_loss": 4.962143421173096,
      "objective/train/theoretical_loss": 4.907951713830082,
      "objective/train/tokens_used": 81920000,
      "objective/train/unlikelihood_loss": 0.07921816408634186,
      "theoretical_loss": 4.907951713830082,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925175526579739,
      "loss": 2.07,
      "theoretical_loss": 4.907951713830082,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924974924774323,
      "loss": 2.0655,
      "theoretical_loss": 4.906833283005785,
      "tokens_seen": 82051072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924774322968907,
      "loss": 2.093,
      "theoretical_loss": 4.90571713673395,
      "tokens_seen": 82182144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924573721163491,
      "loss": 2.3312,
      "theoretical_loss": 4.904603266715578,
      "tokens_seen": 82313216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924373119358075,
      "loss": 2.0983,
      "theoretical_loss": 4.903491664694977,
      "tokens_seen": 82444288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924172517552658,
      "loss": 1.9503,
      "theoretical_loss": 4.902382322459456,
      "tokens_seen": 82575360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923971915747242,
      "loss": 2.1464,
      "theoretical_loss": 4.90127523183904,
      "tokens_seen": 82706432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923771313941825,
      "loss": 2.1856,
      "theoretical_loss": 4.900170384706181,
      "tokens_seen": 82837504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923570712136409,
      "loss": 1.9381,
      "theoretical_loss": 4.899067772975469,
      "tokens_seen": 82968576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923370110330993,
      "loss": 2.3138,
      "theoretical_loss": 4.897967388603346,
      "tokens_seen": 83099648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923169508525576,
      "loss": 2.0845,
      "theoretical_loss": 4.896869223587828,
      "tokens_seen": 83230720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922968906720161,
      "loss": 2.1112,
      "theoretical_loss": 4.895773269968219,
      "tokens_seen": 83361792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922768304914745,
      "loss": 1.9807,
      "theoretical_loss": 4.894679519824841,
      "tokens_seen": 83492864
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.4049975574016571,
      "objective/train/avg_token_score": 0.012781642377376556,
      "objective/train/docs_used": 59310,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9555177688598633,
      "objective/train/original_loss": 5.0520477294921875,
      "objective/train/theoretical_loss": 4.894133468592984,
      "objective/train/tokens_used": 83558400,
      "objective/train/unlikelihood_loss": 0.09169325232505798,
      "theoretical_loss": 4.894133468592984,
      "tokens_seen": 83558400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922567703109328,
      "loss": 1.7533,
      "theoretical_loss": 4.8935879652787495,
      "tokens_seen": 83623936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922367101303912,
      "loss": 1.9519,
      "theoretical_loss": 4.892498598491473,
      "tokens_seen": 83755008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922166499498495,
      "loss": 2.1345,
      "theoretical_loss": 4.891411411664727,
      "tokens_seen": 83886080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921965897693079,
      "loss": 2.1,
      "theoretical_loss": 4.890326397040158,
      "tokens_seen": 84017152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921765295887663,
      "loss": 2.0142,
      "theoretical_loss": 4.8892435468990705,
      "tokens_seen": 84148224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921564694082246,
      "loss": 2.0119,
      "theoretical_loss": 4.888162853562166,
      "tokens_seen": 84279296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492136409227683,
      "loss": 2.2767,
      "theoretical_loss": 4.88708430938928,
      "tokens_seen": 84410368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921163490471415,
      "loss": 2.1413,
      "theoretical_loss": 4.8860079067791204,
      "tokens_seen": 84541440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920962888665998,
      "loss": 2.0281,
      "theoretical_loss": 4.884933638169014,
      "tokens_seen": 84672512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920762286860582,
      "loss": 1.802,
      "theoretical_loss": 4.883861496034644,
      "tokens_seen": 84803584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920561685055166,
      "loss": 1.8336,
      "theoretical_loss": 4.8827914728898065,
      "tokens_seen": 84934656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920361083249749,
      "loss": 2.0471,
      "theoretical_loss": 4.881723561286149,
      "tokens_seen": 85065728
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.35856547951698303,
      "objective/train/avg_token_score": 0.008686823770403862,
      "objective/train/docs_used": 60492,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7955634593963623,
      "objective/train/original_loss": 4.565763473510742,
      "objective/train/theoretical_loss": 4.880657753812926,
      "objective/train/tokens_used": 85196800,
      "objective/train/unlikelihood_loss": 0.19698937237262726,
      "theoretical_loss": 4.880657753812926,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920160481444333,
      "loss": 1.9216,
      "theoretical_loss": 4.880657753812926,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919959879638916,
      "loss": 2.1886,
      "theoretical_loss": 4.879594043096755,
      "tokens_seen": 85327872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049197592778335,
      "loss": 2.0444,
      "theoretical_loss": 4.87853242180136,
      "tokens_seen": 85458944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919558676028084,
      "loss": 1.9928,
      "theoretical_loss": 4.877472882627343,
      "tokens_seen": 85590016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919358074222668,
      "loss": 1.9912,
      "theoretical_loss": 4.876415418311928,
      "tokens_seen": 85721088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919157472417252,
      "loss": 1.9497,
      "theoretical_loss": 4.875360021628733,
      "tokens_seen": 85852160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918956870611836,
      "loss": 2.1111,
      "theoretical_loss": 4.874306685387525,
      "tokens_seen": 85983232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918756268806419,
      "loss": 2.2713,
      "theoretical_loss": 4.87325540243399,
      "tokens_seen": 86114304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918555667001003,
      "loss": 2.0874,
      "theoretical_loss": 4.872206165649493,
      "tokens_seen": 86245376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918355065195586,
      "loss": 1.9654,
      "theoretical_loss": 4.871158967950852,
      "tokens_seen": 86376448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491815446339017,
      "loss": 2.0332,
      "theoretical_loss": 4.8701138022901045,
      "tokens_seen": 86507520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917953861584754,
      "loss": 2.2051,
      "theoretical_loss": 4.8690706616542805,
      "tokens_seen": 86638592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917753259779337,
      "loss": 2.0213,
      "theoretical_loss": 4.868029539065176,
      "tokens_seen": 86769664
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.45521751046180725,
      "objective/train/avg_token_score": 0.008304348215460777,
      "objective/train/docs_used": 61005,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1370649337768555,
      "objective/train/original_loss": 4.633736610412598,
      "objective/train/theoretical_loss": 4.867509732366907,
      "objective/train/tokens_used": 86835200,
      "objective/train/unlikelihood_loss": 0.09587273001670837,
      "theoretical_loss": 4.867509732366907,
      "tokens_seen": 86835200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917552657973922,
      "loss": 1.7873,
      "theoretical_loss": 4.866990427579129,
      "tokens_seen": 86900736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917352056168506,
      "loss": 1.9276,
      "theoretical_loss": 4.8659533202867955,
      "tokens_seen": 87031808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917151454363089,
      "loss": 1.7595,
      "theoretical_loss": 4.864918210312927,
      "tokens_seen": 87162880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916950852557673,
      "loss": 2.029,
      "theoretical_loss": 4.863885090816158,
      "tokens_seen": 87293952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916750250752258,
      "loss": 1.8672,
      "theoretical_loss": 4.862853954988781,
      "tokens_seen": 87425024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916549648946841,
      "loss": 1.9145,
      "theoretical_loss": 4.861824796056533,
      "tokens_seen": 87556096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916349047141425,
      "loss": 1.8864,
      "theoretical_loss": 4.860797607278385,
      "tokens_seen": 87687168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916148445336008,
      "loss": 2.0223,
      "theoretical_loss": 4.859772381946323,
      "tokens_seen": 87818240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915947843530592,
      "loss": 1.8535,
      "theoretical_loss": 4.858749113385144,
      "tokens_seen": 87949312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915747241725176,
      "loss": 1.8087,
      "theoretical_loss": 4.8577277949522415,
      "tokens_seen": 88080384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915546639919759,
      "loss": 1.8308,
      "theoretical_loss": 4.856708420037402,
      "tokens_seen": 88211456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915346038114343,
      "loss": 1.8766,
      "theoretical_loss": 4.8556909820625975,
      "tokens_seen": 88342528
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.44208210706710815,
      "objective/train/avg_token_score": 0.02312779426574707,
      "objective/train/docs_used": 61597,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.139789581298828,
      "objective/train/original_loss": 4.875903129577637,
      "objective/train/theoretical_loss": 4.854675474481779,
      "objective/train/tokens_used": 88473600,
      "objective/train/unlikelihood_loss": 0.07984444499015808,
      "theoretical_loss": 4.854675474481779,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915145436308927,
      "loss": 2.214,
      "theoretical_loss": 4.854675474481779,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914944834503511,
      "loss": 1.9395,
      "theoretical_loss": 4.85366189078068,
      "tokens_seen": 88604672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914744232698095,
      "loss": 1.9757,
      "theoretical_loss": 4.852650224476609,
      "tokens_seen": 88735744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914543630892679,
      "loss": 2.0743,
      "theoretical_loss": 4.851640469118255,
      "tokens_seen": 88866816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914343029087262,
      "loss": 2.041,
      "theoretical_loss": 4.850632618285486,
      "tokens_seen": 88997888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914142427281846,
      "loss": 2.0281,
      "theoretical_loss": 4.849626665589156,
      "tokens_seen": 89128960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913941825476429,
      "loss": 1.8351,
      "theoretical_loss": 4.8486226046709024,
      "tokens_seen": 89260032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913741223671013,
      "loss": 2.1069,
      "theoretical_loss": 4.847620429202967,
      "tokens_seen": 89391104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913540621865597,
      "loss": 1.7571,
      "theoretical_loss": 4.846620132887992,
      "tokens_seen": 89522176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491334002006018,
      "loss": 2.0421,
      "theoretical_loss": 4.845621709458831,
      "tokens_seen": 89653248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913139418254765,
      "loss": 1.859,
      "theoretical_loss": 4.844625152678364,
      "tokens_seen": 89784320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912938816449349,
      "loss": 2.0505,
      "theoretical_loss": 4.84363045633931,
      "tokens_seen": 89915392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912738214643932,
      "loss": 1.7393,
      "theoretical_loss": 4.84263761426404,
      "tokens_seen": 90046464
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.34017595648765564,
      "objective/train/avg_token_score": 0.00968579389154911,
      "objective/train/docs_used": 63098,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.632453441619873,
      "objective/train/original_loss": 4.724969387054443,
      "objective/train/theoretical_loss": 4.842141886652876,
      "objective/train/tokens_used": 90112000,
      "objective/train/unlikelihood_loss": 0.09992973506450653,
      "theoretical_loss": 4.842141886652876,
      "tokens_seen": 90112000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912537612838516,
      "loss": 2.0554,
      "theoretical_loss": 4.841646620304388,
      "tokens_seen": 90177536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049123370110331,
      "loss": 2.1008,
      "theoretical_loss": 4.840657468341476,
      "tokens_seen": 90308608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912136409227683,
      "loss": 1.9921,
      "theoretical_loss": 4.839670152285526,
      "tokens_seen": 90439680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911935807422267,
      "loss": 2.072,
      "theoretical_loss": 4.838684666075682,
      "tokens_seen": 90570752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491173520561685,
      "loss": 1.9095,
      "theoretical_loss": 4.837701003679829,
      "tokens_seen": 90701824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911534603811434,
      "loss": 1.877,
      "theoretical_loss": 4.836719159094422,
      "tokens_seen": 90832896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911334002006019,
      "loss": 1.9558,
      "theoretical_loss": 4.835739126344298,
      "tokens_seen": 90963968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911133400200602,
      "loss": 1.9039,
      "theoretical_loss": 4.834760899482514,
      "tokens_seen": 91095040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910932798395186,
      "loss": 2.1059,
      "theoretical_loss": 4.833784472590165,
      "tokens_seen": 91226112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491073219658977,
      "loss": 1.9337,
      "theoretical_loss": 4.832809839776213,
      "tokens_seen": 91357184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910531594784353,
      "loss": 2.2944,
      "theoretical_loss": 4.831836995177319,
      "tokens_seen": 91488256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910330992978937,
      "loss": 1.9808,
      "theoretical_loss": 4.8308659329576695,
      "tokens_seen": 91619328
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.19037145376205444,
      "objective/train/avg_token_score": 0.031226102262735367,
      "objective/train/docs_used": 63840,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9143338203430176,
      "objective/train/original_loss": 4.837706089019775,
      "objective/train/theoretical_loss": 4.8298966473088125,
      "objective/train/tokens_used": 91750400,
      "objective/train/unlikelihood_loss": 0.17378956079483032,
      "theoretical_loss": 4.8298966473088125,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491013039117352,
      "loss": 1.7083,
      "theoretical_loss": 4.8298966473088125,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909929789368104,
      "loss": 1.9804,
      "theoretical_loss": 4.8289291324494865,
      "tokens_seen": 91881472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909729187562688,
      "loss": 1.9059,
      "theoretical_loss": 4.827963382625454,
      "tokens_seen": 92012544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909528585757272,
      "loss": 1.9218,
      "theoretical_loss": 4.826999392109344,
      "tokens_seen": 92143616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909327983951856,
      "loss": 1.8156,
      "theoretical_loss": 4.826037155200478,
      "tokens_seen": 92274688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490912738214644,
      "loss": 1.9604,
      "theoretical_loss": 4.825076666224717,
      "tokens_seen": 92405760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908926780341023,
      "loss": 1.8492,
      "theoretical_loss": 4.824117919534297,
      "tokens_seen": 92536832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908726178535607,
      "loss": 2.0128,
      "theoretical_loss": 4.823160909507665,
      "tokens_seen": 92667904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490852557673019,
      "loss": 1.7759,
      "theoretical_loss": 4.822205630549329,
      "tokens_seen": 92798976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908324974924774,
      "loss": 1.9094,
      "theoretical_loss": 4.821252077089696,
      "tokens_seen": 92930048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908124373119358,
      "loss": 1.9877,
      "theoretical_loss": 4.820300243584913,
      "tokens_seen": 93061120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907923771313941,
      "loss": 1.8244,
      "theoretical_loss": 4.819350124516717,
      "tokens_seen": 93192192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907723169508526,
      "loss": 1.9063,
      "theoretical_loss": 4.818401714392279,
      "tokens_seen": 93323264
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.49071359634399414,
      "objective/train/avg_token_score": 0.014383784495294094,
      "objective/train/docs_used": 65224,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1994926929473877,
      "objective/train/original_loss": 4.691728115081787,
      "objective/train/theoretical_loss": 4.817928148474378,
      "objective/train/tokens_used": 93388800,
      "objective/train/unlikelihood_loss": 0.06607113778591156,
      "theoretical_loss": 4.817928148474378,
      "tokens_seen": 93388800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490752256770311,
      "loss": 1.9917,
      "theoretical_loss": 4.817455007744052,
      "tokens_seen": 93454336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907321965897693,
      "loss": 2.0181,
      "theoretical_loss": 4.816509999129618,
      "tokens_seen": 93585408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907121364092277,
      "loss": 1.9882,
      "theoretical_loss": 4.815566683131536,
      "tokens_seen": 93716480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490692076228686,
      "loss": 1.9407,
      "theoretical_loss": 4.814625054357199,
      "tokens_seen": 93847552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906720160481444,
      "loss": 1.7596,
      "theoretical_loss": 4.813685107438679,
      "tokens_seen": 93978624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906519558676028,
      "loss": 1.8142,
      "theoretical_loss": 4.812746837032582,
      "tokens_seen": 94109696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906318956870611,
      "loss": 1.6129,
      "theoretical_loss": 4.811810237819904,
      "tokens_seen": 94240768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906118355065195,
      "loss": 1.9627,
      "theoretical_loss": 4.810875304505881,
      "tokens_seen": 94371840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490591775325978,
      "loss": 1.9738,
      "theoretical_loss": 4.809942031819853,
      "tokens_seen": 94502912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905717151454363,
      "loss": 1.7409,
      "theoretical_loss": 4.809010414515113,
      "tokens_seen": 94633984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905516549648947,
      "loss": 2.0122,
      "theoretical_loss": 4.808080447368766,
      "tokens_seen": 94765056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905315947843531,
      "loss": 1.8492,
      "theoretical_loss": 4.807152125181597,
      "tokens_seen": 94896128
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.5782013535499573,
      "objective/train/avg_token_score": 0.0024078439455479383,
      "objective/train/docs_used": 65494,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.5631401538848877,
      "objective/train/original_loss": 4.700851917266846,
      "objective/train/theoretical_loss": 4.8062254427779205,
      "objective/train/tokens_used": 95027200,
      "objective/train/unlikelihood_loss": 0.061158839613199234,
      "theoretical_loss": 4.8062254427779205,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905115346038114,
      "loss": 1.9437,
      "theoretical_loss": 4.8062254427779205,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904914744232698,
      "loss": 1.9594,
      "theoretical_loss": 4.805300395005444,
      "tokens_seen": 95158272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904714142427281,
      "loss": 1.9803,
      "theoretical_loss": 4.8043769767351385,
      "tokens_seen": 95289344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904513540621865,
      "loss": 1.9562,
      "theoretical_loss": 4.803455182861087,
      "tokens_seen": 95420416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904312938816449,
      "loss": 2.0377,
      "theoretical_loss": 4.802535008300364,
      "tokens_seen": 95551488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904112337011033,
      "loss": 2.0507,
      "theoretical_loss": 4.801616447992888,
      "tokens_seen": 95682560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903911735205617,
      "loss": 2.073,
      "theoretical_loss": 4.8006994969013,
      "tokens_seen": 95813632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903711133400201,
      "loss": 2.0056,
      "theoretical_loss": 4.799784150010819,
      "tokens_seen": 95944704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903510531594784,
      "loss": 1.9311,
      "theoretical_loss": 4.798870402329115,
      "tokens_seen": 96075776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903309929789368,
      "loss": 1.7706,
      "theoretical_loss": 4.797958248886179,
      "tokens_seen": 96206848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903109327983952,
      "loss": 1.8998,
      "theoretical_loss": 4.797047684734192,
      "tokens_seen": 96337920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902908726178535,
      "loss": 1.9212,
      "theoretical_loss": 4.796138704947397,
      "tokens_seen": 96468992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902708124373119,
      "loss": 1.6952,
      "theoretical_loss": 4.795231304621968,
      "tokens_seen": 96600064
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.32086998224258423,
      "objective/train/avg_token_score": 0.05917775630950928,
      "objective/train/docs_used": 66055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3923687934875488,
      "objective/train/original_loss": 5.291764259338379,
      "objective/train/theoretical_loss": 4.794778195230787,
      "objective/train/tokens_used": 96665600,
      "objective/train/unlikelihood_loss": 0.06859153509140015,
      "theoretical_loss": 4.794778195230787,
      "tokens_seen": 96665600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902507522567703,
      "loss": 1.8366,
      "theoretical_loss": 4.794325478875885,
      "tokens_seen": 96731136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902306920762287,
      "loss": 2.0257,
      "theoretical_loss": 4.793421222848808,
      "tokens_seen": 96862208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902106318956871,
      "loss": 1.9104,
      "theoretical_loss": 4.792518531701948,
      "tokens_seen": 96993280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901905717151454,
      "loss": 1.9913,
      "theoretical_loss": 4.791617400617948,
      "tokens_seen": 97124352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901705115346038,
      "loss": 1.8546,
      "theoretical_loss": 4.790717824800755,
      "tokens_seen": 97255424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901504513540623,
      "loss": 1.8143,
      "theoretical_loss": 4.789819799475499,
      "tokens_seen": 97386496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901303911735206,
      "loss": 1.8901,
      "theoretical_loss": 4.788923319888369,
      "tokens_seen": 97517568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490110330992979,
      "loss": 1.8487,
      "theoretical_loss": 4.788028381306497,
      "tokens_seen": 97648640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900902708124374,
      "loss": 1.868,
      "theoretical_loss": 4.787134979017832,
      "tokens_seen": 97779712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900702106318957,
      "loss": 1.8523,
      "theoretical_loss": 4.786243108331024,
      "tokens_seen": 97910784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900501504513541,
      "loss": 1.7257,
      "theoretical_loss": 4.785352764575304,
      "tokens_seen": 98041856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900300902708124,
      "loss": 1.9306,
      "theoretical_loss": 4.784463943100367,
      "tokens_seen": 98172928
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.4745234549045563,
      "objective/train/avg_token_score": 0.020105870440602303,
      "objective/train/docs_used": 66835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.3115997314453125,
      "objective/train/original_loss": 4.947286128997803,
      "objective/train/theoretical_loss": 4.783576639276257,
      "objective/train/tokens_used": 98304000,
      "objective/train/unlikelihood_loss": 0.07566744834184647,
      "theoretical_loss": 4.783576639276257,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900100300902708,
      "loss": 1.7746,
      "theoretical_loss": 4.783576639276257,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899899699097292,
      "loss": 1.8378,
      "theoretical_loss": 4.782690848493245,
      "tokens_seen": 98435072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899699097291876,
      "loss": 2.0137,
      "theoretical_loss": 4.781806566161723,
      "tokens_seen": 98566144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489949849548646,
      "loss": 2.0675,
      "theoretical_loss": 4.78092378771208,
      "tokens_seen": 98697216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899297893681044,
      "loss": 2.0216,
      "theoretical_loss": 4.780042508594596,
      "tokens_seen": 98828288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899097291875627,
      "loss": 1.7411,
      "theoretical_loss": 4.779162724279324,
      "tokens_seen": 98959360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898896690070211,
      "loss": 1.7954,
      "theoretical_loss": 4.778284430255981,
      "tokens_seen": 99090432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898696088264794,
      "loss": 1.9314,
      "theoretical_loss": 4.777407622033838,
      "tokens_seen": 99221504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898495486459378,
      "loss": 1.8941,
      "theoretical_loss": 4.776532295141601,
      "tokens_seen": 99352576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898294884653962,
      "loss": 2.0467,
      "theoretical_loss": 4.775658445127318,
      "tokens_seen": 99483648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898094282848545,
      "loss": 1.875,
      "theoretical_loss": 4.7747860675582485,
      "tokens_seen": 99614720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489789368104313,
      "loss": 1.8207,
      "theoretical_loss": 4.773915158020776,
      "tokens_seen": 99745792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897693079237714,
      "loss": 1.8993,
      "theoretical_loss": 4.773045712120284,
      "tokens_seen": 99876864
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.3229472041130066,
      "objective/train/avg_token_score": 0.02979409135878086,
      "objective/train/docs_used": 67525,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.47669517993927,
      "objective/train/original_loss": 4.8227763175964355,
      "objective/train/theoretical_loss": 4.77261153666576,
      "objective/train/tokens_used": 99942400,
      "objective/train/unlikelihood_loss": 0.1576315313577652,
      "theoretical_loss": 4.77261153666576,
      "tokens_seen": 99942400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897492477432297,
      "loss": 1.9601,
      "theoretical_loss": 4.772177725481062,
      "tokens_seen": 100007936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897291875626881,
      "loss": 1.7757,
      "theoretical_loss": 4.771311193746191,
      "tokens_seen": 100139008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897091273821465,
      "loss": 1.8233,
      "theoretical_loss": 4.770446112577445,
      "tokens_seen": 100270080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896890672016048,
      "loss": 1.9513,
      "theoretical_loss": 4.769582477655177,
      "tokens_seen": 100401152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896690070210632,
      "loss": 1.8272,
      "theoretical_loss": 4.768720284678228,
      "tokens_seen": 100532224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896489468405215,
      "loss": 1.764,
      "theoretical_loss": 4.767859529363809,
      "tokens_seen": 100663296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896288866599799,
      "loss": 1.956,
      "theoretical_loss": 4.767000207447417,
      "tokens_seen": 100794368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896088264794384,
      "loss": 1.9563,
      "theoretical_loss": 4.766142314682716,
      "tokens_seen": 100925440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895887662988967,
      "loss": 1.8045,
      "theoretical_loss": 4.765285846841444,
      "tokens_seen": 101056512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895687061183551,
      "loss": 2.0327,
      "theoretical_loss": 4.764430799713314,
      "tokens_seen": 101187584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895486459378135,
      "loss": 1.6355,
      "theoretical_loss": 4.763577169105912,
      "tokens_seen": 101318656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895285857572718,
      "loss": 1.9126,
      "theoretical_loss": 4.762724950844598,
      "tokens_seen": 101449728
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.361131489276886,
      "objective/train/avg_token_score": 0.010854127816855907,
      "objective/train/docs_used": 68617,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5994964838027954,
      "objective/train/original_loss": 4.556946754455566,
      "objective/train/theoretical_loss": 4.761874140772408,
      "objective/train/tokens_used": 101580800,
      "objective/train/unlikelihood_loss": 0.12181789427995682,
      "theoretical_loss": 4.761874140772408,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895085255767302,
      "loss": 2.0572,
      "theoretical_loss": 4.761874140772408,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894884653961885,
      "loss": 1.6931,
      "theoretical_loss": 4.761024734749958,
      "tokens_seen": 101711872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894684052156469,
      "loss": 1.842,
      "theoretical_loss": 4.760176728655345,
      "tokens_seen": 101842944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894483450351053,
      "loss": 1.7938,
      "theoretical_loss": 4.759330118384053,
      "tokens_seen": 101974016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894282848545637,
      "loss": 1.8328,
      "theoretical_loss": 4.758484899848854,
      "tokens_seen": 102105088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894082246740221,
      "loss": 1.8125,
      "theoretical_loss": 4.7576410689797175,
      "tokens_seen": 102236160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893881644934805,
      "loss": 1.6074,
      "theoretical_loss": 4.756798621723712,
      "tokens_seen": 102367232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893681043129388,
      "loss": 1.7143,
      "theoretical_loss": 4.755957554044917,
      "tokens_seen": 102498304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893480441323972,
      "loss": 2.0152,
      "theoretical_loss": 4.755117861924321,
      "tokens_seen": 102629376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893279839518556,
      "loss": 1.8091,
      "theoretical_loss": 4.754279541359738,
      "tokens_seen": 102760448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893079237713139,
      "loss": 1.8848,
      "theoretical_loss": 4.75344258836571,
      "tokens_seen": 102891520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892878635907723,
      "loss": 1.9612,
      "theoretical_loss": 4.752606998973421,
      "tokens_seen": 103022592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892678034102307,
      "loss": 1.7873,
      "theoretical_loss": 4.7517727692306035,
      "tokens_seen": 103153664
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.3702346086502075,
      "objective/train/avg_token_score": 0.0168976578861475,
      "objective/train/docs_used": 69265,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7375133037567139,
      "objective/train/original_loss": 4.848740100860596,
      "objective/train/theoretical_loss": 4.751356162997164,
      "objective/train/tokens_used": 103219200,
      "objective/train/unlikelihood_loss": 0.1075160950422287,
      "theoretical_loss": 4.751356162997164,
      "tokens_seen": 103219200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892477432296891,
      "loss": 2.047,
      "theoretical_loss": 4.750939895201443,
      "tokens_seen": 103284736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892276830491475,
      "loss": 1.7528,
      "theoretical_loss": 4.750108372966501,
      "tokens_seen": 103415808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892076228686058,
      "loss": 1.9577,
      "theoretical_loss": 4.749278198622617,
      "tokens_seen": 103546880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891875626880642,
      "loss": 1.7456,
      "theoretical_loss": 4.748449368282822,
      "tokens_seen": 103677952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891675025075226,
      "loss": 1.6757,
      "theoretical_loss": 4.747621878076252,
      "tokens_seen": 103809024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891474423269809,
      "loss": 1.6139,
      "theoretical_loss": 4.746795724148061,
      "tokens_seen": 103940096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891273821464393,
      "loss": 2.0421,
      "theoretical_loss": 4.745970902659338,
      "tokens_seen": 104071168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891073219658976,
      "loss": 1.6816,
      "theoretical_loss": 4.7451474097870125,
      "tokens_seen": 104202240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890872617853561,
      "loss": 1.763,
      "theoretical_loss": 4.744325241723777,
      "tokens_seen": 104333312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890672016048145,
      "loss": 1.7075,
      "theoretical_loss": 4.743504394678,
      "tokens_seen": 104464384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890471414242728,
      "loss": 1.8281,
      "theoretical_loss": 4.742684864873641,
      "tokens_seen": 104595456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890270812437312,
      "loss": 1.7182,
      "theoretical_loss": 4.741866648550168,
      "tokens_seen": 104726528
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.35599949955940247,
      "objective/train/avg_token_score": 0.05716551095247269,
      "objective/train/docs_used": 70556,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.608228087425232,
      "objective/train/original_loss": 4.970082759857178,
      "objective/train/theoretical_loss": 4.741049741962473,
      "objective/train/tokens_used": 104857600,
      "objective/train/unlikelihood_loss": 0.09267803281545639,
      "theoretical_loss": 4.741049741962473,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890070210631896,
      "loss": 1.6529,
      "theoretical_loss": 4.741049741962473,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889869608826479,
      "loss": 1.8281,
      "theoretical_loss": 4.740234141380794,
      "tokens_seen": 104988672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889669007021063,
      "loss": 1.7417,
      "theoretical_loss": 4.739419843090626,
      "tokens_seen": 105119744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889468405215647,
      "loss": 1.7041,
      "theoretical_loss": 4.738606843392644,
      "tokens_seen": 105250816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488926780341023,
      "loss": 1.5751,
      "theoretical_loss": 4.737795138602624,
      "tokens_seen": 105381888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889067201604815,
      "loss": 1.8483,
      "theoretical_loss": 4.736984725051357,
      "tokens_seen": 105512960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888866599799398,
      "loss": 1.7904,
      "theoretical_loss": 4.736175599084576,
      "tokens_seen": 105644032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888665997993982,
      "loss": 1.9943,
      "theoretical_loss": 4.735367757062869,
      "tokens_seen": 105775104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888465396188566,
      "loss": 1.7043,
      "theoretical_loss": 4.734561195361609,
      "tokens_seen": 105906176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888264794383149,
      "loss": 1.9254,
      "theoretical_loss": 4.733755910370867,
      "tokens_seen": 106037248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888064192577733,
      "loss": 1.7947,
      "theoretical_loss": 4.732951898495341,
      "tokens_seen": 106168320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887863590772317,
      "loss": 1.7915,
      "theoretical_loss": 4.732149156154276,
      "tokens_seen": 106299392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048876629889669,
      "loss": 1.7682,
      "theoretical_loss": 4.731347679781386,
      "tokens_seen": 106430464
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.429068922996521,
      "objective/train/avg_token_score": 0.009868655353784561,
      "objective/train/docs_used": 70950,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9191758632659912,
      "objective/train/original_loss": 4.865683078765869,
      "objective/train/theoretical_loss": 4.730947415222481,
      "objective/train/tokens_used": 106496000,
      "objective/train/unlikelihood_loss": 0.10545031726360321,
      "theoretical_loss": 4.730947415222481,
      "tokens_seen": 106496000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887462387161484,
      "loss": 1.8063,
      "theoretical_loss": 4.730547465824781,
      "tokens_seen": 106561536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887261785356069,
      "loss": 2.0496,
      "theoretical_loss": 4.72974851074689,
      "tokens_seen": 106692608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887061183550652,
      "loss": 1.869,
      "theoretical_loss": 4.728950811024383,
      "tokens_seen": 106823680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886860581745236,
      "loss": 1.82,
      "theoretical_loss": 4.728154363148102,
      "tokens_seen": 106954752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886659979939819,
      "loss": 1.8344,
      "theoretical_loss": 4.72735916362298,
      "tokens_seen": 107085824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886459378134403,
      "loss": 1.6656,
      "theoretical_loss": 4.726565208967973,
      "tokens_seen": 107216896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886258776328988,
      "loss": 1.9221,
      "theoretical_loss": 4.725772495715983,
      "tokens_seen": 107347968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886058174523571,
      "loss": 1.8658,
      "theoretical_loss": 4.724981020413787,
      "tokens_seen": 107479040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885857572718155,
      "loss": 1.8208,
      "theoretical_loss": 4.7241907796219635,
      "tokens_seen": 107610112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885656970912739,
      "loss": 1.748,
      "theoretical_loss": 4.723401769914824,
      "tokens_seen": 107741184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885456369107322,
      "loss": 1.9757,
      "theoretical_loss": 4.722613987880335,
      "tokens_seen": 107872256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885255767301906,
      "loss": 1.7293,
      "theoretical_loss": 4.721827430120053,
      "tokens_seen": 108003328
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.4198435842990875,
      "objective/train/avg_token_score": 0.03244104981422424,
      "objective/train/docs_used": 72030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9712913036346436,
      "objective/train/original_loss": 5.18431282043457,
      "objective/train/theoretical_loss": 4.721042093249051,
      "objective/train/tokens_used": 108134400,
      "objective/train/unlikelihood_loss": 0.07222175598144531,
      "theoretical_loss": 4.721042093249051,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488505516549649,
      "loss": 2.0079,
      "theoretical_loss": 4.721042093249051,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884854563691073,
      "loss": 1.8136,
      "theoretical_loss": 4.7202579738958494,
      "tokens_seen": 108265472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884653961885657,
      "loss": 1.6593,
      "theoretical_loss": 4.719475068702346,
      "tokens_seen": 108396544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884453360080241,
      "loss": 1.8464,
      "theoretical_loss": 4.718693374323747,
      "tokens_seen": 108527616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884252758274825,
      "loss": 1.552,
      "theoretical_loss": 4.717912887428501,
      "tokens_seen": 108658688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884052156469409,
      "loss": 1.7877,
      "theoretical_loss": 4.717133604698222,
      "tokens_seen": 108789760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883851554663992,
      "loss": 1.6993,
      "theoretical_loss": 4.716355522827633,
      "tokens_seen": 108920832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883650952858576,
      "loss": 1.87,
      "theoretical_loss": 4.715578638524491,
      "tokens_seen": 109051904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488345035105316,
      "loss": 1.7024,
      "theoretical_loss": 4.714802948509522,
      "tokens_seen": 109182976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883249749247743,
      "loss": 1.7157,
      "theoretical_loss": 4.714028449516356,
      "tokens_seen": 109314048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883049147442327,
      "loss": 1.8335,
      "theoretical_loss": 4.713255138291454,
      "tokens_seen": 109445120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882848545636911,
      "loss": 1.7999,
      "theoretical_loss": 4.712483011594056,
      "tokens_seen": 109576192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048826479438314946,
      "loss": 1.6147,
      "theoretical_loss": 4.7117120661961005,
      "tokens_seen": 109707264
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.3784213066101074,
      "objective/train/avg_token_score": 0.07327920943498611,
      "objective/train/docs_used": 72698,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.72416353225708,
      "objective/train/original_loss": 4.901858806610107,
      "objective/train/theoretical_loss": 4.711327035479103,
      "objective/train/tokens_used": 109772800,
      "objective/train/unlikelihood_loss": 0.09020828455686569,
      "theoretical_loss": 4.711327035479103,
      "tokens_seen": 109772800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882447342026078,
      "loss": 1.836,
      "theoretical_loss": 4.710942298882169,
      "tokens_seen": 109838336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048822467402206624,
      "loss": 1.8356,
      "theoretical_loss": 4.710173706449419,
      "tokens_seen": 109969408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882046138415246,
      "loss": 1.6668,
      "theoretical_loss": 4.7094062857075185,
      "tokens_seen": 110100480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048818455366098296,
      "loss": 1.843,
      "theoretical_loss": 4.708640033478584,
      "tokens_seen": 110231552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881644934804413,
      "loss": 1.6348,
      "theoretical_loss": 4.7078749465971175,
      "tokens_seen": 110362624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048814443329989974,
      "loss": 1.885,
      "theoretical_loss": 4.707111021909941,
      "tokens_seen": 110493696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881243731193581,
      "loss": 1.4851,
      "theoretical_loss": 4.706348256276138,
      "tokens_seen": 110624768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048810431293881646,
      "loss": 1.9024,
      "theoretical_loss": 4.705586646566987,
      "tokens_seen": 110755840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880842527582748,
      "loss": 1.8789,
      "theoretical_loss": 4.704826189665905,
      "tokens_seen": 110886912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880641925777332,
      "loss": 1.9667,
      "theoretical_loss": 4.7040668824683785,
      "tokens_seen": 111017984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880441323971916,
      "loss": 1.7731,
      "theoretical_loss": 4.7033087218819105,
      "tokens_seen": 111149056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048802407221664997,
      "loss": 1.782,
      "theoretical_loss": 4.702551704825957,
      "tokens_seen": 111280128
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.38880741596221924,
      "objective/train/avg_token_score": 0.002843292895704508,
      "objective/train/docs_used": 74080,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.625144124031067,
      "objective/train/original_loss": 4.229738235473633,
      "objective/train/theoretical_loss": 4.701795828231866,
      "objective/train/tokens_used": 111411200,
      "objective/train/unlikelihood_loss": 0.14396952092647552,
      "theoretical_loss": 4.701795828231866,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048800401203610833,
      "loss": 1.9266,
      "theoretical_loss": 4.701795828231866,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879839518555667,
      "loss": 1.5747,
      "theoretical_loss": 4.701041089042813,
      "tokens_seen": 111542272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879638916750251,
      "loss": 1.7214,
      "theoretical_loss": 4.700287484213753,
      "tokens_seen": 111673344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048794383149448347,
      "loss": 1.9717,
      "theoretical_loss": 4.6995350107113545,
      "tokens_seen": 111804416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048792377131394183,
      "loss": 1.7515,
      "theoretical_loss": 4.698783665513934,
      "tokens_seen": 111935488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879037111334002,
      "loss": 1.8924,
      "theoretical_loss": 4.698033445611415,
      "tokens_seen": 112066560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048788365095285856,
      "loss": 1.9578,
      "theoretical_loss": 4.697284348005253,
      "tokens_seen": 112197632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048786359077231697,
      "loss": 1.5521,
      "theoretical_loss": 4.696536369708386,
      "tokens_seen": 112328704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048784353059177533,
      "loss": 1.7932,
      "theoretical_loss": 4.695789507745176,
      "tokens_seen": 112459776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878234704112337,
      "loss": 1.8788,
      "theoretical_loss": 4.695043759151353,
      "tokens_seen": 112590848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048780341023069206,
      "loss": 1.851,
      "theoretical_loss": 4.694299120973957,
      "tokens_seen": 112721920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877833500501505,
      "loss": 1.775,
      "theoretical_loss": 4.693555590271282,
      "tokens_seen": 112852992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048776328986960884,
      "loss": 1.7527,
      "theoretical_loss": 4.692813164112819,
      "tokens_seen": 112984064
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.32972875237464905,
      "objective/train/avg_token_score": 0.019512729719281197,
      "objective/train/docs_used": 74817,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5578219890594482,
      "objective/train/original_loss": 4.993590354919434,
      "objective/train/theoretical_loss": 4.692442364324931,
      "objective/train/tokens_used": 113049600,
      "objective/train/unlikelihood_loss": 0.07117574661970139,
      "theoretical_loss": 4.692442364324931,
      "tokens_seen": 113049600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877432296890672,
      "loss": 1.6282,
      "theoretical_loss": 4.692071839579201,
      "tokens_seen": 113115136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048772316950852556,
      "loss": 1.7212,
      "theoretical_loss": 4.691331613762153,
      "tokens_seen": 113246208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877031093279839,
      "loss": 1.9666,
      "theoretical_loss": 4.690592483764427,
      "tokens_seen": 113377280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048768304914744234,
      "loss": 1.8759,
      "theoretical_loss": 4.689854446699757,
      "tokens_seen": 113508352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876629889669007,
      "loss": 1.9312,
      "theoretical_loss": 4.689117499692798,
      "tokens_seen": 113639424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048764292878635907,
      "loss": 1.9425,
      "theoretical_loss": 4.688381639879076,
      "tokens_seen": 113770496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048762286860581743,
      "loss": 1.6478,
      "theoretical_loss": 4.687646864404934,
      "tokens_seen": 113901568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048760280842527584,
      "loss": 1.9203,
      "theoretical_loss": 4.686913170427477,
      "tokens_seen": 114032640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875827482447342,
      "loss": 1.5842,
      "theoretical_loss": 4.68618055511452,
      "tokens_seen": 114163712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048756268806419257,
      "loss": 1.7842,
      "theoretical_loss": 4.685449015644537,
      "tokens_seen": 114294784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048754262788365093,
      "loss": 1.6122,
      "theoretical_loss": 4.684718549206607,
      "tokens_seen": 114425856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048752256770310935,
      "loss": 1.6103,
      "theoretical_loss": 4.6839891530003595,
      "tokens_seen": 114556928
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.42418134212493896,
      "objective/train/avg_token_score": 0.013869252055883408,
      "objective/train/docs_used": 76169,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.927834153175354,
      "objective/train/original_loss": 4.830661296844482,
      "objective/train/theoretical_loss": 4.68326082423593,
      "objective/train/tokens_used": 114688000,
      "objective/train/unlikelihood_loss": 0.13057447969913483,
      "theoretical_loss": 4.68326082423593,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875025075225677,
      "loss": 1.8594,
      "theoretical_loss": 4.68326082423593,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048748244734202607,
      "loss": 1.8205,
      "theoretical_loss": 4.682533560133901,
      "tokens_seen": 114819072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048746238716148443,
      "loss": 1.8782,
      "theoretical_loss": 4.681807357925257,
      "tokens_seen": 114950144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874423269809428,
      "loss": 1.8902,
      "theoretical_loss": 4.68108221485133,
      "tokens_seen": 115081216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874222668004012,
      "loss": 1.7744,
      "theoretical_loss": 4.680358128163747,
      "tokens_seen": 115212288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874022066198596,
      "loss": 1.7876,
      "theoretical_loss": 4.6796350951243895,
      "tokens_seen": 115343360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048738214643931794,
      "loss": 1.6889,
      "theoretical_loss": 4.678913113005333,
      "tokens_seen": 115474432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873620862587763,
      "loss": 2.0774,
      "theoretical_loss": 4.678192179088802,
      "tokens_seen": 115605504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873420260782347,
      "loss": 1.8764,
      "theoretical_loss": 4.677472290667122,
      "tokens_seen": 115736576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873219658976931,
      "loss": 1.8666,
      "theoretical_loss": 4.676753445042669,
      "tokens_seen": 115867648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048730190571715144,
      "loss": 1.7627,
      "theoretical_loss": 4.6760356395278215,
      "tokens_seen": 115998720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872818455366098,
      "loss": 1.9188,
      "theoretical_loss": 4.675318871444908,
      "tokens_seen": 116129792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048726178535606816,
      "loss": 1.9219,
      "theoretical_loss": 4.674603138126168,
      "tokens_seen": 116260864
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.44238758087158203,
      "objective/train/avg_token_score": 0.016831381246447563,
      "objective/train/docs_used": 76931,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.068297863006592,
      "objective/train/original_loss": 4.9380340576171875,
      "objective/train/theoretical_loss": 4.674245658672382,
      "objective/train/tokens_used": 116326400,
      "objective/train/unlikelihood_loss": 0.09188218414783478,
      "theoretical_loss": 4.674245658672382,
      "tokens_seen": 116326400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872417251755266,
      "loss": 1.8908,
      "theoretical_loss": 4.673888436913694,
      "tokens_seen": 116391936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487221664994985,
      "loss": 1.5917,
      "theoretical_loss": 4.673174765159393,
      "tokens_seen": 116523008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048720160481444336,
      "loss": 1.7106,
      "theoretical_loss": 4.6724621202249335,
      "tokens_seen": 116654080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871815446339017,
      "loss": 1.7435,
      "theoretical_loss": 4.6717504994817,
      "tokens_seen": 116785152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048716148445336014,
      "loss": 2.0931,
      "theoretical_loss": 4.671039900310747,
      "tokens_seen": 116916224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871414242728185,
      "loss": 1.889,
      "theoretical_loss": 4.670330320102753,
      "tokens_seen": 117047296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048712136409227686,
      "loss": 1.8596,
      "theoretical_loss": 4.669621756257971,
      "tokens_seen": 117178368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871013039117352,
      "loss": 1.8924,
      "theoretical_loss": 4.668914206186189,
      "tokens_seen": 117309440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870812437311936,
      "loss": 1.848,
      "theoretical_loss": 4.66820766730668,
      "tokens_seen": 117440512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487061183550652,
      "loss": 1.6911,
      "theoretical_loss": 4.667502137048155,
      "tokens_seen": 117571584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048704112337011037,
      "loss": 1.7743,
      "theoretical_loss": 4.666797612848723,
      "tokens_seen": 117702656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048702106318956873,
      "loss": 1.6988,
      "theoretical_loss": 4.666094092155843,
      "tokens_seen": 117833728
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.5015273690223694,
      "objective/train/avg_token_score": 0.006345341447740793,
      "objective/train/docs_used": 78354,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.065946578979492,
      "objective/train/original_loss": 3.913723945617676,
      "objective/train/theoretical_loss": 4.665391572426282,
      "objective/train/tokens_used": 117964800,
      "objective/train/unlikelihood_loss": 0.22304238379001617,
      "theoretical_loss": 4.665391572426282,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870010030090271,
      "loss": 1.9779,
      "theoretical_loss": 4.665391572426282,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869809428284855,
      "loss": 1.8665,
      "theoretical_loss": 4.664690051126065,
      "tokens_seen": 118095872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048696088264794387,
      "loss": 1.5626,
      "theoretical_loss": 4.66398952573044,
      "tokens_seen": 118226944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048694082246740223,
      "loss": 1.8862,
      "theoretical_loss": 4.663289993723826,
      "tokens_seen": 118358016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869207622868606,
      "loss": 1.88,
      "theoretical_loss": 4.662591452599774,
      "tokens_seen": 118489088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048690070210631896,
      "loss": 1.8471,
      "theoretical_loss": 4.661893899860923,
      "tokens_seen": 118620160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048688064192577737,
      "loss": 1.5922,
      "theoretical_loss": 4.661197333018957,
      "tokens_seen": 118751232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048686058174523573,
      "loss": 1.8316,
      "theoretical_loss": 4.6605017495945615,
      "tokens_seen": 118882304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868405215646941,
      "loss": 1.7042,
      "theoretical_loss": 4.659807147117382,
      "tokens_seen": 119013376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048682046138415246,
      "loss": 1.6569,
      "theoretical_loss": 4.659113523125981,
      "tokens_seen": 119144448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868004012036109,
      "loss": 1.6006,
      "theoretical_loss": 4.658420875167799,
      "tokens_seen": 119275520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048678034102306924,
      "loss": 1.6682,
      "theoretical_loss": 4.657729200799105,
      "tokens_seen": 119406592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867602808425276,
      "loss": 1.7433,
      "theoretical_loss": 4.657038497584967,
      "tokens_seen": 119537664
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.359115332365036,
      "objective/train/avg_token_score": 0.027856703847646713,
      "objective/train/docs_used": 78905,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.605509638786316,
      "objective/train/original_loss": 4.950808048248291,
      "objective/train/theoretical_loss": 4.656693509402331,
      "objective/train/tokens_used": 119603200,
      "objective/train/unlikelihood_loss": 0.09441498667001724,
      "theoretical_loss": 4.656693509402331,
      "tokens_seen": 119603200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048674022066198596,
      "loss": 1.599,
      "theoretical_loss": 4.6563487630991975,
      "tokens_seen": 119668736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867201604814443,
      "loss": 1.8791,
      "theoretical_loss": 4.655659994924323,
      "tokens_seen": 119799808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048670010030090274,
      "loss": 1.6709,
      "theoretical_loss": 4.65497219065154,
      "tokens_seen": 119930880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866800401203611,
      "loss": 1.7202,
      "theoretical_loss": 4.654285347880672,
      "tokens_seen": 120061952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048665997993981947,
      "loss": 1.8856,
      "theoretical_loss": 4.653599464220129,
      "tokens_seen": 120193024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048663991975927783,
      "loss": 1.8096,
      "theoretical_loss": 4.6529145372868745,
      "tokens_seen": 120324096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048661985957873624,
      "loss": 1.6635,
      "theoretical_loss": 4.652230564706377,
      "tokens_seen": 120455168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865997993981946,
      "loss": 1.6887,
      "theoretical_loss": 4.651547544112575,
      "tokens_seen": 120586240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048657973921765297,
      "loss": 1.6575,
      "theoretical_loss": 4.650865473147837,
      "tokens_seen": 120717312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048655967903711133,
      "loss": 1.7481,
      "theoretical_loss": 4.650184349462922,
      "tokens_seen": 120848384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048653961885656975,
      "loss": 1.7805,
      "theoretical_loss": 4.6495041707169396,
      "tokens_seen": 120979456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865195586760281,
      "loss": 1.5922,
      "theoretical_loss": 4.648824934577313,
      "tokens_seen": 121110528
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.28109726309776306,
      "objective/train/avg_token_score": 0.07095440477132797,
      "objective/train/docs_used": 79692,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.170123815536499,
      "objective/train/original_loss": 4.673448085784912,
      "objective/train/theoretical_loss": 4.648146638719739,
      "objective/train/tokens_used": 121241600,
      "objective/train/unlikelihood_loss": 0.12229027599096298,
      "theoretical_loss": 4.648146638719739,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048649949849548647,
      "loss": 1.7261,
      "theoretical_loss": 4.648146638719739,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048647943831494483,
      "loss": 1.8013,
      "theoretical_loss": 4.647469280828153,
      "tokens_seen": 121372672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864593781344032,
      "loss": 1.9773,
      "theoretical_loss": 4.646792858594686,
      "tokens_seen": 121503744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864393179538616,
      "loss": 1.6408,
      "theoretical_loss": 4.646117369719629,
      "tokens_seen": 121634816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048641925777332,
      "loss": 1.7331,
      "theoretical_loss": 4.6454428119113995,
      "tokens_seen": 121765888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048639919759277834,
      "loss": 1.6656,
      "theoretical_loss": 4.644769182886495,
      "tokens_seen": 121896960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863791374122367,
      "loss": 1.8518,
      "theoretical_loss": 4.644096480369466,
      "tokens_seen": 122028032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863590772316951,
      "loss": 1.8672,
      "theoretical_loss": 4.6434247020928705,
      "tokens_seen": 122159104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863390170511535,
      "loss": 1.8686,
      "theoretical_loss": 4.642753845797243,
      "tokens_seen": 122290176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048631895687061184,
      "loss": 1.9569,
      "theoretical_loss": 4.642083909231053,
      "tokens_seen": 122421248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862988966900702,
      "loss": 1.659,
      "theoretical_loss": 4.641414890150675,
      "tokens_seen": 122552320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048627883650952857,
      "loss": 1.8805,
      "theoretical_loss": 4.6407467863203475,
      "tokens_seen": 122683392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486258776328987,
      "loss": 1.8608,
      "theoretical_loss": 4.6400795955121374,
      "tokens_seen": 122814464
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.46236559748649597,
      "objective/train/avg_token_score": 0.02207406610250473,
      "objective/train/docs_used": 81051,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.061617851257324,
      "objective/train/original_loss": 4.578189849853516,
      "objective/train/theoretical_loss": 4.639746341797229,
      "objective/train/tokens_used": 122880000,
      "objective/train/unlikelihood_loss": 0.10073511302471161,
      "theoretical_loss": 4.639746341797229,
      "tokens_seen": 122880000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048623871614844534,
      "loss": 1.9091,
      "theoretical_loss": 4.639413315505905,
      "tokens_seen": 122945536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862186559679037,
      "loss": 1.7783,
      "theoretical_loss": 4.638747944089273,
      "tokens_seen": 123076608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048619859578736207,
      "loss": 1.9328,
      "theoretical_loss": 4.638083479057579,
      "tokens_seen": 123207680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861785356068205,
      "loss": 1.8101,
      "theoretical_loss": 4.6374199182138565,
      "tokens_seen": 123338752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048615847542627885,
      "loss": 1.7789,
      "theoretical_loss": 4.636757259368787,
      "tokens_seen": 123469824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861384152457372,
      "loss": 1.9448,
      "theoretical_loss": 4.636095500340669,
      "tokens_seen": 123600896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048611835506519557,
      "loss": 1.7024,
      "theoretical_loss": 4.635434638955388,
      "tokens_seen": 123731968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048609829488465393,
      "loss": 1.8975,
      "theoretical_loss": 4.634774673046376,
      "tokens_seen": 123863040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048607823470411235,
      "loss": 1.6453,
      "theoretical_loss": 4.634115600454582,
      "tokens_seen": 123994112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860581745235707,
      "loss": 1.8373,
      "theoretical_loss": 4.6334574190284314,
      "tokens_seen": 124125184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860381143430291,
      "loss": 1.7563,
      "theoretical_loss": 4.632800126623803,
      "tokens_seen": 124256256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048601805416248744,
      "loss": 1.832,
      "theoretical_loss": 4.632143721103983,
      "tokens_seen": 124387328
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.3342497646808624,
      "objective/train/avg_token_score": 0.010767764411866665,
      "objective/train/docs_used": 81846,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.500130295753479,
      "objective/train/original_loss": 4.459724426269531,
      "objective/train/theoretical_loss": 4.631488200339643,
      "objective/train/tokens_used": 124518400,
      "objective/train/unlikelihood_loss": 0.17901530861854553,
      "theoretical_loss": 4.631488200339643,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048599799398194585,
      "loss": 1.8062,
      "theoretical_loss": 4.631488200339643,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859779338014042,
      "loss": 1.8382,
      "theoretical_loss": 4.630833562208797,
      "tokens_seen": 124649472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859578736208626,
      "loss": 1.7831,
      "theoretical_loss": 4.630179804596775,
      "tokens_seen": 124780544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048593781344032094,
      "loss": 1.7324,
      "theoretical_loss": 4.629526925396189,
      "tokens_seen": 124911616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859177532597793,
      "loss": 1.7767,
      "theoretical_loss": 4.628874922506897,
      "tokens_seen": 125042688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858976930792377,
      "loss": 1.801,
      "theoretical_loss": 4.628223793835975,
      "tokens_seen": 125173760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858776328986961,
      "loss": 1.8394,
      "theoretical_loss": 4.627573537297678,
      "tokens_seen": 125304832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048585757271815444,
      "loss": 1.8139,
      "theoretical_loss": 4.6269241508134185,
      "tokens_seen": 125435904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858375125376128,
      "loss": 1.8361,
      "theoretical_loss": 4.6262756323117245,
      "tokens_seen": 125566976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858174523570712,
      "loss": 2.1512,
      "theoretical_loss": 4.625627979728212,
      "tokens_seen": 125698048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857973921765296,
      "loss": 1.7211,
      "theoretical_loss": 4.624981191005554,
      "tokens_seen": 125829120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048577733199598795,
      "loss": 1.7506,
      "theoretical_loss": 4.624335264093447,
      "tokens_seen": 125960192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857572718154463,
      "loss": 1.719,
      "theoretical_loss": 4.623690196948582,
      "tokens_seen": 126091264
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.2474951148033142,
      "objective/train/avg_token_score": 0.018130095675587654,
      "objective/train/docs_used": 83055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1577845811843872,
      "objective/train/original_loss": 4.456354141235352,
      "objective/train/theoretical_loss": 4.623367985152234,
      "objective/train/tokens_used": 126156800,
      "objective/train/unlikelihood_loss": 0.11982196569442749,
      "theoretical_loss": 4.623367985152234,
      "tokens_seen": 126156800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048573721163490467,
      "loss": 1.7005,
      "theoretical_loss": 4.623045987534609,
      "tokens_seen": 126222336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048571715145436314,
      "loss": 2.0355,
      "theoretical_loss": 4.622402633822114,
      "tokens_seen": 126353408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856970912738215,
      "loss": 1.7999,
      "theoretical_loss": 4.62176013378858,
      "tokens_seen": 126484480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048567703109327987,
      "loss": 1.6439,
      "theoretical_loss": 4.621118485418362,
      "tokens_seen": 126615552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048565697091273823,
      "loss": 2.0235,
      "theoretical_loss": 4.620477686702651,
      "tokens_seen": 126746624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048563691073219664,
      "loss": 1.805,
      "theoretical_loss": 4.619837735639452,
      "tokens_seen": 126877696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485616850551655,
      "loss": 1.8306,
      "theoretical_loss": 4.619198630233547,
      "tokens_seen": 127008768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048559679037111337,
      "loss": 1.7948,
      "theoretical_loss": 4.618560368496466,
      "tokens_seen": 127139840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048557673019057173,
      "loss": 1.7666,
      "theoretical_loss": 4.617922948446459,
      "tokens_seen": 127270912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048555667001003015,
      "loss": 1.8654,
      "theoretical_loss": 4.617286368108466,
      "tokens_seen": 127401984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855366098294885,
      "loss": 1.7467,
      "theoretical_loss": 4.616650625514091,
      "tokens_seen": 127533056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048551654964894687,
      "loss": 1.7312,
      "theoretical_loss": 4.616015718701563,
      "tokens_seen": 127664128
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.39668866991996765,
      "objective/train/avg_token_score": 0.008423499763011932,
      "objective/train/docs_used": 83592,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7686882019042969,
      "objective/train/original_loss": 4.729675769805908,
      "objective/train/theoretical_loss": 4.615381645715717,
      "objective/train/tokens_used": 127795200,
      "objective/train/unlikelihood_loss": 0.08303800970315933,
      "theoretical_loss": 4.615381645715717,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048549648946840523,
      "loss": 1.8034,
      "theoretical_loss": 4.615381645715717,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854764292878636,
      "loss": 1.7946,
      "theoretical_loss": 4.61474840460796,
      "tokens_seen": 127926272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485456369107322,
      "loss": 1.6519,
      "theoretical_loss": 4.614115993436242,
      "tokens_seen": 128057344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854363089267804,
      "loss": 1.9487,
      "theoretical_loss": 4.613484410265032,
      "tokens_seen": 128188416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048541624874623874,
      "loss": 1.6912,
      "theoretical_loss": 4.612853653165283,
      "tokens_seen": 128319488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853961885656971,
      "loss": 1.7975,
      "theoretical_loss": 4.612223720214407,
      "tokens_seen": 128450560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853761283851555,
      "loss": 1.6292,
      "theoretical_loss": 4.61159460949625,
      "tokens_seen": 128581632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853560682046139,
      "loss": 1.5301,
      "theoretical_loss": 4.610966319101056,
      "tokens_seen": 128712704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048533600802407224,
      "loss": 1.8054,
      "theoretical_loss": 4.610338847125445,
      "tokens_seen": 128843776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853159478435306,
      "loss": 1.6353,
      "theoretical_loss": 4.6097121916723856,
      "tokens_seen": 128974848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048529588766298897,
      "loss": 1.8334,
      "theoretical_loss": 4.609086350851165,
      "tokens_seen": 129105920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852758274824474,
      "loss": 1.7218,
      "theoretical_loss": 4.6084613227773605,
      "tokens_seen": 129236992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048525576730190574,
      "loss": 1.8052,
      "theoretical_loss": 4.607837105572816,
      "tokens_seen": 129368064
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.5973240733146667,
      "objective/train/avg_token_score": 0.010894504375755787,
      "objective/train/docs_used": 84814,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.5378358364105225,
      "objective/train/original_loss": 4.310730934143066,
      "objective/train/theoretical_loss": 4.607525300461299,
      "objective/train/tokens_used": 129433600,
      "objective/train/unlikelihood_loss": 0.1317676305770874,
      "theoretical_loss": 4.607525300461299,
      "tokens_seen": 129433600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852357071213641,
      "loss": 1.9714,
      "theoretical_loss": 4.607213697365613,
      "tokens_seen": 129499136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048521564694082247,
      "loss": 1.7143,
      "theoretical_loss": 4.6065910962900425,
      "tokens_seen": 129630208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851955867602809,
      "loss": 2.0327,
      "theoretical_loss": 4.60596930048658,
      "tokens_seen": 129761280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048517552657973925,
      "loss": 1.8787,
      "theoretical_loss": 4.6053483081018545,
      "tokens_seen": 129892352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851554663991976,
      "loss": 1.9663,
      "theoretical_loss": 4.604728117288631,
      "tokens_seen": 130023424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048513540621865597,
      "loss": 1.7322,
      "theoretical_loss": 4.604108726205774,
      "tokens_seen": 130154496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048511534603811433,
      "loss": 1.5917,
      "theoretical_loss": 4.6034901330182265,
      "tokens_seen": 130285568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048509528585757275,
      "loss": 1.7381,
      "theoretical_loss": 4.6028723358969845,
      "tokens_seen": 130416640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850752256770311,
      "loss": 1.8999,
      "theoretical_loss": 4.602255333019068,
      "tokens_seen": 130547712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850551654964895,
      "loss": 1.7749,
      "theoretical_loss": 4.601639122567497,
      "tokens_seen": 130678784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048503510531594784,
      "loss": 2.0268,
      "theoretical_loss": 4.601023702731264,
      "tokens_seen": 130809856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048501504513540625,
      "loss": 1.8285,
      "theoretical_loss": 4.600409071705312,
      "tokens_seen": 130940928
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.4414711594581604,
      "objective/train/avg_token_score": 0.013736131601035595,
      "objective/train/docs_used": 85560,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.943097710609436,
      "objective/train/original_loss": 4.5531487464904785,
      "objective/train/theoretical_loss": 4.599795227690505,
      "objective/train/tokens_used": 131072000,
      "objective/train/unlikelihood_loss": 0.18566669523715973,
      "theoretical_loss": 4.599795227690505,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849949849548646,
      "loss": 1.7317,
      "theoretical_loss": 4.599795227690505,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484974924774323,
      "loss": 2.0678,
      "theoretical_loss": 4.599182168893604,
      "tokens_seen": 131203072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048495486459378134,
      "loss": 2.0053,
      "theoretical_loss": 4.5985698935272445,
      "tokens_seen": 131334144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849348044132397,
      "loss": 1.7232,
      "theoretical_loss": 4.597958399809908,
      "tokens_seen": 131465216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849147442326981,
      "loss": 1.7742,
      "theoretical_loss": 4.597347685965897,
      "tokens_seen": 131596288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848946840521565,
      "loss": 1.7775,
      "theoretical_loss": 4.596737750225311,
      "tokens_seen": 131727360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048487462387161484,
      "loss": 1.6808,
      "theoretical_loss": 4.596128590824026,
      "tokens_seen": 131858432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848545636910732,
      "loss": 1.6192,
      "theoretical_loss": 4.595520206003663,
      "tokens_seen": 131989504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848345035105316,
      "loss": 1.7225,
      "theoretical_loss": 4.594912594011566,
      "tokens_seen": 132120576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048481444332999,
      "loss": 1.9255,
      "theoretical_loss": 4.594305753100782,
      "tokens_seen": 132251648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048479438314944835,
      "loss": 1.7942,
      "theoretical_loss": 4.59369968153003,
      "tokens_seen": 132382720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847743229689067,
      "loss": 1.7619,
      "theoretical_loss": 4.593094377563681,
      "tokens_seen": 132513792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048475426278836507,
      "loss": 1.7052,
      "theoretical_loss": 4.592489839471735,
      "tokens_seen": 132644864
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.4741568863391876,
      "objective/train/avg_token_score": 0.008817518129944801,
      "objective/train/docs_used": 86181,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0522453784942627,
      "objective/train/original_loss": 4.405289649963379,
      "objective/train/theoretical_loss": 4.592187857089571,
      "objective/train/tokens_used": 132710400,
      "objective/train/unlikelihood_loss": 0.15506720542907715,
      "theoretical_loss": 4.592187857089571,
      "tokens_seen": 132710400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847342026078235,
      "loss": 1.6979,
      "theoretical_loss": 4.591886065529795,
      "tokens_seen": 132775936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048471414242728185,
      "loss": 1.6689,
      "theoretical_loss": 4.591283054019041,
      "tokens_seen": 132907008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846940822467402,
      "loss": 1.4962,
      "theoretical_loss": 4.590680803226213,
      "tokens_seen": 133038080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846740220661986,
      "loss": 1.6976,
      "theoretical_loss": 4.590079311443583,
      "tokens_seen": 133169152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484653961885657,
      "loss": 1.7468,
      "theoretical_loss": 4.589478576968932,
      "tokens_seen": 133300224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048463390170511535,
      "loss": 1.7804,
      "theoretical_loss": 4.588878598105527,
      "tokens_seen": 133431296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846138415245737,
      "loss": 1.6541,
      "theoretical_loss": 4.588279373162101,
      "tokens_seen": 133562368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845937813440321,
      "loss": 1.8706,
      "theoretical_loss": 4.587680900452824,
      "tokens_seen": 133693440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048457372116349044,
      "loss": 1.8239,
      "theoretical_loss": 4.587083178297288,
      "tokens_seen": 133824512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048455366098294886,
      "loss": 1.8302,
      "theoretical_loss": 4.586486205020474,
      "tokens_seen": 133955584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845336008024072,
      "loss": 1.772,
      "theoretical_loss": 4.585889978952741,
      "tokens_seen": 134086656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845135406218656,
      "loss": 1.6578,
      "theoretical_loss": 4.585294498429796,
      "tokens_seen": 134217728
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.49590665102005005,
      "objective/train/avg_token_score": 0.008106932044029236,
      "objective/train/docs_used": 87535,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1396756172180176,
      "objective/train/original_loss": 4.32695198059082,
      "objective/train/theoretical_loss": 4.584699761792674,
      "objective/train/tokens_used": 134348800,
      "objective/train/unlikelihood_loss": 0.16700398921966553,
      "theoretical_loss": 4.584699761792674,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048449348044132394,
      "loss": 1.8098,
      "theoretical_loss": 4.584699761792674,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048447342026078236,
      "loss": 1.9279,
      "theoretical_loss": 4.5841057673877135,
      "tokens_seen": 134479872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844533600802407,
      "loss": 1.8587,
      "theoretical_loss": 4.5835125135665375,
      "tokens_seen": 134610944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844332998996991,
      "loss": 1.6187,
      "theoretical_loss": 4.58291999868603,
      "tokens_seen": 134742016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048441323971915745,
      "loss": 1.58,
      "theoretical_loss": 4.582328221108318,
      "tokens_seen": 134873088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048439317953861586,
      "loss": 1.9187,
      "theoretical_loss": 4.581737179200739,
      "tokens_seen": 135004160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843731193580742,
      "loss": 1.8012,
      "theoretical_loss": 4.581146871335832,
      "tokens_seen": 135135232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843530591775326,
      "loss": 1.745,
      "theoretical_loss": 4.5805572958913086,
      "tokens_seen": 135266304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048433299899699095,
      "loss": 1.8915,
      "theoretical_loss": 4.579968451250032,
      "tokens_seen": 135397376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843129388164493,
      "loss": 1.745,
      "theoretical_loss": 4.579380335800001,
      "tokens_seen": 135528448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048429287863590773,
      "loss": 1.6448,
      "theoretical_loss": 4.5787929479343195,
      "tokens_seen": 135659520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842728184553661,
      "loss": 1.9537,
      "theoretical_loss": 4.578206286051184,
      "tokens_seen": 135790592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048425275827482445,
      "loss": 1.8237,
      "theoretical_loss": 4.577620348553859,
      "tokens_seen": 135921664
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.3742668628692627,
      "objective/train/avg_token_score": 0.01183388289064169,
      "objective/train/docs_used": 88189,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5616800785064697,
      "objective/train/original_loss": 4.133111476898193,
      "objective/train/theoretical_loss": 4.577327650952276,
      "objective/train/tokens_used": 135987200,
      "objective/train/unlikelihood_loss": 0.14664104580879211,
      "theoretical_loss": 4.577327650952276,
      "tokens_seen": 135987200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048423269809428287,
      "loss": 1.6142,
      "theoretical_loss": 4.5770351338506545,
      "tokens_seen": 136052736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842126379137413,
      "loss": 1.8337,
      "theoretical_loss": 4.57645064035491,
      "tokens_seen": 136183808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048419257773319965,
      "loss": 1.8521,
      "theoretical_loss": 4.575866866484967,
      "tokens_seen": 136314880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484172517552658,
      "loss": 1.6806,
      "theoretical_loss": 4.575283810664155,
      "tokens_seen": 136445952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048415245737211637,
      "loss": 1.6517,
      "theoretical_loss": 4.574701471320768,
      "tokens_seen": 136577024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048413239719157473,
      "loss": 1.5807,
      "theoretical_loss": 4.574119846888045,
      "tokens_seen": 136708096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048411233701103315,
      "loss": 1.6995,
      "theoretical_loss": 4.573538935804146,
      "tokens_seen": 136839168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840922768304915,
      "loss": 1.7618,
      "theoretical_loss": 4.5729587365121365,
      "tokens_seen": 136970240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840722166499499,
      "loss": 1.7269,
      "theoretical_loss": 4.572379247459969,
      "tokens_seen": 137101312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048405215646940824,
      "loss": 1.629,
      "theoretical_loss": 4.571800467100456,
      "tokens_seen": 137232384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048403209628886665,
      "loss": 1.7945,
      "theoretical_loss": 4.571222393891253,
      "tokens_seen": 137363456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484012036108325,
      "loss": 1.6366,
      "theoretical_loss": 4.570645026294844,
      "tokens_seen": 137494528
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.4571114480495453,
      "objective/train/avg_token_score": 0.013456963934004307,
      "objective/train/docs_used": 88907,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0327019691467285,
      "objective/train/original_loss": 4.131555557250977,
      "objective/train/theoretical_loss": 4.570068362778516,
      "objective/train/tokens_used": 137625600,
      "objective/train/unlikelihood_loss": 0.16605950891971588,
      "theoretical_loss": 4.570068362778516,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839919759277834,
      "loss": 1.9336,
      "theoretical_loss": 4.570068362778516,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048397191574724174,
      "loss": 1.7372,
      "theoretical_loss": 4.569492401814339,
      "tokens_seen": 137756672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839518555667001,
      "loss": 1.7945,
      "theoretical_loss": 4.568917141879149,
      "tokens_seen": 137887744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839317953861585,
      "loss": 1.8875,
      "theoretical_loss": 4.568342581454532,
      "tokens_seen": 138018816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839117352056169,
      "loss": 1.5855,
      "theoretical_loss": 4.567768719026797,
      "tokens_seen": 138149888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048389167502507524,
      "loss": 1.796,
      "theoretical_loss": 4.567195553086961,
      "tokens_seen": 138280960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838716148445336,
      "loss": 1.8751,
      "theoretical_loss": 4.566623082130729,
      "tokens_seen": 138412032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483851554663992,
      "loss": 1.7671,
      "theoretical_loss": 4.56605130465848,
      "tokens_seen": 138543104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838314944834504,
      "loss": 1.7724,
      "theoretical_loss": 4.565480219175237,
      "tokens_seen": 138674176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048381143430290875,
      "loss": 1.8898,
      "theoretical_loss": 4.56490982419066,
      "tokens_seen": 138805248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837913741223671,
      "loss": 1.766,
      "theoretical_loss": 4.564340118219022,
      "tokens_seen": 138936320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048377131394182547,
      "loss": 1.6873,
      "theoretical_loss": 4.563771099779187,
      "tokens_seen": 139067392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837512537612839,
      "loss": 1.7903,
      "theoretical_loss": 4.563202767394597,
      "tokens_seen": 139198464
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.4491080045700073,
      "objective/train/avg_token_score": 0.0068944301456213,
      "objective/train/docs_used": 90119,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.943935751914978,
      "objective/train/original_loss": 4.3753509521484375,
      "objective/train/theoretical_loss": 4.562918858012843,
      "objective/train/tokens_used": 139264000,
      "objective/train/unlikelihood_loss": 0.1535024493932724,
      "theoretical_loss": 4.562918858012843,
      "tokens_seen": 139264000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048373119358074225,
      "loss": 1.6177,
      "theoretical_loss": 4.562635119593255,
      "tokens_seen": 139329536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837111334002006,
      "loss": 1.5881,
      "theoretical_loss": 4.5620681549076965,
      "tokens_seen": 139460608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483691073219659,
      "loss": 1.8352,
      "theoretical_loss": 4.561501871874984,
      "tokens_seen": 139591680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836710130391174,
      "loss": 1.7988,
      "theoretical_loss": 4.560936269036679,
      "tokens_seen": 139722752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048365095285857575,
      "loss": 1.694,
      "theoretical_loss": 4.560371344938831,
      "tokens_seen": 139853824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836308926780341,
      "loss": 1.5882,
      "theoretical_loss": 4.559807098131953,
      "tokens_seen": 139984896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836108324974925,
      "loss": 1.6214,
      "theoretical_loss": 4.559243527171011,
      "tokens_seen": 140115968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048359077231695084,
      "loss": 1.8377,
      "theoretical_loss": 4.558680630615397,
      "tokens_seen": 140247040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048357071213640926,
      "loss": 1.8401,
      "theoretical_loss": 4.558118407028921,
      "tokens_seen": 140378112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835506519558676,
      "loss": 1.6105,
      "theoretical_loss": 4.557556854979786,
      "tokens_seen": 140509184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483530591775326,
      "loss": 1.7269,
      "theoretical_loss": 4.556995973040574,
      "tokens_seen": 140640256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048351053159478434,
      "loss": 1.6745,
      "theoretical_loss": 4.556435759788229,
      "tokens_seen": 140771328
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.3440249264240265,
      "objective/train/avg_token_score": 0.03104650229215622,
      "objective/train/docs_used": 90792,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5081690549850464,
      "objective/train/original_loss": 4.372777462005615,
      "objective/train/theoretical_loss": 4.555876213804037,
      "objective/train/tokens_used": 140902400,
      "objective/train/unlikelihood_loss": 0.17843031883239746,
      "theoretical_loss": 4.555876213804037,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048349047141424276,
      "loss": 1.682,
      "theoretical_loss": 4.555876213804037,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834704112337011,
      "loss": 1.8494,
      "theoretical_loss": 4.555317333673611,
      "tokens_seen": 141033472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834503510531595,
      "loss": 1.6859,
      "theoretical_loss": 4.5547591179868725,
      "tokens_seen": 141164544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048343029087261785,
      "loss": 1.7201,
      "theoretical_loss": 4.554201565338033,
      "tokens_seen": 141295616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048341023069207626,
      "loss": 1.6618,
      "theoretical_loss": 4.553644674325584,
      "tokens_seen": 141426688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833901705115346,
      "loss": 1.6429,
      "theoretical_loss": 4.553088443552269,
      "tokens_seen": 141557760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483370110330993,
      "loss": 1.6037,
      "theoretical_loss": 4.552532871625077,
      "tokens_seen": 141688832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048335005015045135,
      "loss": 1.564,
      "theoretical_loss": 4.551977957155217,
      "tokens_seen": 141819904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833299899699097,
      "loss": 1.7074,
      "theoretical_loss": 4.551423698758111,
      "tokens_seen": 141950976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048330992978936813,
      "loss": 1.6408,
      "theoretical_loss": 4.550870095053366,
      "tokens_seen": 142082048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832898696088265,
      "loss": 1.6711,
      "theoretical_loss": 4.550317144664766,
      "tokens_seen": 142213120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048326980942828485,
      "loss": 1.649,
      "theoretical_loss": 4.5497648462202545,
      "tokens_seen": 142344192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832497492477432,
      "loss": 1.714,
      "theoretical_loss": 4.549213198351914,
      "tokens_seen": 142475264
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.31085044145584106,
      "objective/train/avg_token_score": 0.008633834309875965,
      "objective/train/docs_used": 92175,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2824527025222778,
      "objective/train/original_loss": 4.095202922821045,
      "objective/train/theoretical_loss": 4.548937617957463,
      "objective/train/tokens_used": 142540800,
      "objective/train/unlikelihood_loss": 0.16536656022071838,
      "theoretical_loss": 4.548937617957463,
      "tokens_seen": 142540800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048322968906720163,
      "loss": 1.6286,
      "theoretical_loss": 4.548662199695954,
      "tokens_seen": 142606336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048320962888666,
      "loss": 1.7102,
      "theoretical_loss": 4.548111848892693,
      "tokens_seen": 142737408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048318956870611836,
      "loss": 1.6976,
      "theoretical_loss": 4.547562144586539,
      "tokens_seen": 142868480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831695085255767,
      "loss": 1.8329,
      "theoretical_loss": 4.5470130854259825,
      "tokens_seen": 142999552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831494483450351,
      "loss": 1.8789,
      "theoretical_loss": 4.546464670063569,
      "tokens_seen": 143130624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831293881644935,
      "loss": 1.9697,
      "theoretical_loss": 4.545916897155894,
      "tokens_seen": 143261696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048310932798395186,
      "loss": 1.7363,
      "theoretical_loss": 4.545369765363578,
      "tokens_seen": 143392768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830892678034102,
      "loss": 1.889,
      "theoretical_loss": 4.544823273351257,
      "tokens_seen": 143523840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830692076228686,
      "loss": 1.6188,
      "theoretical_loss": 4.544277419787566,
      "tokens_seen": 143654912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483049147442327,
      "loss": 1.8209,
      "theoretical_loss": 4.543732203345119,
      "tokens_seen": 143785984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048302908726178536,
      "loss": 1.6722,
      "theoretical_loss": 4.543187622700497,
      "tokens_seen": 143917056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830090270812437,
      "loss": 1.6843,
      "theoretical_loss": 4.542643676534234,
      "tokens_seen": 144048128
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.35123410820961,
      "objective/train/avg_token_score": 0.024587556719779968,
      "objective/train/docs_used": 92798,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.466112732887268,
      "objective/train/original_loss": 4.362775802612305,
      "objective/train/theoretical_loss": 4.542100363530799,
      "objective/train/tokens_used": 144179200,
      "objective/train/unlikelihood_loss": 0.18418121337890625,
      "theoretical_loss": 4.542100363530799,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829889669007021,
      "loss": 1.8888,
      "theoretical_loss": 4.542100363530799,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048296890672016045,
      "loss": 1.8105,
      "theoretical_loss": 4.5415576823785795,
      "tokens_seen": 144310272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048294884653961887,
      "loss": 1.7692,
      "theoretical_loss": 4.541015631769872,
      "tokens_seen": 144441344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048292878635907723,
      "loss": 1.7058,
      "theoretical_loss": 4.540474210400859,
      "tokens_seen": 144572416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829087261785356,
      "loss": 1.8536,
      "theoretical_loss": 4.5399334169716,
      "tokens_seen": 144703488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048288866599799395,
      "loss": 1.7873,
      "theoretical_loss": 4.539393250186015,
      "tokens_seen": 144834560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048286860581745237,
      "loss": 1.7762,
      "theoretical_loss": 4.538853708751866,
      "tokens_seen": 144965632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048284854563691073,
      "loss": 1.7058,
      "theoretical_loss": 4.538314791380748,
      "tokens_seen": 145096704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828284854563691,
      "loss": 1.9382,
      "theoretical_loss": 4.537776496788071,
      "tokens_seen": 145227776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048280842527582746,
      "loss": 1.8217,
      "theoretical_loss": 4.537238823693045,
      "tokens_seen": 145358848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827883650952858,
      "loss": 1.7635,
      "theoretical_loss": 4.536701770818665,
      "tokens_seen": 145489920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048276830491474423,
      "loss": 1.7094,
      "theoretical_loss": 4.536165336891699,
      "tokens_seen": 145620992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827482447342026,
      "loss": 1.8869,
      "theoretical_loss": 4.53562952064267,
      "tokens_seen": 145752064
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.4389662742614746,
      "objective/train/avg_token_score": 0.00564260920509696,
      "objective/train/docs_used": 94036,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9034613370895386,
      "objective/train/original_loss": 4.32117223739624,
      "objective/train/theoretical_loss": 4.535361843751696,
      "objective/train/tokens_used": 145817600,
      "objective/train/unlikelihood_loss": 0.2049456685781479,
      "theoretical_loss": 4.535361843751696,
      "tokens_seen": 145817600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482728184553661,
      "loss": 1.8672,
      "theoretical_loss": 4.535094320805847,
      "tokens_seen": 145883136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827081243731194,
      "loss": 1.8281,
      "theoretical_loss": 4.5345597361192205,
      "tokens_seen": 146014208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826880641925778,
      "loss": 1.8617,
      "theoretical_loss": 4.5340257653244995,
      "tokens_seen": 146145280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048266800401203615,
      "loss": 1.8899,
      "theoretical_loss": 4.533492407167093,
      "tokens_seen": 146276352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826479438314945,
      "loss": 1.6389,
      "theoretical_loss": 4.5329596603960916,
      "tokens_seen": 146407424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826278836509529,
      "loss": 1.831,
      "theoretical_loss": 4.532427523764261,
      "tokens_seen": 146538496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048260782347041124,
      "loss": 1.67,
      "theoretical_loss": 4.5318959960280205,
      "tokens_seen": 146669568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048258776328986966,
      "loss": 1.4795,
      "theoretical_loss": 4.531365075947434,
      "tokens_seen": 146800640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482567703109328,
      "loss": 1.9673,
      "theoretical_loss": 4.5308347622861955,
      "tokens_seen": 146931712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825476429287864,
      "loss": 1.6775,
      "theoretical_loss": 4.530305053811611,
      "tokens_seen": 147062784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048252758274824474,
      "loss": 1.8903,
      "theoretical_loss": 4.529775949294593,
      "tokens_seen": 147193856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048250752256770316,
      "loss": 1.5669,
      "theoretical_loss": 4.529247447509637,
      "tokens_seen": 147324928
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.5044599175453186,
      "objective/train/avg_token_score": 0.009645281359553337,
      "objective/train/docs_used": 94726,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1166269779205322,
      "objective/train/original_loss": 4.383207321166992,
      "objective/train/theoretical_loss": 4.528719547234816,
      "objective/train/tokens_used": 147456000,
      "objective/train/unlikelihood_loss": 0.12163429707288742,
      "theoretical_loss": 4.528719547234816,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824874623871615,
      "loss": 1.874,
      "theoretical_loss": 4.528719547234816,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824674022066199,
      "loss": 1.5731,
      "theoretical_loss": 4.528192247251763,
      "tokens_seen": 147587072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048244734202607825,
      "loss": 1.7182,
      "theoretical_loss": 4.527665546345656,
      "tokens_seen": 147718144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048242728184553666,
      "loss": 1.6724,
      "theoretical_loss": 4.527139443305209,
      "tokens_seen": 147849216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482407221664995,
      "loss": 1.8344,
      "theoretical_loss": 4.526613936922654,
      "tokens_seen": 147980288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823871614844534,
      "loss": 1.729,
      "theoretical_loss": 4.526089025993732,
      "tokens_seen": 148111360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048236710130391175,
      "loss": 1.914,
      "theoretical_loss": 4.525564709317678,
      "tokens_seen": 148242432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823470411233701,
      "loss": 1.8067,
      "theoretical_loss": 4.525040985697203,
      "tokens_seen": 148373504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048232698094282853,
      "loss": 1.7183,
      "theoretical_loss": 4.524517853938489,
      "tokens_seen": 148504576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004823069207622869,
      "loss": 1.6597,
      "theoretical_loss": 4.523995312851174,
      "tokens_seen": 148635648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048228686058174525,
      "loss": 1.8709,
      "theoretical_loss": 4.52347336124833,
      "tokens_seen": 148766720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822668004012036,
      "loss": 1.8413,
      "theoretical_loss": 4.522951997946466,
      "tokens_seen": 148897792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048224674022066203,
      "loss": 1.938,
      "theoretical_loss": 4.522431221765498,
      "tokens_seen": 149028864
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.5625610947608948,
      "objective/train/avg_token_score": 0.0176337081938982,
      "objective/train/docs_used": 95869,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.122063159942627,
      "objective/train/original_loss": 4.070887088775635,
      "objective/train/theoretical_loss": 4.522171053477507,
      "objective/train/tokens_used": 149094400,
      "objective/train/unlikelihood_loss": 0.15213339030742645,
      "theoretical_loss": 4.522171053477507,
      "tokens_seen": 149094400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822266800401204,
      "loss": 1.9062,
      "theoretical_loss": 4.5219110315287505,
      "tokens_seen": 149159936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048220661985957876,
      "loss": 1.7105,
      "theoretical_loss": 4.521391426062934,
      "tokens_seen": 149291008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821865596790371,
      "loss": 1.8112,
      "theoretical_loss": 4.520872404198139,
      "tokens_seen": 149422080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821664994984955,
      "loss": 1.6997,
      "theoretical_loss": 4.520353964767814,
      "tokens_seen": 149553152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821464393179539,
      "loss": 1.6329,
      "theoretical_loss": 4.519836106608768,
      "tokens_seen": 149684224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048212637913741226,
      "loss": 1.7948,
      "theoretical_loss": 4.519318828561142,
      "tokens_seen": 149815296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821063189568706,
      "loss": 1.9672,
      "theoretical_loss": 4.518802129468405,
      "tokens_seen": 149946368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482086258776329,
      "loss": 1.7178,
      "theoretical_loss": 4.518286008177341,
      "tokens_seen": 150077440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820661985957874,
      "loss": 1.8528,
      "theoretical_loss": 4.517770463538038,
      "tokens_seen": 150208512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048204613841524576,
      "loss": 1.7364,
      "theoretical_loss": 4.517255494403868,
      "tokens_seen": 150339584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820260782347041,
      "loss": 1.7236,
      "theoretical_loss": 4.516741099631485,
      "tokens_seen": 150470656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820060180541625,
      "loss": 1.6208,
      "theoretical_loss": 4.5162272780808035,
      "tokens_seen": 150601728
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.4750122129917145,
      "objective/train/avg_token_score": 0.008325014263391495,
      "objective/train/docs_used": 96442,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0012364387512207,
      "objective/train/original_loss": 4.350079536437988,
      "objective/train/theoretical_loss": 4.515714028614996,
      "objective/train/tokens_used": 150732800,
      "objective/train/unlikelihood_loss": 0.13368818163871765,
      "theoretical_loss": 4.515714028614996,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048198595787362085,
      "loss": 1.7642,
      "theoretical_loss": 4.515714028614996,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048196589769307927,
      "loss": 1.986,
      "theoretical_loss": 4.515201350100471,
      "tokens_seen": 150863872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048194583751253763,
      "loss": 1.6876,
      "theoretical_loss": 4.5146892414068684,
      "tokens_seen": 150994944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481925777331996,
      "loss": 1.3922,
      "theoretical_loss": 4.514177701407042,
      "tokens_seen": 151126016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048190571715145435,
      "loss": 1.7374,
      "theoretical_loss": 4.513666728977054,
      "tokens_seen": 151257088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048188565697091277,
      "loss": 1.7462,
      "theoretical_loss": 4.513156322996155,
      "tokens_seen": 151388160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048186559679037113,
      "loss": 1.5728,
      "theoretical_loss": 4.5126464823467805,
      "tokens_seen": 151519232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818455366098295,
      "loss": 1.7909,
      "theoretical_loss": 4.5121372059145335,
      "tokens_seen": 151650304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048182547642928786,
      "loss": 1.5978,
      "theoretical_loss": 4.511628492588174,
      "tokens_seen": 151781376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818054162487462,
      "loss": 1.6053,
      "theoretical_loss": 4.511120341259608,
      "tokens_seen": 151912448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048178535606820463,
      "loss": 1.8598,
      "theoretical_loss": 4.510612750823878,
      "tokens_seen": 152043520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481765295887663,
      "loss": 1.864,
      "theoretical_loss": 4.510105720179144,
      "tokens_seen": 152174592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048174523570712136,
      "loss": 1.7646,
      "theoretical_loss": 4.509599248226683,
      "tokens_seen": 152305664
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.5002443790435791,
      "objective/train/avg_token_score": 0.033866036683321,
      "objective/train/docs_used": 97778,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1443092823028564,
      "objective/train/original_loss": 4.667900562286377,
      "objective/train/theoretical_loss": 4.509346221417552,
      "objective/train/tokens_used": 152371200,
      "objective/train/unlikelihood_loss": 0.1134328842163086,
      "theoretical_loss": 4.509346221417552,
      "tokens_seen": 152371200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817251755265797,
      "loss": 1.7244,
      "theoretical_loss": 4.509093333870869,
      "tokens_seen": 152436736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048170511534603814,
      "loss": 1.5927,
      "theoretical_loss": 4.508587976019164,
      "tokens_seen": 152567808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816850551654965,
      "loss": 1.6639,
      "theoretical_loss": 4.508083173582105,
      "tokens_seen": 152698880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048166499498495486,
      "loss": 1.8064,
      "theoretical_loss": 4.5075789254733,
      "tokens_seen": 152829952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816449348044132,
      "loss": 1.6705,
      "theoretical_loss": 4.507075230609407,
      "tokens_seen": 152961024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816248746238716,
      "loss": 1.6887,
      "theoretical_loss": 4.506572087910127,
      "tokens_seen": 153092096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048160481444333,
      "loss": 1.7163,
      "theoretical_loss": 4.506069496298198,
      "tokens_seen": 153223168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048158475426278837,
      "loss": 1.5463,
      "theoretical_loss": 4.505567454699373,
      "tokens_seen": 153354240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048156469408224673,
      "loss": 1.7644,
      "theoretical_loss": 4.505065962042418,
      "tokens_seen": 153485312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815446339017051,
      "loss": 1.5152,
      "theoretical_loss": 4.504565017259097,
      "tokens_seen": 153616384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815245737211635,
      "loss": 1.6447,
      "theoretical_loss": 4.504064619284163,
      "tokens_seen": 153747456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048150451354062187,
      "loss": 1.4444,
      "theoretical_loss": 4.5035647670553445,
      "tokens_seen": 153878528
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.584799587726593,
      "objective/train/avg_token_score": 0.008578718639910221,
      "objective/train/docs_used": 98099,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.4139657020568848,
      "objective/train/original_loss": 4.249912261962891,
      "objective/train/theoretical_loss": 4.503065459513339,
      "objective/train/tokens_used": 154009600,
      "objective/train/unlikelihood_loss": 0.09866306930780411,
      "theoretical_loss": 4.503065459513339,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048148445336008023,
      "loss": 1.8128,
      "theoretical_loss": 4.503065459513339,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814643931795386,
      "loss": 1.8889,
      "theoretical_loss": 4.502566695601795,
      "tokens_seen": 154140672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048144433299899696,
      "loss": 1.8156,
      "theoretical_loss": 4.502068474267309,
      "tokens_seen": 154271744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048142427281845537,
      "loss": 1.8757,
      "theoretical_loss": 4.501570794459411,
      "tokens_seen": 154402816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048140421263791373,
      "loss": 1.8008,
      "theoretical_loss": 4.501073655130554,
      "tokens_seen": 154533888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813841524573721,
      "loss": 1.781,
      "theoretical_loss": 4.500577055236104,
      "tokens_seen": 154664960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048136409227683046,
      "loss": 1.7119,
      "theoretical_loss": 4.500080993734329,
      "tokens_seen": 154796032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813440320962889,
      "loss": 1.8122,
      "theoretical_loss": 4.499585469586387,
      "tokens_seen": 154927104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048132397191574724,
      "loss": 1.9573,
      "theoretical_loss": 4.499090481756321,
      "tokens_seen": 155058176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813039117352056,
      "loss": 1.6949,
      "theoretical_loss": 4.498596029211041,
      "tokens_seen": 155189248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048128385155466396,
      "loss": 1.8511,
      "theoretical_loss": 4.4981021109203185,
      "tokens_seen": 155320320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812637913741223,
      "loss": 1.716,
      "theoretical_loss": 4.497608725856776,
      "tokens_seen": 155451392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048124373119358074,
      "loss": 1.7856,
      "theoretical_loss": 4.497115872995876,
      "tokens_seen": 155582464
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.4549120366573334,
      "objective/train/avg_token_score": 0.0024577039293944836,
      "objective/train/docs_used": 99496,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9255015850067139,
      "objective/train/original_loss": 4.110994338989258,
      "objective/train/theoretical_loss": 4.496869645822022,
      "objective/train/tokens_used": 155648000,
      "objective/train/unlikelihood_loss": 0.1798098087310791,
      "theoretical_loss": 4.496869645822022,
      "tokens_seen": 155648000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048122367101303916,
      "loss": 1.6885,
      "theoretical_loss": 4.496623551315908,
      "tokens_seen": 155713536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812036108324975,
      "loss": 1.7518,
      "theoretical_loss": 4.496131759797984,
      "tokens_seen": 155844608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811835506519559,
      "loss": 1.696,
      "theoretical_loss": 4.495640497426023,
      "tokens_seen": 155975680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811634904714143,
      "loss": 1.7442,
      "theoretical_loss": 4.4951497631867445,
      "tokens_seen": 156106752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048114343029087266,
      "loss": 1.6777,
      "theoretical_loss": 4.4946595560696565,
      "tokens_seen": 156237824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481123370110331,
      "loss": 1.6765,
      "theoretical_loss": 4.494169875067046,
      "tokens_seen": 156368896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811033099297894,
      "loss": 1.6892,
      "theoretical_loss": 4.493680719173968,
      "tokens_seen": 156499968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810832497492478,
      "loss": 1.5711,
      "theoretical_loss": 4.493192087388239,
      "tokens_seen": 156631040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048106318956870616,
      "loss": 1.6435,
      "theoretical_loss": 4.4927039787104235,
      "tokens_seen": 156762112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810431293881645,
      "loss": 1.4946,
      "theoretical_loss": 4.492216392143826,
      "tokens_seen": 156893184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810230692076229,
      "loss": 1.9506,
      "theoretical_loss": 4.49172932669448,
      "tokens_seen": 157024256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048100300902708125,
      "loss": 1.5955,
      "theoretical_loss": 4.491242781371138,
      "tokens_seen": 157155328
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.3397482931613922,
      "objective/train/avg_token_score": 0.02855628915131092,
      "objective/train/docs_used": 100000,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4549492597579956,
      "objective/train/original_loss": 4.49015998840332,
      "objective/train/theoretical_loss": 4.4907567551852665,
      "objective/train/tokens_used": 157286400,
      "objective/train/unlikelihood_loss": 0.11378736048936844,
      "theoretical_loss": 4.4907567551852665,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048098294884653967,
      "loss": 1.5109,
      "theoretical_loss": 4.4907567551852665,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048096288866599803,
      "loss": 1.7852,
      "theoretical_loss": 4.490271247151027,
      "tokens_seen": 157417472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809428284854564,
      "loss": 1.6,
      "theoretical_loss": 4.489786256285276,
      "tokens_seen": 157548544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048092276830491475,
      "loss": 1.8271,
      "theoretical_loss": 4.489301781607551,
      "tokens_seen": 157679616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048090270812437317,
      "loss": 1.8531,
      "theoretical_loss": 4.48881782214006,
      "tokens_seen": 157810688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048088264794383153,
      "loss": 1.7978,
      "theoretical_loss": 4.488334376907673,
      "tokens_seen": 157941760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808625877632899,
      "loss": 1.5617,
      "theoretical_loss": 4.487851444937916,
      "tokens_seen": 158072832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048084252758274826,
      "loss": 1.7742,
      "theoretical_loss": 4.487369025260954,
      "tokens_seen": 158203904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808224674022066,
      "loss": 1.7228,
      "theoretical_loss": 4.48688711690959,
      "tokens_seen": 158334976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048080240722166503,
      "loss": 1.819,
      "theoretical_loss": 4.48640571891925,
      "tokens_seen": 158466048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807823470411234,
      "loss": 1.7198,
      "theoretical_loss": 4.485924830327974,
      "tokens_seen": 158597120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048076228686058176,
      "loss": 1.7289,
      "theoretical_loss": 4.485444450176413,
      "tokens_seen": 158728192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807422266800401,
      "loss": 1.694,
      "theoretical_loss": 4.484964577507808,
      "tokens_seen": 158859264
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.47953322529792786,
      "objective/train/avg_token_score": 0.00853870902210474,
      "objective/train/docs_used": 101476,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9475176334381104,
      "objective/train/original_loss": 4.236670970916748,
      "objective/train/theoretical_loss": 4.484724831181337,
      "objective/train/tokens_used": 158924800,
      "objective/train/unlikelihood_loss": 0.1266087293624878,
      "theoretical_loss": 4.484724831181337,
      "tokens_seen": 158924800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048072216649949854,
      "loss": 1.8989,
      "theoretical_loss": 4.4844852113679945,
      "tokens_seen": 158990336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807021063189569,
      "loss": 1.5987,
      "theoretical_loss": 4.484006350805385,
      "tokens_seen": 159121408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048068204613841526,
      "loss": 1.7375,
      "theoretical_loss": 4.483527994870958,
      "tokens_seen": 159252480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806619859578736,
      "loss": 1.5585,
      "theoretical_loss": 4.483050142618255,
      "tokens_seen": 159383552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480641925777332,
      "loss": 1.7186,
      "theoretical_loss": 4.482572793103373,
      "tokens_seen": 159514624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806218655967904,
      "loss": 1.6185,
      "theoretical_loss": 4.482095945384946,
      "tokens_seen": 159645696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048060180541624877,
      "loss": 1.6661,
      "theoretical_loss": 4.4816195985241425,
      "tokens_seen": 159776768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048058174523570713,
      "loss": 1.7461,
      "theoretical_loss": 4.481143751584659,
      "tokens_seen": 159907840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805616850551655,
      "loss": 1.8502,
      "theoretical_loss": 4.480668403632706,
      "tokens_seen": 160038912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805416248746239,
      "loss": 1.7936,
      "theoretical_loss": 4.480193553736999,
      "tokens_seen": 160169984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048052156469408227,
      "loss": 1.9024,
      "theoretical_loss": 4.479719200968757,
      "tokens_seen": 160301056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048050150451354063,
      "loss": 1.7003,
      "theoretical_loss": 4.479245344401685,
      "tokens_seen": 160432128
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.53665691614151,
      "objective/train/avg_token_score": 0.011694345623254776,
      "objective/train/docs_used": 102054,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.992788553237915,
      "objective/train/original_loss": 3.9252665042877197,
      "objective/train/theoretical_loss": 4.478771983111967,
      "objective/train/tokens_used": 160563200,
      "objective/train/unlikelihood_loss": 0.1462445855140686,
      "theoretical_loss": 4.478771983111967,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480481444332999,
      "loss": 1.6986,
      "theoretical_loss": 4.478771983111967,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048046138415245736,
      "loss": 1.4835,
      "theoretical_loss": 4.478299116178265,
      "tokens_seen": 160694272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048044132397191577,
      "loss": 1.5806,
      "theoretical_loss": 4.4778267426817,
      "tokens_seen": 160825344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048042126379137413,
      "loss": 1.7454,
      "theoretical_loss": 4.47735486170585,
      "tokens_seen": 160956416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804012036108325,
      "loss": 1.6559,
      "theoretical_loss": 4.4768834723367394,
      "tokens_seen": 161087488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048038114343029086,
      "loss": 1.7753,
      "theoretical_loss": 4.476412573662829,
      "tokens_seen": 161218560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803610832497493,
      "loss": 1.6249,
      "theoretical_loss": 4.475942164775013,
      "tokens_seen": 161349632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048034102306920764,
      "loss": 1.6971,
      "theoretical_loss": 4.475472244766601,
      "tokens_seen": 161480704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480320962888666,
      "loss": 1.6015,
      "theoretical_loss": 4.47500281273332,
      "tokens_seen": 161611776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048030090270812436,
      "loss": 1.8606,
      "theoretical_loss": 4.474533867773299,
      "tokens_seen": 161742848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802808425275827,
      "loss": 1.615,
      "theoretical_loss": 4.474065408987063,
      "tokens_seen": 161873920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048026078234704114,
      "loss": 1.5313,
      "theoretical_loss": 4.473597435477526,
      "tokens_seen": 162004992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802407221664995,
      "loss": 1.8292,
      "theoretical_loss": 4.473129946349982,
      "tokens_seen": 162136064
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.436950147151947,
      "objective/train/avg_token_score": 0.005409137345850468,
      "objective/train/docs_used": 102718,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9111195802688599,
      "objective/train/original_loss": 4.506403923034668,
      "objective/train/theoretical_loss": 4.472896383150508,
      "objective/train/tokens_used": 162201600,
      "objective/train/unlikelihood_loss": 0.16208896040916443,
      "theoretical_loss": 4.472896383150508,
      "tokens_seen": 162201600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048022066198595786,
      "loss": 1.6402,
      "theoretical_loss": 4.472662940712091,
      "tokens_seen": 162267136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048020060180541623,
      "loss": 1.5722,
      "theoretical_loss": 4.472196417673883,
      "tokens_seen": 162398208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048018054162487464,
      "loss": 1.6823,
      "theoretical_loss": 4.471730376347738,
      "tokens_seen": 162529280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480160481444333,
      "loss": 1.6869,
      "theoretical_loss": 4.471264815848384,
      "tokens_seen": 162660352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048014042126379137,
      "loss": 1.794,
      "theoretical_loss": 4.470799735292889,
      "tokens_seen": 162791424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048012036108324973,
      "loss": 1.7949,
      "theoretical_loss": 4.470335133800649,
      "tokens_seen": 162922496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048010030090270815,
      "loss": 1.6686,
      "theoretical_loss": 4.469871010493383,
      "tokens_seen": 163053568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800802407221665,
      "loss": 1.7866,
      "theoretical_loss": 4.4694073644951295,
      "tokens_seen": 163184640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048006018054162487,
      "loss": 1.6188,
      "theoretical_loss": 4.468944194932225,
      "tokens_seen": 163315712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048004012036108323,
      "loss": 1.7558,
      "theoretical_loss": 4.46848150093331,
      "tokens_seen": 163446784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800200601805416,
      "loss": 1.9446,
      "theoretical_loss": 4.468019281629316,
      "tokens_seen": 163577856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048,
      "loss": 1.9535,
      "theoretical_loss": 4.467557536153457,
      "tokens_seen": 163708928
    },
    {
      "debugging/length": 491.6,
      "debugging/num_segments": 5,
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.416177898645401,
      "objective/train/avg_token_score": 0.005864487960934639,
      "objective/train/docs_used": 104069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5953806638717651,
      "objective/train/original_loss": 3.946671962738037,
      "objective/train/theoretical_loss": 4.467096263641219,
      "objective/train/tokens_used": 163840000,
      "objective/train/unlikelihood_loss": 0.20240885019302368,
      "theoretical_loss": 4.467096263641219,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799799398194584,
      "loss": 1.8006,
      "theoretical_loss": 4.467096263641219,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047995987963891674,
      "loss": 1.7427,
      "theoretical_loss": 4.466635463230359,
      "tokens_seen": 163971072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799398194583751,
      "loss": 1.7973,
      "theoretical_loss": 4.466175134060894,
      "tokens_seen": 164102144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799197592778335,
      "loss": 1.7691,
      "theoretical_loss": 4.4657152752750875,
      "tokens_seen": 164233216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798996990972919,
      "loss": 1.8554,
      "theoretical_loss": 4.465255886017452,
      "tokens_seen": 164364288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047987963891675024,
      "loss": 1.9336,
      "theoretical_loss": 4.464796965434738,
      "tokens_seen": 164495360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798595787362086,
      "loss": 1.5818,
      "theoretical_loss": 4.464338512675919,
      "tokens_seen": 164626432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047983951855566696,
      "loss": 1.8526,
      "theoretical_loss": 4.4638805268921935,
      "tokens_seen": 164757504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798194583751254,
      "loss": 1.8442,
      "theoretical_loss": 4.463423007236974,
      "tokens_seen": 164888576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047979939819458374,
      "loss": 1.418,
      "theoretical_loss": 4.462965952865879,
      "tokens_seen": 165019648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797793380140421,
      "loss": 1.7218,
      "theoretical_loss": 4.462509362936723,
      "tokens_seen": 165150720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047975927783350047,
      "loss": 1.7421,
      "theoretical_loss": 4.462053236609516,
      "tokens_seen": 165281792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797392176529589,
      "loss": 1.7061,
      "theoretical_loss": 4.461597573046449,
      "tokens_seen": 165412864
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.4883309006690979,
      "objective/train/avg_token_score": 0.004766791593283415,
      "objective/train/docs_used": 105300,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9741791486740112,
      "objective/train/original_loss": 3.7647905349731445,
      "objective/train/theoretical_loss": 4.461369914540247,
      "objective/train/tokens_used": 165478400,
      "objective/train/unlikelihood_loss": 0.15376895666122437,
      "theoretical_loss": 4.461369914540247,
      "tokens_seen": 165478400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797191574724173,
      "loss": 1.7765,
      "theoretical_loss": 4.4611423714118885,
      "tokens_seen": 165543936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047969909729187566,
      "loss": 2.1214,
      "theoretical_loss": 4.460687630872371,
      "tokens_seen": 165675008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479679037111334,
      "loss": 1.8148,
      "theoretical_loss": 4.460233350596599,
      "tokens_seen": 165806080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796589769307924,
      "loss": 1.8356,
      "theoretical_loss": 4.459779529755423,
      "tokens_seen": 165937152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796389167502508,
      "loss": 1.847,
      "theoretical_loss": 4.459326167521844,
      "tokens_seen": 166068224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047961885656970917,
      "loss": 1.5851,
      "theoretical_loss": 4.458873263071002,
      "tokens_seen": 166199296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047959879638916753,
      "loss": 1.87,
      "theoretical_loss": 4.458420815580169,
      "tokens_seen": 166330368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795787362086259,
      "loss": 1.7718,
      "theoretical_loss": 4.457968824228743,
      "tokens_seen": 166461440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795586760280843,
      "loss": 1.9016,
      "theoretical_loss": 4.4575172881982414,
      "tokens_seen": 166592512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047953861584754267,
      "loss": 1.5086,
      "theoretical_loss": 4.457066206672291,
      "tokens_seen": 166723584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047951855566700103,
      "loss": 1.8205,
      "theoretical_loss": 4.456615578836625,
      "tokens_seen": 166854656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794984954864594,
      "loss": 1.617,
      "theoretical_loss": 4.45616540387907,
      "tokens_seen": 166985728
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.5096529722213745,
      "objective/train/avg_token_score": 0.009488563984632492,
      "objective/train/docs_used": 106148,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0118892192840576,
      "objective/train/original_loss": 4.2763590812683105,
      "objective/train/theoretical_loss": 4.455715680989545,
      "objective/train/tokens_used": 167116800,
      "objective/train/unlikelihood_loss": 0.12322863936424255,
      "theoretical_loss": 4.455715680989545,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047947843530591776,
      "loss": 1.6269,
      "theoretical_loss": 4.455715680989545,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047945837512537617,
      "loss": 1.809,
      "theoretical_loss": 4.4552664093600525,
      "tokens_seen": 167247872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047943831494483453,
      "loss": 1.7982,
      "theoretical_loss": 4.454817588184669,
      "tokens_seen": 167378944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794182547642929,
      "loss": 1.6824,
      "theoretical_loss": 4.454369216659542,
      "tokens_seen": 167510016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047939819458375126,
      "loss": 1.82,
      "theoretical_loss": 4.453921293982877,
      "tokens_seen": 167641088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793781344032097,
      "loss": 1.4183,
      "theoretical_loss": 4.453473819354942,
      "tokens_seen": 167772160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047935807422266804,
      "loss": 1.6876,
      "theoretical_loss": 4.453026791978045,
      "tokens_seen": 167903232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793380140421264,
      "loss": 1.4265,
      "theoretical_loss": 4.452580211056542,
      "tokens_seen": 168034304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047931795386158476,
      "loss": 1.7324,
      "theoretical_loss": 4.45213407579682,
      "tokens_seen": 168165376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792978936810431,
      "loss": 1.7566,
      "theoretical_loss": 4.451688385407296,
      "tokens_seen": 168296448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047927783350050154,
      "loss": 1.6371,
      "theoretical_loss": 4.4512431390984055,
      "tokens_seen": 168427520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792577733199599,
      "loss": 1.6714,
      "theoretical_loss": 4.450798336082601,
      "tokens_seen": 168558592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047923771313941827,
      "loss": 1.6945,
      "theoretical_loss": 4.450353975574341,
      "tokens_seen": 168689664
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.3742057681083679,
      "objective/train/avg_token_score": 0.02583669312298298,
      "objective/train/docs_used": 106755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.461486577987671,
      "objective/train/original_loss": 4.418996334075928,
      "objective/train/theoretical_loss": 4.450131961015606,
      "objective/train/tokens_used": 168755200,
      "objective/train/unlikelihood_loss": 0.12685106694698334,
      "theoretical_loss": 4.450131961015606,
      "tokens_seen": 168755200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047921765295887663,
      "loss": 1.7889,
      "theoretical_loss": 4.449910056790086,
      "tokens_seen": 168820736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047919759277833504,
      "loss": 1.6048,
      "theoretical_loss": 4.4494665789482895,
      "tokens_seen": 168951808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791775325977934,
      "loss": 1.5793,
      "theoretical_loss": 4.449023541269395,
      "tokens_seen": 169082880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047915747241725177,
      "loss": 1.6751,
      "theoretical_loss": 4.448580942975825,
      "tokens_seen": 169213952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047913741223671013,
      "loss": 1.7455,
      "theoretical_loss": 4.448138783291979,
      "tokens_seen": 169345024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047911735205616855,
      "loss": 1.7699,
      "theoretical_loss": 4.44769706144422,
      "tokens_seen": 169476096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790972918756269,
      "loss": 1.6299,
      "theoretical_loss": 4.447255776660878,
      "tokens_seen": 169607168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047907723169508527,
      "loss": 1.5542,
      "theoretical_loss": 4.446814928172234,
      "tokens_seen": 169738240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047905717151454363,
      "loss": 1.7944,
      "theoretical_loss": 4.446374515210521,
      "tokens_seen": 169869312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479037111334002,
      "loss": 1.6775,
      "theoretical_loss": 4.445934537009911,
      "tokens_seen": 170000384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790170511534604,
      "loss": 1.8059,
      "theoretical_loss": 4.445494992806513,
      "tokens_seen": 170131456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789969909729188,
      "loss": 1.8668,
      "theoretical_loss": 4.445055881838365,
      "tokens_seen": 170262528
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.41752198338508606,
      "objective/train/avg_token_score": 0.02483731135725975,
      "objective/train/docs_used": 107830,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6519962549209595,
      "objective/train/original_loss": 4.723080635070801,
      "objective/train/theoretical_loss": 4.44461720334543,
      "objective/train/tokens_used": 170393600,
      "objective/train/unlikelihood_loss": 0.11609357595443726,
      "theoretical_loss": 4.44461720334543,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047897693079237714,
      "loss": 1.7401,
      "theoretical_loss": 4.44461720334543,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789568706118355,
      "loss": 1.6675,
      "theoretical_loss": 4.444178956569585,
      "tokens_seen": 170524672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789368104312939,
      "loss": 1.7224,
      "theoretical_loss": 4.44374114075462,
      "tokens_seen": 170655744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789167502507523,
      "loss": 1.6234,
      "theoretical_loss": 4.443303755146225,
      "tokens_seen": 170786816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047889669007021064,
      "loss": 1.6098,
      "theoretical_loss": 4.442866798991993,
      "tokens_seen": 170917888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478876629889669,
      "loss": 1.749,
      "theoretical_loss": 4.442430271541404,
      "tokens_seen": 171048960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047885656970912736,
      "loss": 1.6635,
      "theoretical_loss": 4.4419941720458285,
      "tokens_seen": 171180032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788365095285858,
      "loss": 1.8943,
      "theoretical_loss": 4.441558499758511,
      "tokens_seen": 171311104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047881644934804414,
      "loss": 1.6301,
      "theoretical_loss": 4.441123253934572,
      "tokens_seen": 171442176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787963891675025,
      "loss": 1.5459,
      "theoretical_loss": 4.440688433830999,
      "tokens_seen": 171573248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047877632898696087,
      "loss": 1.7699,
      "theoretical_loss": 4.440254038706639,
      "tokens_seen": 171704320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787562688064193,
      "loss": 1.6063,
      "theoretical_loss": 4.439820067822195,
      "tokens_seen": 171835392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047873620862587765,
      "loss": 1.7779,
      "theoretical_loss": 4.439386520440218,
      "tokens_seen": 171966464
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.3144550323486328,
      "objective/train/avg_token_score": 0.00989236868917942,
      "objective/train/docs_used": 108471,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1930038928985596,
      "objective/train/original_loss": 4.118610382080078,
      "objective/train/theoretical_loss": 4.439169905332706,
      "objective/train/tokens_used": 172032000,
      "objective/train/unlikelihood_loss": 0.1678958386182785,
      "theoretical_loss": 4.439169905332706,
      "tokens_seen": 172032000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478716148445336,
      "loss": 1.4924,
      "theoretical_loss": 4.438953395825102,
      "tokens_seen": 172097536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047869608826479437,
      "loss": 1.6309,
      "theoretical_loss": 4.438520693243079,
      "tokens_seen": 172228608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047867602808425273,
      "loss": 1.5903,
      "theoretical_loss": 4.438088411962211,
      "tokens_seen": 172359680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047865596790371115,
      "loss": 1.5514,
      "theoretical_loss": 4.437656551252381,
      "tokens_seen": 172490752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786359077231695,
      "loss": 1.8921,
      "theoretical_loss": 4.437225110385297,
      "tokens_seen": 172621824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786158475426279,
      "loss": 1.6624,
      "theoretical_loss": 4.436794088634477,
      "tokens_seen": 172752896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047859578736208624,
      "loss": 1.653,
      "theoretical_loss": 4.436363485275246,
      "tokens_seen": 172883968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047857572718154465,
      "loss": 1.6515,
      "theoretical_loss": 4.435933299584729,
      "tokens_seen": 173015040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478555667001003,
      "loss": 1.6443,
      "theoretical_loss": 4.435503530841849,
      "tokens_seen": 173146112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785356068204614,
      "loss": 1.6424,
      "theoretical_loss": 4.4350741783273175,
      "tokens_seen": 173277184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047851554663991974,
      "loss": 1.7407,
      "theoretical_loss": 4.434645241323629,
      "tokens_seen": 173408256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784954864593781,
      "loss": 1.6988,
      "theoretical_loss": 4.434216719115057,
      "tokens_seen": 173539328
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.38972383737564087,
      "objective/train/avg_token_score": 0.009662693366408348,
      "objective/train/docs_used": 109645,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5914331674575806,
      "objective/train/original_loss": 4.367075443267822,
      "objective/train/theoretical_loss": 4.433788610987646,
      "objective/train/tokens_used": 173670400,
      "objective/train/unlikelihood_loss": 0.1352538764476776,
      "theoretical_loss": 4.433788610987646,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784754262788365,
      "loss": 1.7224,
      "theoretical_loss": 4.433788610987646,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784553660982949,
      "loss": 1.7154,
      "theoretical_loss": 4.433360916229209,
      "tokens_seen": 173801472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047843530591775324,
      "loss": 1.6758,
      "theoretical_loss": 4.432933634129318,
      "tokens_seen": 173932544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784152457372116,
      "loss": 1.6071,
      "theoretical_loss": 4.432506763979299,
      "tokens_seen": 174063616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047839518555667,
      "loss": 1.6384,
      "theoretical_loss": 4.432080305072233,
      "tokens_seen": 174194688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783751253761284,
      "loss": 1.5316,
      "theoretical_loss": 4.431654256702938,
      "tokens_seen": 174325760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047835506519558675,
      "loss": 1.614,
      "theoretical_loss": 4.4312286181679745,
      "tokens_seen": 174456832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783350050150451,
      "loss": 1.9548,
      "theoretical_loss": 4.430803388765636,
      "tokens_seen": 174587904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047831494483450347,
      "loss": 1.5284,
      "theoretical_loss": 4.430378567795938,
      "tokens_seen": 174718976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782948846539619,
      "loss": 1.6997,
      "theoretical_loss": 4.429954154560624,
      "tokens_seen": 174850048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047827482447342025,
      "loss": 1.6483,
      "theoretical_loss": 4.429530148363151,
      "tokens_seen": 174981120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782547642928786,
      "loss": 1.5161,
      "theoretical_loss": 4.429106548508685,
      "tokens_seen": 175112192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478234704112337,
      "loss": 1.7931,
      "theoretical_loss": 4.428683354304098,
      "tokens_seen": 175243264
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.3965053856372833,
      "objective/train/avg_token_score": 0.01913469098508358,
      "objective/train/docs_used": 110166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7019623517990112,
      "objective/train/original_loss": 4.4854278564453125,
      "objective/train/theoretical_loss": 4.428471909104372,
      "objective/train/tokens_used": 175308800,
      "objective/train/unlikelihood_loss": 0.11818500608205795,
      "theoretical_loss": 4.428471909104372,
      "tokens_seen": 175308800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047821464393179544,
      "loss": 1.5788,
      "theoretical_loss": 4.428260565057964,
      "tokens_seen": 175374336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781945837512538,
      "loss": 1.786,
      "theoretical_loss": 4.427838180080547,
      "tokens_seen": 175505408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047817452357071217,
      "loss": 1.7903,
      "theoretical_loss": 4.427416198683803,
      "tokens_seen": 175636480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047815446339017053,
      "loss": 1.7291,
      "theoretical_loss": 4.42699462018137,
      "tokens_seen": 175767552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047813440320962895,
      "loss": 1.8156,
      "theoretical_loss": 4.426573443888563,
      "tokens_seen": 175898624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781143430290873,
      "loss": 1.6172,
      "theoretical_loss": 4.426152669122374,
      "tokens_seen": 176029696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047809428284854567,
      "loss": 1.5866,
      "theoretical_loss": 4.425732295201455,
      "tokens_seen": 176160768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047807422266800403,
      "loss": 1.6032,
      "theoretical_loss": 4.425312321446127,
      "tokens_seen": 176291840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780541624874624,
      "loss": 1.7259,
      "theoretical_loss": 4.424892747178365,
      "tokens_seen": 176422912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780341023069208,
      "loss": 1.5495,
      "theoretical_loss": 4.424473571721794,
      "tokens_seen": 176553984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780140421263792,
      "loss": 1.5823,
      "theoretical_loss": 4.424054794401689,
      "tokens_seen": 176685056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047799398194583754,
      "loss": 1.5334,
      "theoretical_loss": 4.42363641454496,
      "tokens_seen": 176816128
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.27394917607307434,
      "objective/train/avg_token_score": 0.024683641269803047,
      "objective/train/docs_used": 110915,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1538923978805542,
      "objective/train/original_loss": 4.269205570220947,
      "objective/train/theoretical_loss": 4.42321843148016,
      "objective/train/tokens_used": 176947200,
      "objective/train/unlikelihood_loss": 0.17650169134140015,
      "theoretical_loss": 4.42321843148016,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779739217652959,
      "loss": 1.5801,
      "theoretical_loss": 4.42321843148016,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779538615847543,
      "loss": 1.743,
      "theoretical_loss": 4.422800844537466,
      "tokens_seen": 177078272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779338014042127,
      "loss": 1.7594,
      "theoretical_loss": 4.422383653048685,
      "tokens_seen": 177209344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047791374122367104,
      "loss": 1.7924,
      "theoretical_loss": 4.421966856347243,
      "tokens_seen": 177340416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778936810431294,
      "loss": 1.9118,
      "theoretical_loss": 4.421550453768181,
      "tokens_seen": 177471488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047787362086258776,
      "loss": 1.6296,
      "theoretical_loss": 4.42113444464815,
      "tokens_seen": 177602560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778535606820462,
      "loss": 1.512,
      "theoretical_loss": 4.420718828325403,
      "tokens_seen": 177733632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047783350050150454,
      "loss": 1.7201,
      "theoretical_loss": 4.4203036041398,
      "tokens_seen": 177864704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778134403209629,
      "loss": 1.6327,
      "theoretical_loss": 4.419888771432789,
      "tokens_seen": 177995776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047779338014042127,
      "loss": 1.5254,
      "theoretical_loss": 4.419474329547413,
      "tokens_seen": 178126848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777733199598797,
      "loss": 1.8315,
      "theoretical_loss": 4.419060277828295,
      "tokens_seen": 178257920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047775325977933805,
      "loss": 1.6243,
      "theoretical_loss": 4.4186466156216415,
      "tokens_seen": 178388992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777331995987964,
      "loss": 1.6067,
      "theoretical_loss": 4.418233342275233,
      "tokens_seen": 178520064
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.4303519129753113,
      "objective/train/avg_token_score": 0.005776857491582632,
      "objective/train/docs_used": 112413,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6707626581192017,
      "objective/train/original_loss": 3.843381643295288,
      "objective/train/theoretical_loss": 4.418026851221231,
      "objective/train/tokens_used": 178585600,
      "objective/train/unlikelihood_loss": 0.19406434893608093,
      "theoretical_loss": 4.418026851221231,
      "tokens_seen": 178585600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047771313941825477,
      "loss": 1.5482,
      "theoretical_loss": 4.4178204571384185,
      "tokens_seen": 178651136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047769307923771313,
      "loss": 1.5164,
      "theoretical_loss": 4.417407959562116,
      "tokens_seen": 178782208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047767301905717155,
      "loss": 1.6927,
      "theoretical_loss": 4.416995848898797,
      "tokens_seen": 178913280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776529588766299,
      "loss": 1.496,
      "theoretical_loss": 4.416584124502495,
      "tokens_seen": 179044352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776328986960883,
      "loss": 1.5685,
      "theoretical_loss": 4.41617278572879,
      "tokens_seen": 179175424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047761283851554664,
      "loss": 1.4586,
      "theoretical_loss": 4.415761831934808,
      "tokens_seen": 179306496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047759277833500505,
      "loss": 1.4191,
      "theoretical_loss": 4.415351262479216,
      "tokens_seen": 179437568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775727181544634,
      "loss": 1.5665,
      "theoretical_loss": 4.414941076722219,
      "tokens_seen": 179568640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775526579739218,
      "loss": 1.545,
      "theoretical_loss": 4.414531274025548,
      "tokens_seen": 179699712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047753259779338014,
      "loss": 1.5296,
      "theoretical_loss": 4.414121853752466,
      "tokens_seen": 179830784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775125376128385,
      "loss": 1.5963,
      "theoretical_loss": 4.413712815267752,
      "tokens_seen": 179961856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774924774322969,
      "loss": 1.8008,
      "theoretical_loss": 4.4133041579377075,
      "tokens_seen": 180092928
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.39986559748649597,
      "objective/train/avg_token_score": 0.06428433954715729,
      "objective/train/docs_used": 112982,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4951305389404297,
      "objective/train/original_loss": 4.130880355834961,
      "objective/train/theoretical_loss": 4.412895881130142,
      "objective/train/tokens_used": 180224000,
      "objective/train/unlikelihood_loss": 0.13160917162895203,
      "theoretical_loss": 4.412895881130142,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774724172517553,
      "loss": 1.6337,
      "theoretical_loss": 4.412895881130142,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047745235707121364,
      "loss": 1.658,
      "theoretical_loss": 4.412487984214373,
      "tokens_seen": 180355072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000477432296890672,
      "loss": 1.5547,
      "theoretical_loss": 4.412080466561221,
      "tokens_seen": 180486144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774122367101304,
      "loss": 1.7956,
      "theoretical_loss": 4.411673327543005,
      "tokens_seen": 180617216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773921765295888,
      "loss": 1.7189,
      "theoretical_loss": 4.411266566533539,
      "tokens_seen": 180748288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047737211634904715,
      "loss": 1.6131,
      "theoretical_loss": 4.41086018290812,
      "tokens_seen": 180879360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773520561685055,
      "loss": 1.7809,
      "theoretical_loss": 4.410454176043537,
      "tokens_seen": 181010432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047733199598796387,
      "loss": 1.5577,
      "theoretical_loss": 4.410048545318052,
      "tokens_seen": 181141504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773119358074223,
      "loss": 1.4906,
      "theoretical_loss": 4.409643290111404,
      "tokens_seen": 181272576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047729187562688065,
      "loss": 1.5827,
      "theoretical_loss": 4.409238409804804,
      "tokens_seen": 181403648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477271815446339,
      "loss": 1.6267,
      "theoretical_loss": 4.408833903780926,
      "tokens_seen": 181534720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772517552657974,
      "loss": 1.6512,
      "theoretical_loss": 4.408429771423909,
      "tokens_seen": 181665792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772316950852558,
      "loss": 1.7169,
      "theoretical_loss": 4.408026012119344,
      "tokens_seen": 181796864
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.3759775161743164,
      "objective/train/avg_token_score": 0.011839372105896473,
      "objective/train/docs_used": 114216,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4656167030334473,
      "objective/train/original_loss": 4.162503719329834,
      "objective/train/theoretical_loss": 4.407824272170128,
      "objective/train/tokens_used": 181862400,
      "objective/train/unlikelihood_loss": 0.16261805593967438,
      "theoretical_loss": 4.407824272170128,
      "tokens_seen": 181862400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047721163490471415,
      "loss": 1.7268,
      "theoretical_loss": 4.407622625254279,
      "tokens_seen": 181927936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771915747241725,
      "loss": 1.6943,
      "theoretical_loss": 4.407219610217206,
      "tokens_seen": 182059008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771715145436309,
      "loss": 1.7274,
      "theoretical_loss": 4.406816966398064,
      "tokens_seen": 182190080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047715145436308924,
      "loss": 1.539,
      "theoretical_loss": 4.4064146931882275,
      "tokens_seen": 182321152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047713139418254766,
      "loss": 1.6115,
      "theoretical_loss": 4.406012789980506,
      "tokens_seen": 182452224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477111334002006,
      "loss": 1.6946,
      "theoretical_loss": 4.405611256169143,
      "tokens_seen": 182583296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770912738214644,
      "loss": 1.69,
      "theoretical_loss": 4.405210091149802,
      "tokens_seen": 182714368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047707121364092274,
      "loss": 1.7164,
      "theoretical_loss": 4.404809294319572,
      "tokens_seen": 182845440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047705115346038116,
      "loss": 1.7673,
      "theoretical_loss": 4.404408865076955,
      "tokens_seen": 182976512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770310932798395,
      "loss": 1.7797,
      "theoretical_loss": 4.404008802821871,
      "tokens_seen": 183107584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770110330992979,
      "loss": 1.7046,
      "theoretical_loss": 4.403609106955645,
      "tokens_seen": 183238656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047699097291875625,
      "loss": 1.5983,
      "theoretical_loss": 4.403209776881004,
      "tokens_seen": 183369728
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.4590664803981781,
      "objective/train/avg_token_score": 0.02654801867902279,
      "objective/train/docs_used": 114775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.934823751449585,
      "objective/train/original_loss": 4.4886064529418945,
      "objective/train/theoretical_loss": 4.4028108120020795,
      "objective/train/tokens_used": 183500800,
      "objective/train/unlikelihood_loss": 0.1138959601521492,
      "theoretical_loss": 4.4028108120020795,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047697091273821466,
      "loss": 1.7135,
      "theoretical_loss": 4.4028108120020795,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476950852557673,
      "loss": 1.6734,
      "theoretical_loss": 4.4024122117243945,
      "tokens_seen": 183631872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769307923771314,
      "loss": 1.8335,
      "theoretical_loss": 4.4020139754548655,
      "tokens_seen": 183762944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047691073219658975,
      "loss": 1.6858,
      "theoretical_loss": 4.4016161026017935,
      "tokens_seen": 183894016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768906720160481,
      "loss": 1.7323,
      "theoretical_loss": 4.401218592574865,
      "tokens_seen": 184025088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047687061183550653,
      "loss": 1.8249,
      "theoretical_loss": 4.400821444785143,
      "tokens_seen": 184156160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768505516549649,
      "loss": 1.588,
      "theoretical_loss": 4.400424658645065,
      "tokens_seen": 184287232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047683049147442325,
      "loss": 1.6622,
      "theoretical_loss": 4.40002823356844,
      "tokens_seen": 184418304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768104312938816,
      "loss": 1.6254,
      "theoretical_loss": 4.39963216897044,
      "tokens_seen": 184549376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047679037111334003,
      "loss": 1.7437,
      "theoretical_loss": 4.399236464267602,
      "tokens_seen": 184680448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767703109327984,
      "loss": 1.6554,
      "theoretical_loss": 4.398841118877819,
      "tokens_seen": 184811520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047675025075225676,
      "loss": 1.6995,
      "theoretical_loss": 4.398446132220338,
      "tokens_seen": 184942592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767301905717151,
      "loss": 1.9342,
      "theoretical_loss": 4.398051503715753,
      "tokens_seen": 185073664
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.5482649207115173,
      "objective/train/avg_token_score": 0.00839152093976736,
      "objective/train/docs_used": 115939,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.310957908630371,
      "objective/train/original_loss": 4.36484956741333,
      "objective/train/theoretical_loss": 4.397854323590102,
      "objective/train/tokens_used": 185139200,
      "objective/train/unlikelihood_loss": 0.10018301010131836,
      "theoretical_loss": 4.397854323590102,
      "tokens_seen": 185139200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047671013039117353,
      "loss": 1.7113,
      "theoretical_loss": 4.397657232786008,
      "tokens_seen": 185204736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047669007021063195,
      "loss": 1.6987,
      "theoretical_loss": 4.397263318854384,
      "tokens_seen": 185335808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766700100300903,
      "loss": 1.5346,
      "theoretical_loss": 4.396869761345503,
      "tokens_seen": 185466880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766499498495487,
      "loss": 1.7575,
      "theoretical_loss": 4.396476559685315,
      "tokens_seen": 185597952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047662988966900704,
      "loss": 1.6408,
      "theoretical_loss": 4.3960837133011035,
      "tokens_seen": 185729024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047660982948846545,
      "loss": 1.8307,
      "theoretical_loss": 4.395691221621476,
      "tokens_seen": 185860096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765897693079238,
      "loss": 1.8426,
      "theoretical_loss": 4.39529908407636,
      "tokens_seen": 185991168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765697091273822,
      "loss": 1.6831,
      "theoretical_loss": 4.394907300097002,
      "tokens_seen": 186122240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047654964894684054,
      "loss": 1.6989,
      "theoretical_loss": 4.39451586911596,
      "tokens_seen": 186253312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765295887662989,
      "loss": 1.7103,
      "theoretical_loss": 4.394124790567101,
      "tokens_seen": 186384384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765095285857573,
      "loss": 1.7273,
      "theoretical_loss": 4.393734063885599,
      "tokens_seen": 186515456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764894684052157,
      "loss": 1.6129,
      "theoretical_loss": 4.39334368850793,
      "tokens_seen": 186646528
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.47330155968666077,
      "objective/train/avg_token_score": 0.007206100504845381,
      "objective/train/docs_used": 116562,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.913322925567627,
      "objective/train/original_loss": 4.169583797454834,
      "objective/train/theoretical_loss": 4.392953663871862,
      "objective/train/tokens_used": 186777600,
      "objective/train/unlikelihood_loss": 0.11674590408802032,
      "theoretical_loss": 4.392953663871862,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047646940822467404,
      "loss": 1.7131,
      "theoretical_loss": 4.392953663871862,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764493480441324,
      "loss": 1.7657,
      "theoretical_loss": 4.392563989416462,
      "tokens_seen": 186908672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764292878635908,
      "loss": 1.7479,
      "theoretical_loss": 4.392174664582085,
      "tokens_seen": 187039744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764092276830492,
      "loss": 1.7493,
      "theoretical_loss": 4.391785688810373,
      "tokens_seen": 187170816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047638916750250755,
      "loss": 1.7387,
      "theoretical_loss": 4.391397061544247,
      "tokens_seen": 187301888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763691073219659,
      "loss": 1.6152,
      "theoretical_loss": 4.3910087822279085,
      "tokens_seen": 187432960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047634904714142427,
      "loss": 1.5353,
      "theoretical_loss": 4.390620850306832,
      "tokens_seen": 187564032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763289869608827,
      "loss": 1.5725,
      "theoretical_loss": 4.390233265227764,
      "tokens_seen": 187695104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047630892678034105,
      "loss": 1.631,
      "theoretical_loss": 4.389846026438715,
      "tokens_seen": 187826176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762888665997994,
      "loss": 1.7898,
      "theoretical_loss": 4.389459133388962,
      "tokens_seen": 187957248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762688064192578,
      "loss": 1.5287,
      "theoretical_loss": 4.389072585529037,
      "tokens_seen": 188088320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762487462387162,
      "loss": 1.4302,
      "theoretical_loss": 4.388686382310732,
      "tokens_seen": 188219392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047622868605817455,
      "loss": 1.8187,
      "theoretical_loss": 4.388300523187087,
      "tokens_seen": 188350464
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.37365591526031494,
      "objective/train/avg_token_score": 0.0153102558106184,
      "objective/train/docs_used": 117858,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5269169807434082,
      "objective/train/original_loss": 4.470926284790039,
      "objective/train/theoretical_loss": 4.3881077224901786,
      "objective/train/tokens_used": 188416000,
      "objective/train/unlikelihood_loss": 0.11220584064722061,
      "theoretical_loss": 4.3881077224901786,
      "tokens_seen": 188416000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762086258776329,
      "loss": 1.5687,
      "theoretical_loss": 4.38791500761239,
      "tokens_seen": 188481536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761885656970913,
      "loss": 1.5935,
      "theoretical_loss": 4.3875298350421765,
      "tokens_seen": 188612608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047616850551654964,
      "loss": 1.7741,
      "theoretical_loss": 4.387145004933218,
      "tokens_seen": 188743680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047614844533600806,
      "loss": 1.4192,
      "theoretical_loss": 4.386760516743526,
      "tokens_seen": 188874752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761283851554664,
      "loss": 1.8067,
      "theoretical_loss": 4.386376369932344,
      "tokens_seen": 189005824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761083249749248,
      "loss": 1.6904,
      "theoretical_loss": 4.385992563960145,
      "tokens_seen": 189136896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047608826479438314,
      "loss": 1.7294,
      "theoretical_loss": 4.385609098288628,
      "tokens_seen": 189267968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047606820461384156,
      "loss": 1.5139,
      "theoretical_loss": 4.385225972380715,
      "tokens_seen": 189399040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760481444332999,
      "loss": 1.5383,
      "theoretical_loss": 4.384843185700544,
      "tokens_seen": 189530112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760280842527583,
      "loss": 1.745,
      "theoretical_loss": 4.384460737713471,
      "tokens_seen": 189661184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047600802407221665,
      "loss": 1.6476,
      "theoretical_loss": 4.384078627886062,
      "tokens_seen": 189792256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047598796389167506,
      "loss": 1.7629,
      "theoretical_loss": 4.38369685568609,
      "tokens_seen": 189923328
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.3760386109352112,
      "objective/train/avg_token_score": 0.010193110443651676,
      "objective/train/docs_used": 118609,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4706860780715942,
      "objective/train/original_loss": 3.930041790008545,
      "objective/train/theoretical_loss": 4.383315420582533,
      "objective/train/tokens_used": 190054400,
      "objective/train/unlikelihood_loss": 0.24754564464092255,
      "theoretical_loss": 4.383315420582533,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759679037111334,
      "loss": 1.862,
      "theoretical_loss": 4.383315420582533,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759478435305918,
      "loss": 1.5947,
      "theoretical_loss": 4.38293432204557,
      "tokens_seen": 190185472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047592778335005015,
      "loss": 1.7051,
      "theoretical_loss": 4.3825535595465785,
      "tokens_seen": 190316544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759077231695085,
      "loss": 1.6088,
      "theoretical_loss": 4.382173132558126,
      "tokens_seen": 190447616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047588766298896693,
      "loss": 1.6397,
      "theoretical_loss": 4.381793040553973,
      "tokens_seen": 190578688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758676028084253,
      "loss": 1.662,
      "theoretical_loss": 4.381413283009065,
      "tokens_seen": 190709760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047584754262788365,
      "loss": 1.5988,
      "theoretical_loss": 4.381033859399532,
      "tokens_seen": 190840832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475827482447342,
      "loss": 1.7682,
      "theoretical_loss": 4.380654769202683,
      "tokens_seen": 190971904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047580742226680043,
      "loss": 1.6793,
      "theoretical_loss": 4.380276011897003,
      "tokens_seen": 191102976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757873620862588,
      "loss": 1.5176,
      "theoretical_loss": 4.379897586962148,
      "tokens_seen": 191234048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047576730190571716,
      "loss": 1.5289,
      "theoretical_loss": 4.379519493878948,
      "tokens_seen": 191365120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757472417251755,
      "loss": 1.646,
      "theoretical_loss": 4.379141732129394,
      "tokens_seen": 191496192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757271815446339,
      "loss": 1.6823,
      "theoretical_loss": 4.378764301196642,
      "tokens_seen": 191627264
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.46615347266197205,
      "objective/train/avg_token_score": 0.02856089174747467,
      "objective/train/docs_used": 119339,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6883468627929688,
      "objective/train/original_loss": 4.204484462738037,
      "objective/train/theoretical_loss": 4.37857570962537,
      "objective/train/tokens_used": 191692800,
      "objective/train/unlikelihood_loss": 0.11059554666280746,
      "theoretical_loss": 4.37857570962537,
      "tokens_seen": 191692800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757071213640923,
      "loss": 1.699,
      "theoretical_loss": 4.378387200565006,
      "tokens_seen": 191758336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047568706118355066,
      "loss": 1.8261,
      "theoretical_loss": 4.378010429719957,
      "tokens_seen": 191889408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475667001003009,
      "loss": 1.7145,
      "theoretical_loss": 4.377633988148117,
      "tokens_seen": 192020480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756469408224674,
      "loss": 1.5378,
      "theoretical_loss": 4.377257875337257,
      "tokens_seen": 192151552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756268806419258,
      "loss": 1.8218,
      "theoretical_loss": 4.376882090776293,
      "tokens_seen": 192282624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047560682046138416,
      "loss": 1.7853,
      "theoretical_loss": 4.376506633955286,
      "tokens_seen": 192413696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755867602808425,
      "loss": 1.4124,
      "theoretical_loss": 4.37613150436543,
      "tokens_seen": 192544768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755667001003009,
      "loss": 1.6803,
      "theoretical_loss": 4.375756701499063,
      "tokens_seen": 192675840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047554663991975925,
      "loss": 1.5873,
      "theoretical_loss": 4.375382224849648,
      "tokens_seen": 192806912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047552657973921766,
      "loss": 1.5583,
      "theoretical_loss": 4.375008073911781,
      "tokens_seen": 192937984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047550651955867603,
      "loss": 1.6886,
      "theoretical_loss": 4.374634248181182,
      "tokens_seen": 193069056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754864593781344,
      "loss": 1.6532,
      "theoretical_loss": 4.374260747154692,
      "tokens_seen": 193200128
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.4188660681247711,
      "objective/train/avg_token_score": 0.00520719401538372,
      "objective/train/docs_used": 120055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6985862255096436,
      "objective/train/original_loss": 4.125784873962402,
      "objective/train/theoretical_loss": 4.373887570330275,
      "objective/train/tokens_used": 193331200,
      "objective/train/unlikelihood_loss": 0.16551171243190765,
      "theoretical_loss": 4.373887570330275,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047546639919759275,
      "loss": 1.6246,
      "theoretical_loss": 4.373887570330275,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047544633901705117,
      "loss": 1.4885,
      "theoretical_loss": 4.373514717207009,
      "tokens_seen": 193462272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047542627883650953,
      "loss": 1.4034,
      "theoretical_loss": 4.373142187285083,
      "tokens_seen": 193593344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754062186559679,
      "loss": 1.4975,
      "theoretical_loss": 4.372769980065797,
      "tokens_seen": 193724416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047538615847542625,
      "loss": 1.6481,
      "theoretical_loss": 4.372398095051559,
      "tokens_seen": 193855488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753660982948846,
      "loss": 1.4715,
      "theoretical_loss": 4.372026531745877,
      "tokens_seen": 193986560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047534603811434303,
      "loss": 1.4546,
      "theoretical_loss": 4.37165528965336,
      "tokens_seen": 194117632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753259779338014,
      "loss": 1.4513,
      "theoretical_loss": 4.371284368279714,
      "tokens_seen": 194248704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047530591775325976,
      "loss": 1.7305,
      "theoretical_loss": 4.3709137671317375,
      "tokens_seen": 194379776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752858575727181,
      "loss": 1.4355,
      "theoretical_loss": 4.370543485717322,
      "tokens_seen": 194510848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047526579739217654,
      "loss": 1.5039,
      "theoretical_loss": 4.370173523545443,
      "tokens_seen": 194641920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752457372116349,
      "loss": 1.6506,
      "theoretical_loss": 4.369803880126162,
      "tokens_seen": 194772992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047522567703109326,
      "loss": 1.4202,
      "theoretical_loss": 4.369434554970621,
      "tokens_seen": 194904064
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.42467010021209717,
      "objective/train/avg_token_score": 0.023077575489878654,
      "objective/train/docs_used": 121265,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6124404668807983,
      "objective/train/original_loss": 3.863985300064087,
      "objective/train/theoretical_loss": 4.369250011589279,
      "objective/train/tokens_used": 194969600,
      "objective/train/unlikelihood_loss": 0.3769527077674866,
      "theoretical_loss": 4.369250011589279,
      "tokens_seen": 194969600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752056168505517,
      "loss": 1.6853,
      "theoretical_loss": 4.369065547591038,
      "tokens_seen": 195035136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047518555667001004,
      "loss": 1.7059,
      "theoretical_loss": 4.36869685750071,
      "tokens_seen": 195166208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047516549648946846,
      "loss": 1.6439,
      "theoretical_loss": 4.368328484214002,
      "tokens_seen": 195297280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751454363089268,
      "loss": 1.5728,
      "theoretical_loss": 4.3679604272463495,
      "tokens_seen": 195428352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751253761283852,
      "loss": 1.6146,
      "theoretical_loss": 4.367592686114252,
      "tokens_seen": 195559424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047510531594784354,
      "loss": 1.563,
      "theoretical_loss": 4.367225260335272,
      "tokens_seen": 195690496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047508525576730196,
      "loss": 1.6781,
      "theoretical_loss": 4.366858149428032,
      "tokens_seen": 195821568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750651955867603,
      "loss": 1.5683,
      "theoretical_loss": 4.366491352912211,
      "tokens_seen": 195952640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750451354062187,
      "loss": 1.7887,
      "theoretical_loss": 4.366124870308541,
      "tokens_seen": 196083712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047502507522567705,
      "loss": 1.5592,
      "theoretical_loss": 4.3657587011388035,
      "tokens_seen": 196214784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047500501504513546,
      "loss": 1.7372,
      "theoretical_loss": 4.365392844925829,
      "tokens_seen": 196345856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749849548645938,
      "loss": 1.6439,
      "theoretical_loss": 4.365027301193491,
      "tokens_seen": 196476928
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.4978005886077881,
      "objective/train/avg_token_score": 0.0068794176913797855,
      "objective/train/docs_used": 121711,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8076094388961792,
      "objective/train/original_loss": 4.219160556793213,
      "objective/train/theoretical_loss": 4.364662069466704,
      "objective/train/tokens_used": 196608000,
      "objective/train/unlikelihood_loss": 0.12803588807582855,
      "theoretical_loss": 4.364662069466704,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749648946840522,
      "loss": 1.6509,
      "theoretical_loss": 4.364662069466704,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047494483450351055,
      "loss": 1.4878,
      "theoretical_loss": 4.364297149271423,
      "tokens_seen": 196739072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749247743229689,
      "loss": 1.5957,
      "theoretical_loss": 4.363932540134638,
      "tokens_seen": 196870144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047490471414242733,
      "loss": 1.705,
      "theoretical_loss": 4.363568241584368,
      "tokens_seen": 197001216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748846539618857,
      "loss": 1.5305,
      "theoretical_loss": 4.363204253149667,
      "tokens_seen": 197132288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047486459378134405,
      "loss": 1.7872,
      "theoretical_loss": 4.362840574360612,
      "tokens_seen": 197263360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748445336008024,
      "loss": 1.7201,
      "theoretical_loss": 4.362477204748305,
      "tokens_seen": 197394432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047482447342026083,
      "loss": 1.7059,
      "theoretical_loss": 4.362114143844867,
      "tokens_seen": 197525504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748044132397192,
      "loss": 1.557,
      "theoretical_loss": 4.3617513911834385,
      "tokens_seen": 197656576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047478435305917756,
      "loss": 1.6276,
      "theoretical_loss": 4.3613889462981765,
      "tokens_seen": 197787648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747642928786359,
      "loss": 1.5583,
      "theoretical_loss": 4.361026808724247,
      "tokens_seen": 197918720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747442326980943,
      "loss": 1.7629,
      "theoretical_loss": 4.360664977997828,
      "tokens_seen": 198049792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747241725175527,
      "loss": 1.4869,
      "theoretical_loss": 4.360303453656103,
      "tokens_seen": 198180864
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.3821480870246887,
      "objective/train/avg_token_score": 0.010640091262757778,
      "objective/train/docs_used": 122248,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6288254261016846,
      "objective/train/original_loss": 4.3907952308654785,
      "objective/train/theoretical_loss": 4.360122806235145,
      "objective/train/tokens_used": 198246400,
      "objective/train/unlikelihood_loss": 0.1243598461151123,
      "theoretical_loss": 4.360122806235145,
      "tokens_seen": 198246400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047470411233701106,
      "loss": 1.6806,
      "theoretical_loss": 4.359942235237257,
      "tokens_seen": 198311936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746840521564694,
      "loss": 1.6821,
      "theoretical_loss": 4.359581322280479,
      "tokens_seen": 198443008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746639919759278,
      "loss": 1.7165,
      "theoretical_loss": 4.359220714325954,
      "tokens_seen": 198574080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746439317953862,
      "loss": 1.6884,
      "theoretical_loss": 4.358860410914861,
      "tokens_seen": 198705152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047462387161484456,
      "loss": 1.6764,
      "theoretical_loss": 4.358500411589375,
      "tokens_seen": 198836224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746038114343029,
      "loss": 1.874,
      "theoretical_loss": 4.358140715892658,
      "tokens_seen": 198967296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745837512537613,
      "loss": 1.5337,
      "theoretical_loss": 4.357781323368857,
      "tokens_seen": 199098368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047456369107321965,
      "loss": 1.916,
      "theoretical_loss": 4.357422233563106,
      "tokens_seen": 199229440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047454363089267807,
      "loss": 1.5149,
      "theoretical_loss": 4.35706344602152,
      "tokens_seen": 199360512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047452357071213643,
      "loss": 1.6638,
      "theoretical_loss": 4.356704960291191,
      "tokens_seen": 199491584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745035105315948,
      "loss": 1.7224,
      "theoretical_loss": 4.356346775920185,
      "tokens_seen": 199622656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047448345035105315,
      "loss": 1.6415,
      "theoretical_loss": 4.355988892457546,
      "tokens_seen": 199753728
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.5615224838256836,
      "objective/train/avg_token_score": 0.008392857387661934,
      "objective/train/docs_used": 123553,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.331106185913086,
      "objective/train/original_loss": 4.441834449768066,
      "objective/train/theoretical_loss": 4.355631309453283,
      "objective/train/tokens_used": 199884800,
      "objective/train/unlikelihood_loss": 0.08705087006092072,
      "theoretical_loss": 4.355631309453283,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047446339017051157,
      "loss": 1.9122,
      "theoretical_loss": 4.355631309453283,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047444332998996993,
      "loss": 1.8121,
      "theoretical_loss": 4.355274026458375,
      "tokens_seen": 200015872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744232698094283,
      "loss": 1.7729,
      "theoretical_loss": 4.354917043024765,
      "tokens_seen": 200146944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047440320962888666,
      "loss": 1.6087,
      "theoretical_loss": 4.354560358705358,
      "tokens_seen": 200278016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474383149448345,
      "loss": 1.4376,
      "theoretical_loss": 4.35420397305402,
      "tokens_seen": 200409088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047436308926780343,
      "loss": 1.8095,
      "theoretical_loss": 4.353847885625571,
      "tokens_seen": 200540160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743430290872618,
      "loss": 1.672,
      "theoretical_loss": 4.353492095975787,
      "tokens_seen": 200671232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047432296890672016,
      "loss": 1.7417,
      "theoretical_loss": 4.353136603661392,
      "tokens_seen": 200802304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743029087261785,
      "loss": 1.6302,
      "theoretical_loss": 4.352781408240065,
      "tokens_seen": 200933376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047428284854563694,
      "loss": 1.7316,
      "theoretical_loss": 4.352426509270425,
      "tokens_seen": 201064448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742627883650953,
      "loss": 1.8957,
      "theoretical_loss": 4.352071906312037,
      "tokens_seen": 201195520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047424272818455366,
      "loss": 1.7722,
      "theoretical_loss": 4.351717598925406,
      "tokens_seen": 201326592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474222668004012,
      "loss": 1.6547,
      "theoretical_loss": 4.351363586671976,
      "tokens_seen": 201457664
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.4540567100048065,
      "objective/train/avg_token_score": 0.013070252723991871,
      "objective/train/docs_used": 124253,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9555007219314575,
      "objective/train/original_loss": 4.570303916931152,
      "objective/train/theoretical_loss": 4.351186691083417,
      "objective/train/tokens_used": 201523200,
      "objective/train/unlikelihood_loss": 0.11967580020427704,
      "theoretical_loss": 4.351186691083417,
      "tokens_seen": 201523200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742026078234704,
      "loss": 1.6592,
      "theoretical_loss": 4.351009869114124,
      "tokens_seen": 201588736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741825476429288,
      "loss": 1.652,
      "theoretical_loss": 4.350656445815164,
      "tokens_seen": 201719808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047416248746238716,
      "loss": 1.6998,
      "theoretical_loss": 4.350303316339337,
      "tokens_seen": 201850880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741424272818455,
      "loss": 1.637,
      "theoretical_loss": 4.349950480251813,
      "tokens_seen": 201981952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741223671013039,
      "loss": 1.443,
      "theoretical_loss": 4.349597937118687,
      "tokens_seen": 202113024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741023069207623,
      "loss": 1.719,
      "theoretical_loss": 4.349245686506976,
      "tokens_seen": 202244096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047408224674022067,
      "loss": 1.6887,
      "theoretical_loss": 4.348893727984619,
      "tokens_seen": 202375168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047406218655967903,
      "loss": 1.7137,
      "theoretical_loss": 4.348542061120469,
      "tokens_seen": 202506240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740421263791374,
      "loss": 1.5756,
      "theoretical_loss": 4.3481906854842975,
      "tokens_seen": 202637312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047402206619859575,
      "loss": 1.8621,
      "theoretical_loss": 4.347839600646786,
      "tokens_seen": 202768384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047400200601805417,
      "loss": 1.6352,
      "theoretical_loss": 4.347488806179528,
      "tokens_seen": 202899456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047398194583751253,
      "loss": 1.7664,
      "theoretical_loss": 4.347138301655021,
      "tokens_seen": 203030528
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.4085410535335541,
      "objective/train/avg_token_score": 0.01795566827058792,
      "objective/train/docs_used": 125838,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5744194984436035,
      "objective/train/original_loss": 3.834841251373291,
      "objective/train/theoretical_loss": 4.346788086646671,
      "objective/train/tokens_used": 203161600,
      "objective/train/unlikelihood_loss": 0.23488780856132507,
      "theoretical_loss": 4.346788086646671,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739618856569709,
      "loss": 1.4374,
      "theoretical_loss": 4.346788086646671,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047394182547642926,
      "loss": 1.6623,
      "theoretical_loss": 4.346438160728785,
      "tokens_seen": 203292672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739217652958877,
      "loss": 1.8131,
      "theoretical_loss": 4.346088523476569,
      "tokens_seen": 203423744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047390170511534604,
      "loss": 1.6644,
      "theoretical_loss": 4.345739174466127,
      "tokens_seen": 203554816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738816449348044,
      "loss": 1.6942,
      "theoretical_loss": 4.3453901132744575,
      "tokens_seen": 203685888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047386158475426276,
      "loss": 1.6128,
      "theoretical_loss": 4.345041339479453,
      "tokens_seen": 203816960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738415245737211,
      "loss": 1.8607,
      "theoretical_loss": 4.344692852659895,
      "tokens_seen": 203948032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047382146439317954,
      "loss": 1.6397,
      "theoretical_loss": 4.344344652395451,
      "tokens_seen": 204079104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738014042126379,
      "loss": 1.5671,
      "theoretical_loss": 4.343996738266677,
      "tokens_seen": 204210176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047378134403209626,
      "loss": 1.6248,
      "theoretical_loss": 4.343649109855009,
      "tokens_seen": 204341248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737612838515546,
      "loss": 1.6748,
      "theoretical_loss": 4.343301766742763,
      "tokens_seen": 204472320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047374122367101304,
      "loss": 1.6671,
      "theoretical_loss": 4.342954708513136,
      "tokens_seen": 204603392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047372116349047146,
      "loss": 1.5574,
      "theoretical_loss": 4.3426079347501965,
      "tokens_seen": 204734464
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.3620479106903076,
      "objective/train/avg_token_score": 0.014170818962156773,
      "objective/train/docs_used": 126453,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4098846912384033,
      "objective/train/original_loss": 3.9272265434265137,
      "objective/train/theoretical_loss": 4.342434654413995,
      "objective/train/tokens_used": 204800000,
      "objective/train/unlikelihood_loss": 0.20976464450359344,
      "theoretical_loss": 4.342434654413995,
      "tokens_seen": 204800000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737011033099298,
      "loss": 1.5437,
      "theoretical_loss": 4.342261445038888,
      "tokens_seen": 204865536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736810431293882,
      "loss": 1.5255,
      "theoretical_loss": 4.341915238965026,
      "tokens_seen": 204996608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736609829488466,
      "loss": 1.7084,
      "theoretical_loss": 4.34156931611529,
      "tokens_seen": 205127680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047364092276830496,
      "loss": 1.6739,
      "theoretical_loss": 4.341223676077232,
      "tokens_seen": 205258752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736208625877633,
      "loss": 1.6936,
      "theoretical_loss": 4.340878318439261,
      "tokens_seen": 205389824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736008024072217,
      "loss": 1.7022,
      "theoretical_loss": 4.3405332427906504,
      "tokens_seen": 205520896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047358074222668005,
      "loss": 1.3979,
      "theoretical_loss": 4.340188448721532,
      "tokens_seen": 205651968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047356068204613847,
      "loss": 1.7449,
      "theoretical_loss": 4.339843935822895,
      "tokens_seen": 205783040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047354062186559683,
      "loss": 1.5221,
      "theoretical_loss": 4.339499703686579,
      "tokens_seen": 205914112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735205616850552,
      "loss": 1.6496,
      "theoretical_loss": 4.339155751905282,
      "tokens_seen": 206045184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047350050150451355,
      "loss": 1.6696,
      "theoretical_loss": 4.338812080072545,
      "tokens_seen": 206176256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047348044132397197,
      "loss": 1.7524,
      "theoretical_loss": 4.3384686877827585,
      "tokens_seen": 206307328
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.37756597995758057,
      "objective/train/avg_token_score": 0.015428772196173668,
      "objective/train/docs_used": 127848,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4948502779006958,
      "objective/train/original_loss": 4.056982040405273,
      "objective/train/theoretical_loss": 4.33812557463116,
      "objective/train/tokens_used": 206438400,
      "objective/train/unlikelihood_loss": 0.15866130590438843,
      "theoretical_loss": 4.33812557463116,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047346038114343033,
      "loss": 1.5575,
      "theoretical_loss": 4.33812557463116,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734403209628887,
      "loss": 1.4098,
      "theoretical_loss": 4.337782740213827,
      "tokens_seen": 206569472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047342026078234706,
      "loss": 1.6327,
      "theoretical_loss": 4.337440184127679,
      "tokens_seen": 206700544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734002006018054,
      "loss": 1.3983,
      "theoretical_loss": 4.337097905970471,
      "tokens_seen": 206831616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047338014042126383,
      "loss": 1.7989,
      "theoretical_loss": 4.336755905340797,
      "tokens_seen": 206962688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733600802407222,
      "loss": 1.6815,
      "theoretical_loss": 4.336414181838082,
      "tokens_seen": 207093760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047334002006018056,
      "loss": 1.6673,
      "theoretical_loss": 4.336072735062583,
      "tokens_seen": 207224832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733199598796389,
      "loss": 1.7324,
      "theoretical_loss": 4.335731564615387,
      "tokens_seen": 207355904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047329989969909734,
      "loss": 1.5223,
      "theoretical_loss": 4.335390670098407,
      "tokens_seen": 207486976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732798395185557,
      "loss": 1.7556,
      "theoretical_loss": 4.335050051114379,
      "tokens_seen": 207618048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047325977933801406,
      "loss": 1.6275,
      "theoretical_loss": 4.334709707266865,
      "tokens_seen": 207749120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732397191574724,
      "loss": 1.7843,
      "theoretical_loss": 4.334369638160242,
      "tokens_seen": 207880192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732196589769308,
      "loss": 1.6139,
      "theoretical_loss": 4.334029843399709,
      "tokens_seen": 208011264
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.4079912006855011,
      "objective/train/avg_token_score": 0.016537239775061607,
      "objective/train/docs_used": 128338,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6732548475265503,
      "objective/train/original_loss": 4.391384124755859,
      "objective/train/theoretical_loss": 4.333860048776074,
      "objective/train/tokens_used": 208076800,
      "objective/train/unlikelihood_loss": 0.11190986633300781,
      "theoretical_loss": 4.333860048776074,
      "tokens_seen": 208076800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731995987963892,
      "loss": 1.8005,
      "theoretical_loss": 4.33369032259128,
      "tokens_seen": 208142336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047317953861584757,
      "loss": 1.4683,
      "theoretical_loss": 4.33335107534178,
      "tokens_seen": 208273408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047315947843530593,
      "loss": 1.5492,
      "theoretical_loss": 4.3330121012588485,
      "tokens_seen": 208404480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731394182547643,
      "loss": 1.6122,
      "theoretical_loss": 4.332673399950932,
      "tokens_seen": 208535552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731193580742227,
      "loss": 1.6955,
      "theoretical_loss": 4.332334971027284,
      "tokens_seen": 208666624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047309929789368107,
      "loss": 1.4314,
      "theoretical_loss": 4.331996814097963,
      "tokens_seen": 208797696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047307923771313943,
      "loss": 1.599,
      "theoretical_loss": 4.331658928773831,
      "tokens_seen": 208928768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730591775325978,
      "loss": 1.7127,
      "theoretical_loss": 4.33132131466655,
      "tokens_seen": 209059840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047303911735205615,
      "loss": 1.5623,
      "theoretical_loss": 4.330983971388578,
      "tokens_seen": 209190912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047301905717151457,
      "loss": 1.4685,
      "theoretical_loss": 4.330646898553173,
      "tokens_seen": 209321984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047299899699097293,
      "loss": 1.5697,
      "theoretical_loss": 4.330310095774383,
      "tokens_seen": 209453056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729789368104313,
      "loss": 1.652,
      "theoretical_loss": 4.329973562667053,
      "tokens_seen": 209584128
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.45869991183280945,
      "objective/train/avg_token_score": 0.025695886462926865,
      "objective/train/docs_used": 129483,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.688183307647705,
      "objective/train/original_loss": 4.105484485626221,
      "objective/train/theoretical_loss": 4.329637298846812,
      "objective/train/tokens_used": 209715200,
      "objective/train/unlikelihood_loss": 0.14652951061725616,
      "theoretical_loss": 4.329637298846812,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047295887662988966,
      "loss": 1.675,
      "theoretical_loss": 4.329637298846812,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729388164493481,
      "loss": 1.5683,
      "theoretical_loss": 4.32930130393008,
      "tokens_seen": 209846272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047291875626880644,
      "loss": 1.6511,
      "theoretical_loss": 4.3289655775340625,
      "tokens_seen": 209977344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728986960882648,
      "loss": 1.5679,
      "theoretical_loss": 4.328630119276747,
      "tokens_seen": 210108416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047287863590772316,
      "loss": 1.6666,
      "theoretical_loss": 4.328294928776903,
      "tokens_seen": 210239488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728585757271815,
      "loss": 1.6358,
      "theoretical_loss": 4.327960005654081,
      "tokens_seen": 210370560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047283851554663994,
      "loss": 1.5076,
      "theoretical_loss": 4.327625349528605,
      "tokens_seen": 210501632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728184553660983,
      "loss": 1.5038,
      "theoretical_loss": 4.3272909600215765,
      "tokens_seen": 210632704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047279839518555666,
      "loss": 1.7096,
      "theoretical_loss": 4.326956836754871,
      "tokens_seen": 210763776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472778335005015,
      "loss": 1.4226,
      "theoretical_loss": 4.326622979351132,
      "tokens_seen": 210894848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047275827482447344,
      "loss": 1.5037,
      "theoretical_loss": 4.326289387433776,
      "tokens_seen": 211025920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727382146439318,
      "loss": 1.7344,
      "theoretical_loss": 4.325956060626982,
      "tokens_seen": 211156992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047271815446339017,
      "loss": 1.7053,
      "theoretical_loss": 4.325622998555697,
      "tokens_seen": 211288064
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.433956503868103,
      "objective/train/avg_token_score": 0.010776055976748466,
      "objective/train/docs_used": 130263,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.608245849609375,
      "objective/train/original_loss": 4.053385257720947,
      "objective/train/theoretical_loss": 4.3254565666788745,
      "objective/train/tokens_used": 211353600,
      "objective/train/unlikelihood_loss": 0.1744779348373413,
      "theoretical_loss": 4.3254565666788745,
      "tokens_seen": 211353600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047269809428284853,
      "loss": 1.6731,
      "theoretical_loss": 4.325290200845629,
      "tokens_seen": 211419136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047267803410230695,
      "loss": 1.7337,
      "theoretical_loss": 4.324957667123249,
      "tokens_seen": 211550208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726579739217653,
      "loss": 1.7319,
      "theoretical_loss": 4.324625397015783,
      "tokens_seen": 211681280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047263791374122367,
      "loss": 1.6882,
      "theoretical_loss": 4.324293390151218,
      "tokens_seen": 211812352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047261785356068203,
      "loss": 1.5198,
      "theoretical_loss": 4.323961646158294,
      "tokens_seen": 211943424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725977933801404,
      "loss": 1.5977,
      "theoretical_loss": 4.323630164666502,
      "tokens_seen": 212074496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725777331995988,
      "loss": 1.6756,
      "theoretical_loss": 4.323298945306089,
      "tokens_seen": 212205568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725576730190572,
      "loss": 1.6358,
      "theoretical_loss": 4.322967987708043,
      "tokens_seen": 212336640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047253761283851554,
      "loss": 1.6253,
      "theoretical_loss": 4.322637291504106,
      "tokens_seen": 212467712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725175526579739,
      "loss": 1.5415,
      "theoretical_loss": 4.322306856326761,
      "tokens_seen": 212598784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724974924774323,
      "loss": 1.7128,
      "theoretical_loss": 4.321976681809236,
      "tokens_seen": 212729856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724774322968907,
      "loss": 1.658,
      "theoretical_loss": 4.321646767585497,
      "tokens_seen": 212860928
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.32704055309295654,
      "objective/train/avg_token_score": 0.014311397448182106,
      "objective/train/docs_used": 131630,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.350407600402832,
      "objective/train/original_loss": 4.261425971984863,
      "objective/train/theoretical_loss": 4.321317113290252,
      "objective/train/tokens_used": 212992000,
      "objective/train/unlikelihood_loss": 0.19203782081604004,
      "theoretical_loss": 4.321317113290252,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047245737211634904,
      "loss": 1.4489,
      "theoretical_loss": 4.321317113290252,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724373119358074,
      "loss": 1.6253,
      "theoretical_loss": 4.320987718558945,
      "tokens_seen": 213123072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047241725175526576,
      "loss": 1.6789,
      "theoretical_loss": 4.320658583027755,
      "tokens_seen": 213254144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723971915747242,
      "loss": 1.4694,
      "theoretical_loss": 4.3203297063335935,
      "tokens_seen": 213385216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047237713139418254,
      "loss": 1.515,
      "theoretical_loss": 4.320001088114105,
      "tokens_seen": 213516288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723570712136409,
      "loss": 1.7089,
      "theoretical_loss": 4.31967272800766,
      "tokens_seen": 213647360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047233701103309927,
      "loss": 1.527,
      "theoretical_loss": 4.319344625653361,
      "tokens_seen": 213778432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723169508525577,
      "loss": 1.7138,
      "theoretical_loss": 4.319016780691033,
      "tokens_seen": 213909504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047229689067201605,
      "loss": 1.7267,
      "theoretical_loss": 4.318689192761225,
      "tokens_seen": 214040576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722768304914744,
      "loss": 1.6687,
      "theoretical_loss": 4.318361861505207,
      "tokens_seen": 214171648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047225677031093277,
      "loss": 1.6014,
      "theoretical_loss": 4.318034786564971,
      "tokens_seen": 214302720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047223671013039113,
      "loss": 1.5851,
      "theoretical_loss": 4.3177079675832255,
      "tokens_seen": 214433792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004722166499498496,
      "loss": 1.3161,
      "theoretical_loss": 4.3173814042033944,
      "tokens_seen": 214564864
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.4389662742614746,
      "objective/train/avg_token_score": 0.023373806849122047,
      "objective/train/docs_used": 132228,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6921892166137695,
      "objective/train/original_loss": 4.019752502441406,
      "objective/train/theoretical_loss": 4.317218218252963,
      "objective/train/tokens_used": 214630400,
      "objective/train/unlikelihood_loss": 0.20437592267990112,
      "theoretical_loss": 4.317218218252963,
      "tokens_seen": 214630400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047219658976930797,
      "loss": 1.6002,
      "theoretical_loss": 4.317055096069618,
      "tokens_seen": 214695936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047217652958876633,
      "loss": 1.6065,
      "theoretical_loss": 4.316729042826745,
      "tokens_seen": 214827008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721564694082247,
      "loss": 1.7483,
      "theoretical_loss": 4.316403244120339,
      "tokens_seen": 214958080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721364092276831,
      "loss": 1.631,
      "theoretical_loss": 4.316077699596671,
      "tokens_seen": 215089152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047211634904714147,
      "loss": 1.8237,
      "theoretical_loss": 4.315752408902716,
      "tokens_seen": 215220224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047209628886659983,
      "loss": 1.6567,
      "theoretical_loss": 4.315427371686157,
      "tokens_seen": 215351296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720762286860582,
      "loss": 1.6742,
      "theoretical_loss": 4.315102587595379,
      "tokens_seen": 215482368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047205616850551656,
      "loss": 1.7232,
      "theoretical_loss": 4.314778056279468,
      "tokens_seen": 215613440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047203610832497497,
      "loss": 1.5351,
      "theoretical_loss": 4.314453777388209,
      "tokens_seen": 215744512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047201604814443333,
      "loss": 1.5639,
      "theoretical_loss": 4.314129750572087,
      "tokens_seen": 215875584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719959879638917,
      "loss": 1.4961,
      "theoretical_loss": 4.313805975482278,
      "tokens_seen": 216006656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047197592778335006,
      "loss": 1.7063,
      "theoretical_loss": 4.313482451770659,
      "tokens_seen": 216137728
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.47482892870903015,
      "objective/train/avg_token_score": 0.027175776660442352,
      "objective/train/docs_used": 133529,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9989675283432007,
      "objective/train/original_loss": 4.329239845275879,
      "objective/train/theoretical_loss": 4.3131591790897925,
      "objective/train/tokens_used": 216268800,
      "objective/train/unlikelihood_loss": 0.155557319521904,
      "theoretical_loss": 4.3131591790897925,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719558676028085,
      "loss": 1.4553,
      "theoretical_loss": 4.3131591790897925,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047193580742226684,
      "loss": 1.6868,
      "theoretical_loss": 4.312836157092934,
      "tokens_seen": 216399872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719157472417252,
      "loss": 1.4115,
      "theoretical_loss": 4.31251338543403,
      "tokens_seen": 216530944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047189568706118356,
      "loss": 1.6705,
      "theoretical_loss": 4.312190863767708,
      "tokens_seen": 216662016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718756268806419,
      "loss": 1.6034,
      "theoretical_loss": 4.311868591749287,
      "tokens_seen": 216793088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047185556670010034,
      "loss": 1.7697,
      "theoretical_loss": 4.311546569034767,
      "tokens_seen": 216924160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718355065195587,
      "loss": 1.6095,
      "theoretical_loss": 4.311224795280825,
      "tokens_seen": 217055232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047181544633901706,
      "loss": 1.7137,
      "theoretical_loss": 4.310903270144825,
      "tokens_seen": 217186304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717953861584754,
      "loss": 1.5225,
      "theoretical_loss": 4.310581993284805,
      "tokens_seen": 217317376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047177532597793384,
      "loss": 1.8289,
      "theoretical_loss": 4.310260964359479,
      "tokens_seen": 217448448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717552657973922,
      "loss": 1.6959,
      "theoretical_loss": 4.309940183028236,
      "tokens_seen": 217579520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047173520561685057,
      "loss": 1.604,
      "theoretical_loss": 4.309619648951139,
      "tokens_seen": 217710592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047171514543630893,
      "loss": 1.6158,
      "theoretical_loss": 4.30929936178892,
      "tokens_seen": 217841664
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.5201612710952759,
      "objective/train/avg_token_score": 0.025473540648818016,
      "objective/train/docs_used": 134228,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.066497325897217,
      "objective/train/original_loss": 3.987381935119629,
      "objective/train/theoretical_loss": 4.309139310695053,
      "objective/train/tokens_used": 217907200,
      "objective/train/unlikelihood_loss": 0.17281749844551086,
      "theoretical_loss": 4.309139310695053,
      "tokens_seen": 217907200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047169508525576735,
      "loss": 1.7519,
      "theoretical_loss": 4.308979321202983,
      "tokens_seen": 217972736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716750250752257,
      "loss": 1.7713,
      "theoretical_loss": 4.308659526855396,
      "tokens_seen": 218103808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047165496489468407,
      "loss": 1.6655,
      "theoretical_loss": 4.308339978408897,
      "tokens_seen": 218234880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047163490471414243,
      "loss": 1.7458,
      "theoretical_loss": 4.308020675526883,
      "tokens_seen": 218365952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716148445336008,
      "loss": 1.6181,
      "theoretical_loss": 4.30770161787342,
      "tokens_seen": 218497024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715947843530592,
      "loss": 1.5035,
      "theoretical_loss": 4.307382805113228,
      "tokens_seen": 218628096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715747241725176,
      "loss": 1.616,
      "theoretical_loss": 4.307064236911692,
      "tokens_seen": 218759168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047155466399197594,
      "loss": 1.806,
      "theoretical_loss": 4.306745912934849,
      "tokens_seen": 218890240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715346038114343,
      "loss": 1.6041,
      "theoretical_loss": 4.306427832849394,
      "tokens_seen": 219021312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715145436308927,
      "loss": 1.8418,
      "theoretical_loss": 4.306109996322679,
      "tokens_seen": 219152384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714944834503511,
      "loss": 1.5547,
      "theoretical_loss": 4.305792403022703,
      "tokens_seen": 219283456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047147442326980944,
      "loss": 1.2979,
      "theoretical_loss": 4.305475052618119,
      "tokens_seen": 219414528
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.4022482931613922,
      "objective/train/avg_token_score": 0.006948302499949932,
      "objective/train/docs_used": 135676,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5557243824005127,
      "objective/train/original_loss": 3.7284510135650635,
      "objective/train/theoretical_loss": 4.305157944778228,
      "objective/train/tokens_used": 219545600,
      "objective/train/unlikelihood_loss": 0.15928247570991516,
      "theoretical_loss": 4.305157944778228,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714543630892678,
      "loss": 1.86,
      "theoretical_loss": 4.305157944778228,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047143430290872616,
      "loss": 1.7673,
      "theoretical_loss": 4.304841079172979,
      "tokens_seen": 219676672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714142427281846,
      "loss": 1.6245,
      "theoretical_loss": 4.304524455472965,
      "tokens_seen": 219807744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047139418254764294,
      "loss": 1.8468,
      "theoretical_loss": 4.304208073349426,
      "tokens_seen": 219938816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713741223671013,
      "loss": 1.5617,
      "theoretical_loss": 4.30389193247424,
      "tokens_seen": 220069888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047135406218655967,
      "loss": 1.7598,
      "theoretical_loss": 4.303576032519931,
      "tokens_seen": 220200960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713340020060181,
      "loss": 1.7091,
      "theoretical_loss": 4.303260373159659,
      "tokens_seen": 220332032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047131394182547645,
      "loss": 1.7758,
      "theoretical_loss": 4.30294495406722,
      "tokens_seen": 220463104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712938816449348,
      "loss": 1.6308,
      "theoretical_loss": 4.302629774917049,
      "tokens_seen": 220594176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047127382146439317,
      "loss": 1.6361,
      "theoretical_loss": 4.302314835384214,
      "tokens_seen": 220725248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047125376128385153,
      "loss": 1.6959,
      "theoretical_loss": 4.302000135144416,
      "tokens_seen": 220856320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047123370110330995,
      "loss": 1.7791,
      "theoretical_loss": 4.301685673873987,
      "tokens_seen": 220987392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712136409227683,
      "loss": 1.6592,
      "theoretical_loss": 4.301371451249888,
      "tokens_seen": 221118464
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.4833821952342987,
      "objective/train/avg_token_score": 0.009588075801730156,
      "objective/train/docs_used": 136371,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.763645887374878,
      "objective/train/original_loss": 3.7894089221954346,
      "objective/train/theoretical_loss": 4.301214429329433,
      "objective/train/tokens_used": 221184000,
      "objective/train/unlikelihood_loss": 0.19660082459449768,
      "theoretical_loss": 4.301214429329433,
      "tokens_seen": 221184000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711935807422267,
      "loss": 1.6616,
      "theoretical_loss": 4.301057466949707,
      "tokens_seen": 221249536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047117352056168504,
      "loss": 1.7754,
      "theoretical_loss": 4.30074372065166,
      "tokens_seen": 221380608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047115346038114345,
      "loss": 1.5822,
      "theoretical_loss": 4.300430212034587,
      "tokens_seen": 221511680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711334002006018,
      "loss": 1.7638,
      "theoretical_loss": 4.300116940777951,
      "tokens_seen": 221642752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711133400200602,
      "loss": 1.8272,
      "theoretical_loss": 4.299803906561835,
      "tokens_seen": 221773824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047109327983951854,
      "loss": 1.6185,
      "theoretical_loss": 4.299491109066947,
      "tokens_seen": 221904896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710732196589769,
      "loss": 1.6955,
      "theoretical_loss": 4.2991785479746065,
      "tokens_seen": 222035968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710531594784353,
      "loss": 1.7621,
      "theoretical_loss": 4.298866222966755,
      "tokens_seen": 222167040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710330992978937,
      "loss": 1.5053,
      "theoretical_loss": 4.298554133725946,
      "tokens_seen": 222298112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047101303911735204,
      "loss": 1.7638,
      "theoretical_loss": 4.298242279935349,
      "tokens_seen": 222429184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709929789368104,
      "loss": 1.7638,
      "theoretical_loss": 4.297930661278745,
      "tokens_seen": 222560256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709729187562688,
      "loss": 1.6024,
      "theoretical_loss": 4.297619277440523,
      "tokens_seen": 222691328
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.4119012653827667,
      "objective/train/avg_token_score": 0.012975508347153664,
      "objective/train/docs_used": 137650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5277689695358276,
      "objective/train/original_loss": 4.178741931915283,
      "objective/train/theoretical_loss": 4.297308128105687,
      "objective/train/tokens_used": 222822400,
      "objective/train/unlikelihood_loss": 0.15160509943962097,
      "theoretical_loss": 4.297308128105687,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709528585757272,
      "loss": 1.4614,
      "theoretical_loss": 4.297308128105687,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047093279839518555,
      "loss": 1.6429,
      "theoretical_loss": 4.296997212959842,
      "tokens_seen": 222953472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709127382146439,
      "loss": 1.6713,
      "theoretical_loss": 4.296686531689204,
      "tokens_seen": 223084544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047089267803410227,
      "loss": 1.7281,
      "theoretical_loss": 4.296376083980589,
      "tokens_seen": 223215616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708726178535607,
      "loss": 1.7329,
      "theoretical_loss": 4.296065869521421,
      "tokens_seen": 223346688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047085255767301905,
      "loss": 1.6212,
      "theoretical_loss": 4.2957558879997215,
      "tokens_seen": 223477760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708324974924774,
      "loss": 1.7222,
      "theoretical_loss": 4.295446139104112,
      "tokens_seen": 223608832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708124373119358,
      "loss": 1.5936,
      "theoretical_loss": 4.295136622523817,
      "tokens_seen": 223739904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707923771313942,
      "loss": 1.5582,
      "theoretical_loss": 4.294827337948651,
      "tokens_seen": 223870976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047077231695085255,
      "loss": 1.621,
      "theoretical_loss": 4.29451828506903,
      "tokens_seen": 224002048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707522567703109,
      "loss": 1.7204,
      "theoretical_loss": 4.29420946357596,
      "tokens_seen": 224133120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707321965897693,
      "loss": 1.6165,
      "theoretical_loss": 4.293900873161043,
      "tokens_seen": 224264192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047071213640922775,
      "loss": 1.6774,
      "theoretical_loss": 4.293592513516469,
      "tokens_seen": 224395264
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.3668743968009949,
      "objective/train/avg_token_score": 0.00708651402965188,
      "objective/train/docs_used": 138458,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4711482524871826,
      "objective/train/original_loss": 4.133864879608154,
      "objective/train/theoretical_loss": 4.293438420137031,
      "objective/train/tokens_used": 224460800,
      "objective/train/unlikelihood_loss": 0.21583233773708344,
      "theoretical_loss": 4.293438420137031,
      "tokens_seen": 224460800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706920762286861,
      "loss": 1.4837,
      "theoretical_loss": 4.293284384335017,
      "tokens_seen": 224526336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047067201604814447,
      "loss": 1.646,
      "theoretical_loss": 4.292976485310057,
      "tokens_seen": 224657408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047065195586760283,
      "loss": 1.4897,
      "theoretical_loss": 4.292668816135545,
      "tokens_seen": 224788480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706318956870612,
      "loss": 1.6201,
      "theoretical_loss": 4.29236137650602,
      "tokens_seen": 224919552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706118355065196,
      "loss": 1.7653,
      "theoretical_loss": 4.292054166116605,
      "tokens_seen": 225050624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470591775325978,
      "loss": 1.5181,
      "theoretical_loss": 4.2917471846630075,
      "tokens_seen": 225181696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047057171514543634,
      "loss": 1.5244,
      "theoretical_loss": 4.291440431841513,
      "tokens_seen": 225312768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705516549648947,
      "loss": 1.3208,
      "theoretical_loss": 4.291133907348989,
      "tokens_seen": 225443840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705315947843531,
      "loss": 1.6231,
      "theoretical_loss": 4.29082761088288,
      "tokens_seen": 225574912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705115346038115,
      "loss": 1.4922,
      "theoretical_loss": 4.290521542141203,
      "tokens_seen": 225705984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047049147442326984,
      "loss": 1.63,
      "theoretical_loss": 4.290215700822556,
      "tokens_seen": 225837056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704714142427282,
      "loss": 1.6671,
      "theoretical_loss": 4.289910086626108,
      "tokens_seen": 225968128
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.519305944442749,
      "objective/train/avg_token_score": 0.011661575175821781,
      "objective/train/docs_used": 139641,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.168332815170288,
      "objective/train/original_loss": 4.076672077178955,
      "objective/train/theoretical_loss": 4.2896046992515995,
      "objective/train/tokens_used": 226099200,
      "objective/train/unlikelihood_loss": 0.17042392492294312,
      "theoretical_loss": 4.2896046992515995,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047045135406218656,
      "loss": 1.5969,
      "theoretical_loss": 4.2896046992515995,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470431293881645,
      "loss": 1.5568,
      "theoretical_loss": 4.289299538399341,
      "tokens_seen": 226230272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047041123370110334,
      "loss": 1.654,
      "theoretical_loss": 4.288994603770215,
      "tokens_seen": 226361344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703911735205617,
      "loss": 1.651,
      "theoretical_loss": 4.28868989506567,
      "tokens_seen": 226492416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047037111334002007,
      "loss": 1.8949,
      "theoretical_loss": 4.288385411987722,
      "tokens_seen": 226623488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703510531594785,
      "loss": 1.5494,
      "theoretical_loss": 4.2880811542389505,
      "tokens_seen": 226754560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047033099297893685,
      "loss": 1.4705,
      "theoretical_loss": 4.287777121522501,
      "tokens_seen": 226885632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703109327983952,
      "loss": 1.6097,
      "theoretical_loss": 4.287473313542077,
      "tokens_seen": 227016704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047029087261785357,
      "loss": 1.7498,
      "theoretical_loss": 4.287169730001949,
      "tokens_seen": 227147776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047027081243731193,
      "loss": 1.7013,
      "theoretical_loss": 4.286866370606943,
      "tokens_seen": 227278848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047025075225677035,
      "loss": 1.8046,
      "theoretical_loss": 4.286563235062444,
      "tokens_seen": 227409920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702306920762287,
      "loss": 1.6791,
      "theoretical_loss": 4.286260323074394,
      "tokens_seen": 227540992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702106318956871,
      "loss": 1.578,
      "theoretical_loss": 4.285957634349289,
      "tokens_seen": 227672064
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.46957477927207947,
      "objective/train/avg_token_score": 0.018808742985129356,
      "objective/train/docs_used": 140376,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8826521635055542,
      "objective/train/original_loss": 4.291516304016113,
      "objective/train/theoretical_loss": 4.285806373618774,
      "objective/train/tokens_used": 227737600,
      "objective/train/unlikelihood_loss": 0.15742744505405426,
      "theoretical_loss": 4.285806373618774,
      "tokens_seen": 227737600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047019057171514544,
      "loss": 1.8432,
      "theoretical_loss": 4.285655168594182,
      "tokens_seen": 227803136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047017051153460385,
      "loss": 1.539,
      "theoretical_loss": 4.285352925516676,
      "tokens_seen": 227934208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701504513540622,
      "loss": 1.6382,
      "theoretical_loss": 4.285050904824925,
      "tokens_seen": 228065280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701303911735206,
      "loss": 1.5511,
      "theoretical_loss": 4.284749106227636,
      "tokens_seen": 228196352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047011033099297894,
      "loss": 1.6545,
      "theoretical_loss": 4.284447529434061,
      "tokens_seen": 228327424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700902708124373,
      "loss": 1.6648,
      "theoretical_loss": 4.284146174154003,
      "tokens_seen": 228458496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700702106318957,
      "loss": 1.5866,
      "theoretical_loss": 4.283845040097807,
      "tokens_seen": 228589568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700501504513541,
      "loss": 1.6361,
      "theoretical_loss": 4.2835441269763646,
      "tokens_seen": 228720640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047003009027081244,
      "loss": 1.6915,
      "theoretical_loss": 4.283243434501112,
      "tokens_seen": 228851712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700100300902708,
      "loss": 1.6013,
      "theoretical_loss": 4.282942962384023,
      "tokens_seen": 228982784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699899699097292,
      "loss": 1.6088,
      "theoretical_loss": 4.282642710337618,
      "tokens_seen": 229113856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699699097291876,
      "loss": 1.7422,
      "theoretical_loss": 4.282342678074951,
      "tokens_seen": 229244928
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.34762951731681824,
      "objective/train/avg_token_score": 0.020571375265717506,
      "objective/train/docs_used": 141693,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.228954553604126,
      "objective/train/original_loss": 4.2740325927734375,
      "objective/train/theoretical_loss": 4.282042865309616,
      "objective/train/tokens_used": 229376000,
      "objective/train/unlikelihood_loss": 0.13970565795898438,
      "theoretical_loss": 4.282042865309616,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046994984954864595,
      "loss": 1.5481,
      "theoretical_loss": 4.282042865309616,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699297893681043,
      "loss": 1.7614,
      "theoretical_loss": 4.2817432717557455,
      "tokens_seen": 229507072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046990972918756267,
      "loss": 1.4883,
      "theoretical_loss": 4.281443897128004,
      "tokens_seen": 229638144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698896690070211,
      "loss": 1.5377,
      "theoretical_loss": 4.281144741141593,
      "tokens_seen": 229769216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046986960882647945,
      "loss": 1.6032,
      "theoretical_loss": 4.2808458035122445,
      "tokens_seen": 229900288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698495486459378,
      "loss": 1.6431,
      "theoretical_loss": 4.280547083956224,
      "tokens_seen": 230031360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698294884653962,
      "loss": 1.7657,
      "theoretical_loss": 4.280248582190324,
      "tokens_seen": 230162432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698094282848546,
      "loss": 1.6122,
      "theoretical_loss": 4.279950297931869,
      "tokens_seen": 230293504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046978936810431295,
      "loss": 1.5755,
      "theoretical_loss": 4.279652230898709,
      "tokens_seen": 230424576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697693079237713,
      "loss": 1.557,
      "theoretical_loss": 4.27935438080922,
      "tokens_seen": 230555648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697492477432297,
      "loss": 1.722,
      "theoretical_loss": 4.279056747382306,
      "tokens_seen": 230686720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046972918756268804,
      "loss": 1.5124,
      "theoretical_loss": 4.27875933033739,
      "tokens_seen": 230817792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046970912738214646,
      "loss": 1.7483,
      "theoretical_loss": 4.278462129394423,
      "tokens_seen": 230948864
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.44086021184921265,
      "objective/train/avg_token_score": 0.02034401148557663,
      "objective/train/docs_used": 142407,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9175533056259155,
      "objective/train/original_loss": 4.461835861206055,
      "objective/train/theoretical_loss": 4.278313609873795,
      "objective/train/tokens_used": 231014400,
      "objective/train/unlikelihood_loss": 0.13161636888980865,
      "theoretical_loss": 4.278313609873795,
      "tokens_seen": 231014400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696890672016048,
      "loss": 1.6861,
      "theoretical_loss": 4.278165144273871,
      "tokens_seen": 231079936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696690070210632,
      "loss": 1.6634,
      "theoretical_loss": 4.277868374696725,
      "tokens_seen": 231211008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046964894684052154,
      "loss": 1.7192,
      "theoretical_loss": 4.277571820384491,
      "tokens_seen": 231342080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046962888665997996,
      "loss": 1.5473,
      "theoretical_loss": 4.277275481059195,
      "tokens_seen": 231473152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696088264794383,
      "loss": 1.6103,
      "theoretical_loss": 4.276979356443377,
      "tokens_seen": 231604224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695887662988967,
      "loss": 1.6004,
      "theoretical_loss": 4.276683446260093,
      "tokens_seen": 231735296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046956870611835505,
      "loss": 1.5718,
      "theoretical_loss": 4.276387750232913,
      "tokens_seen": 231866368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046954864593781346,
      "loss": 1.4526,
      "theoretical_loss": 4.276092268085918,
      "tokens_seen": 231997440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695285857572718,
      "loss": 1.7252,
      "theoretical_loss": 4.275796999543703,
      "tokens_seen": 232128512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695085255767302,
      "loss": 1.6623,
      "theoretical_loss": 4.275501944331367,
      "tokens_seen": 232259584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046948846539618855,
      "loss": 1.6335,
      "theoretical_loss": 4.275207102174525,
      "tokens_seen": 232390656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694684052156469,
      "loss": 1.9984,
      "theoretical_loss": 4.2749124727992935,
      "tokens_seen": 232521728
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.44855815172195435,
      "objective/train/avg_token_score": 0.007418564986437559,
      "objective/train/docs_used": 143093,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7657521963119507,
      "objective/train/original_loss": 3.696604013442993,
      "objective/train/theoretical_loss": 4.274618055932298,
      "objective/train/tokens_used": 232652800,
      "objective/train/unlikelihood_loss": 0.1868204027414322,
      "theoretical_loss": 4.274618055932298,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694483450351053,
      "loss": 1.7135,
      "theoretical_loss": 4.274618055932298,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694282848545637,
      "loss": 1.6273,
      "theoretical_loss": 4.27432385130067,
      "tokens_seen": 232783872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046940822467402205,
      "loss": 1.6464,
      "theoretical_loss": 4.2740298586320415,
      "tokens_seen": 232914944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693881644934804,
      "loss": 1.6506,
      "theoretical_loss": 4.27373607765455,
      "tokens_seen": 233046016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046936810431293883,
      "loss": 1.5552,
      "theoretical_loss": 4.273442508096833,
      "tokens_seen": 233177088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693480441323972,
      "loss": 1.6676,
      "theoretical_loss": 4.273149149688028,
      "tokens_seen": 233308160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046932798395185555,
      "loss": 1.9249,
      "theoretical_loss": 4.272856002157772,
      "tokens_seen": 233439232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693079237713139,
      "loss": 1.7331,
      "theoretical_loss": 4.2725630652362,
      "tokens_seen": 233570304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692878635907723,
      "loss": 1.6774,
      "theoretical_loss": 4.272270338653942,
      "tokens_seen": 233701376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692678034102307,
      "loss": 1.5886,
      "theoretical_loss": 4.271977822142125,
      "tokens_seen": 233832448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046924774322968906,
      "loss": 1.6229,
      "theoretical_loss": 4.27168551543237,
      "tokens_seen": 233963520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692276830491474,
      "loss": 1.6078,
      "theoretical_loss": 4.27139341825679,
      "tokens_seen": 234094592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046920762286860584,
      "loss": 1.672,
      "theoretical_loss": 4.27110153034799,
      "tokens_seen": 234225664
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.4761730134487152,
      "objective/train/avg_token_score": 0.01180332899093628,
      "objective/train/docs_used": 144323,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8356056213378906,
      "objective/train/original_loss": 4.213988780975342,
      "objective/train/theoretical_loss": 4.270955664785207,
      "objective/train/tokens_used": 234291200,
      "objective/train/unlikelihood_loss": 0.13796842098236084,
      "theoretical_loss": 4.270955664785207,
      "tokens_seen": 234291200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046918756268806425,
      "loss": 1.6133,
      "theoretical_loss": 4.2708098514390676,
      "tokens_seen": 234356736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691675025075226,
      "loss": 1.7033,
      "theoretical_loss": 4.2705183812636065,
      "tokens_seen": 234487808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469147442326981,
      "loss": 1.7059,
      "theoretical_loss": 4.270227119555681,
      "tokens_seen": 234618880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046912738214643934,
      "loss": 1.7052,
      "theoretical_loss": 4.269936066049852,
      "tokens_seen": 234749952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691073219658977,
      "loss": 1.4875,
      "theoretical_loss": 4.269645220481166,
      "tokens_seen": 234881024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690872617853561,
      "loss": 1.7967,
      "theoretical_loss": 4.269354582585156,
      "tokens_seen": 235012096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690672016048145,
      "loss": 1.5462,
      "theoretical_loss": 4.269064152097835,
      "tokens_seen": 235143168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046904714142427284,
      "loss": 1.642,
      "theoretical_loss": 4.268773928755701,
      "tokens_seen": 235274240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690270812437312,
      "loss": 1.7563,
      "theoretical_loss": 4.268483912295735,
      "tokens_seen": 235405312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690070210631896,
      "loss": 1.7788,
      "theoretical_loss": 4.268194102455395,
      "tokens_seen": 235536384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468986960882648,
      "loss": 1.758,
      "theoretical_loss": 4.267904498972618,
      "tokens_seen": 235667456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046896690070210635,
      "loss": 1.5842,
      "theoretical_loss": 4.267615101585821,
      "tokens_seen": 235798528
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.48765885829925537,
      "objective/train/avg_token_score": 0.008590947836637497,
      "objective/train/docs_used": 145084,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9980686902999878,
      "objective/train/original_loss": 4.121566295623779,
      "objective/train/theoretical_loss": 4.267325910033897,
      "objective/train/tokens_used": 235929600,
      "objective/train/unlikelihood_loss": 0.1698838621377945,
      "theoretical_loss": 4.267325910033897,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689468405215647,
      "loss": 1.6593,
      "theoretical_loss": 4.267325910033897,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046892678034102307,
      "loss": 1.6164,
      "theoretical_loss": 4.267036924056215,
      "tokens_seen": 236060672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689067201604815,
      "loss": 1.5894,
      "theoretical_loss": 4.266748143392617,
      "tokens_seen": 236191744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046888665997993985,
      "loss": 1.6229,
      "theoretical_loss": 4.26645956778342,
      "tokens_seen": 236322816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688665997993982,
      "loss": 1.805,
      "theoretical_loss": 4.266171196969412,
      "tokens_seen": 236453888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688465396188566,
      "loss": 1.7046,
      "theoretical_loss": 4.265883030691853,
      "tokens_seen": 236584960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468826479438315,
      "loss": 1.6652,
      "theoretical_loss": 4.265595068692473,
      "tokens_seen": 236716032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046880641925777335,
      "loss": 1.7922,
      "theoretical_loss": 4.265307310713471,
      "tokens_seen": 236847104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687863590772317,
      "loss": 1.5168,
      "theoretical_loss": 4.265019756497512,
      "tokens_seen": 236978176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687662988966901,
      "loss": 1.7535,
      "theoretical_loss": 4.264732405787731,
      "tokens_seen": 237109248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046874623871614844,
      "loss": 1.6919,
      "theoretical_loss": 4.264445258327724,
      "tokens_seen": 237240320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046872617853560686,
      "loss": 1.7658,
      "theoretical_loss": 4.264158313861557,
      "tokens_seen": 237371392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687061183550652,
      "loss": 1.6452,
      "theoretical_loss": 4.2638715721337554,
      "tokens_seen": 237502464
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.4448924660682678,
      "objective/train/avg_token_score": 0.019487597048282623,
      "objective/train/docs_used": 146518,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6499921083450317,
      "objective/train/original_loss": 4.144484519958496,
      "objective/train/theoretical_loss": 4.263728277217032,
      "objective/train/tokens_used": 237568000,
      "objective/train/unlikelihood_loss": 0.14790794253349304,
      "theoretical_loss": 4.263728277217032,
      "tokens_seen": 237568000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686860581745236,
      "loss": 1.6665,
      "theoretical_loss": 4.263585032889306,
      "tokens_seen": 237633536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046866599799398194,
      "loss": 1.5501,
      "theoretical_loss": 4.2632986958736625,
      "tokens_seen": 237764608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046864593781344036,
      "loss": 1.788,
      "theoretical_loss": 4.263012560832733,
      "tokens_seen": 237895680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686258776328987,
      "loss": 1.2767,
      "theoretical_loss": 4.262726627512886,
      "tokens_seen": 238026752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686058174523571,
      "loss": 1.7164,
      "theoretical_loss": 4.26244089566095,
      "tokens_seen": 238157824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046858575727181545,
      "loss": 1.6782,
      "theoretical_loss": 4.262155365024207,
      "tokens_seen": 238288896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685656970912738,
      "loss": 1.6008,
      "theoretical_loss": 4.261870035350399,
      "tokens_seen": 238419968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685456369107322,
      "loss": 1.5916,
      "theoretical_loss": 4.26158490638772,
      "tokens_seen": 238551040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685255767301906,
      "loss": 1.6809,
      "theoretical_loss": 4.261299977884816,
      "tokens_seen": 238682112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046850551654964895,
      "loss": 1.5857,
      "theoretical_loss": 4.261015249590789,
      "tokens_seen": 238813184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684854563691073,
      "loss": 1.5282,
      "theoretical_loss": 4.260730721255191,
      "tokens_seen": 238944256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046846539618856573,
      "loss": 1.7735,
      "theoretical_loss": 4.260446392628026,
      "tokens_seen": 239075328
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.43187928199768066,
      "objective/train/avg_token_score": 0.025535602122545242,
      "objective/train/docs_used": 147039,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7066162824630737,
      "objective/train/original_loss": 4.2467427253723145,
      "objective/train/theoretical_loss": 4.260162263459744,
      "objective/train/tokens_used": 239206400,
      "objective/train/unlikelihood_loss": 0.11330122500658035,
      "theoretical_loss": 4.260162263459744,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684453360080241,
      "loss": 1.475,
      "theoretical_loss": 4.260162263459744,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046842527582748245,
      "loss": 1.5486,
      "theoretical_loss": 4.259878333501247,
      "tokens_seen": 239337472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684052156469408,
      "loss": 1.4455,
      "theoretical_loss": 4.259594602503881,
      "tokens_seen": 239468544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046838515546639923,
      "loss": 1.6137,
      "theoretical_loss": 4.259311070219441,
      "tokens_seen": 239599616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683650952858576,
      "loss": 1.4483,
      "theoretical_loss": 4.259027736400165,
      "tokens_seen": 239730688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046834503510531595,
      "loss": 1.618,
      "theoretical_loss": 4.258744600798739,
      "tokens_seen": 239861760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683249749247743,
      "loss": 1.6814,
      "theoretical_loss": 4.258461663168285,
      "tokens_seen": 239992832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683049147442327,
      "loss": 1.5598,
      "theoretical_loss": 4.258178923262376,
      "tokens_seen": 240123904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682848545636911,
      "loss": 1.6794,
      "theoretical_loss": 4.257896380835018,
      "tokens_seen": 240254976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046826479438314946,
      "loss": 1.5066,
      "theoretical_loss": 4.257614035640662,
      "tokens_seen": 240386048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682447342026078,
      "loss": 1.7484,
      "theoretical_loss": 4.257331887434198,
      "tokens_seen": 240517120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682246740220662,
      "loss": 1.6482,
      "theoretical_loss": 4.25704993597095,
      "tokens_seen": 240648192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682046138415246,
      "loss": 1.5145,
      "theoretical_loss": 4.256768181006683,
      "tokens_seen": 240779264
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.35654935240745544,
      "objective/train/avg_token_score": 0.024610688909888268,
      "objective/train/docs_used": 148118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2514936923980713,
      "objective/train/original_loss": 3.9919378757476807,
      "objective/train/theoretical_loss": 4.2566273771354615,
      "objective/train/tokens_used": 240844800,
      "objective/train/unlikelihood_loss": 0.20598264038562775,
      "theoretical_loss": 4.2566273771354615,
      "tokens_seen": 240844800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046818455366098296,
      "loss": 1.5226,
      "theoretical_loss": 4.256486622297595,
      "tokens_seen": 240910336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681644934804413,
      "loss": 1.7672,
      "theoretical_loss": 4.256205259600321,
      "tokens_seen": 241041408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681444332998997,
      "loss": 1.6025,
      "theoretical_loss": 4.255924092671931,
      "tokens_seen": 241172480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046812437311935805,
      "loss": 1.5735,
      "theoretical_loss": 4.255643121269924,
      "tokens_seen": 241303552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046810431293881646,
      "loss": 1.8846,
      "theoretical_loss": 4.255362345152234,
      "tokens_seen": 241434624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680842527582748,
      "loss": 1.5332,
      "theoretical_loss": 4.255081764077224,
      "tokens_seen": 241565696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680641925777332,
      "loss": 1.6185,
      "theoretical_loss": 4.254801377803689,
      "tokens_seen": 241696768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046804413239719155,
      "loss": 1.5774,
      "theoretical_loss": 4.254521186090852,
      "tokens_seen": 241827840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046802407221664997,
      "loss": 1.728,
      "theoretical_loss": 4.254241188698361,
      "tokens_seen": 241958912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046800401203610833,
      "loss": 1.5526,
      "theoretical_loss": 4.2539613853862965,
      "tokens_seen": 242089984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679839518555667,
      "loss": 1.6072,
      "theoretical_loss": 4.253681775915161,
      "tokens_seen": 242221056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046796389167502505,
      "loss": 1.7814,
      "theoretical_loss": 4.253402360045882,
      "tokens_seen": 242352128
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.2602028250694275,
      "objective/train/avg_token_score": 0.019304223358631134,
      "objective/train/docs_used": 148826,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9984758496284485,
      "objective/train/original_loss": 4.527791500091553,
      "objective/train/theoretical_loss": 4.253123137539814,
      "objective/train/tokens_used": 242483200,
      "objective/train/unlikelihood_loss": 0.15593910217285156,
      "theoretical_loss": 4.253123137539814,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679438314944834,
      "loss": 1.4314,
      "theoretical_loss": 4.253123137539814,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046792377131394183,
      "loss": 1.5239,
      "theoretical_loss": 4.25284410815873,
      "tokens_seen": 242614272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679037111334002,
      "loss": 1.5362,
      "theoretical_loss": 4.252565271664828,
      "tokens_seen": 242745344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046788365095285856,
      "loss": 1.663,
      "theoretical_loss": 4.252286627820727,
      "tokens_seen": 242876416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678635907723169,
      "loss": 1.4544,
      "theoretical_loss": 4.252008176389465,
      "tokens_seen": 243007488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046784353059177534,
      "loss": 1.5947,
      "theoretical_loss": 4.251729917134498,
      "tokens_seen": 243138560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678234704112337,
      "loss": 1.4033,
      "theoretical_loss": 4.251451849819704,
      "tokens_seen": 243269632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046780341023069206,
      "loss": 1.618,
      "theoretical_loss": 4.251173974209375,
      "tokens_seen": 243400704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677833500501504,
      "loss": 1.5327,
      "theoretical_loss": 4.250896290068218,
      "tokens_seen": 243531776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677632898696088,
      "loss": 1.7856,
      "theoretical_loss": 4.25061879716136,
      "tokens_seen": 243662848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677432296890672,
      "loss": 1.9125,
      "theoretical_loss": 4.250341495254337,
      "tokens_seen": 243793920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046772316950852556,
      "loss": 1.6721,
      "theoretical_loss": 4.250064384113102,
      "tokens_seen": 243924992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467703109327984,
      "loss": 1.4838,
      "theoretical_loss": 4.249787463504019,
      "tokens_seen": 244056064
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.20821113884449005,
      "objective/train/avg_token_score": 0.05237054452300072,
      "objective/train/docs_used": 150000,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8499055504798889,
      "objective/train/original_loss": 5.0128703117370605,
      "objective/train/theoretical_loss": 4.249649074576134,
      "objective/train/tokens_used": 244121600,
      "objective/train/unlikelihood_loss": 0.0956830307841301,
      "theoretical_loss": 4.249649074576134,
      "tokens_seen": 244121600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046768304914744234,
      "loss": 1.5865,
      "theoretical_loss": 4.249510733193862,
      "tokens_seen": 244187136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046766298896690076,
      "loss": 1.6801,
      "theoretical_loss": 4.249234192949818,
      "tokens_seen": 244318208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676429287863591,
      "loss": 1.5263,
      "theoretical_loss": 4.248957842539484,
      "tokens_seen": 244449280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676228686058175,
      "loss": 1.5385,
      "theoretical_loss": 4.2486816817308615,
      "tokens_seen": 244580352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046760280842527585,
      "loss": 1.5574,
      "theoretical_loss": 4.248405710292364,
      "tokens_seen": 244711424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675827482447342,
      "loss": 1.5474,
      "theoretical_loss": 4.248129927992808,
      "tokens_seen": 244842496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675626880641926,
      "loss": 1.64,
      "theoretical_loss": 4.247854334601421,
      "tokens_seen": 244973568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467542627883651,
      "loss": 1.6484,
      "theoretical_loss": 4.247578929887829,
      "tokens_seen": 245104640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046752256770310935,
      "loss": 1.5901,
      "theoretical_loss": 4.247303713622067,
      "tokens_seen": 245235712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675025075225677,
      "loss": 1.697,
      "theoretical_loss": 4.247028685574569,
      "tokens_seen": 245366784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046748244734202613,
      "loss": 1.8329,
      "theoretical_loss": 4.246753845516174,
      "tokens_seen": 245497856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674623871614845,
      "loss": 1.582,
      "theoretical_loss": 4.246479193218123,
      "tokens_seen": 245628928
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.35355570912361145,
      "objective/train/avg_token_score": 0.023441534489393234,
      "objective/train/docs_used": 150657,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.178994059562683,
      "objective/train/original_loss": 3.2855653762817383,
      "objective/train/theoretical_loss": 4.246204728452055,
      "objective/train/tokens_used": 245760000,
      "objective/train/unlikelihood_loss": 0.2683996260166168,
      "theoretical_loss": 4.246204728452055,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046744232698094285,
      "loss": 1.4603,
      "theoretical_loss": 4.246204728452055,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674222668004012,
      "loss": 1.5375,
      "theoretical_loss": 4.245930450990007,
      "tokens_seen": 245891072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046740220661985963,
      "loss": 1.4696,
      "theoretical_loss": 4.245656360604417,
      "tokens_seen": 246022144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467382146439318,
      "loss": 1.641,
      "theoretical_loss": 4.2453824570681205,
      "tokens_seen": 246153216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046736208625877636,
      "loss": 1.5838,
      "theoretical_loss": 4.2451087401543495,
      "tokens_seen": 246284288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673420260782347,
      "loss": 1.736,
      "theoretical_loss": 4.24483520963673,
      "tokens_seen": 246415360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673219658976931,
      "loss": 1.4829,
      "theoretical_loss": 4.244561865289285,
      "tokens_seen": 246546432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673019057171515,
      "loss": 1.6817,
      "theoretical_loss": 4.2442887068864295,
      "tokens_seen": 246677504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046728184553660986,
      "loss": 1.5743,
      "theoretical_loss": 4.244015734202973,
      "tokens_seen": 246808576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672617853560682,
      "loss": 1.6287,
      "theoretical_loss": 4.243742947014117,
      "tokens_seen": 246939648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672417251755266,
      "loss": 1.5413,
      "theoretical_loss": 4.243470345095453,
      "tokens_seen": 247070720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467221664994985,
      "loss": 1.5116,
      "theoretical_loss": 4.243197928222964,
      "tokens_seen": 247201792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046720160481444336,
      "loss": 1.6491,
      "theoretical_loss": 4.242925696173021,
      "tokens_seen": 247332864
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.2963709533214569,
      "objective/train/avg_token_score": 0.031853631138801575,
      "objective/train/docs_used": 151862,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1483128070831299,
      "objective/train/original_loss": 4.083296775817871,
      "objective/train/theoretical_loss": 4.2427896493867285,
      "objective/train/tokens_used": 247398400,
      "objective/train/unlikelihood_loss": 0.13969635963439941,
      "theoretical_loss": 4.2427896493867285,
      "tokens_seen": 247398400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004671815446339017,
      "loss": 1.597,
      "theoretical_loss": 4.242653648722387,
      "tokens_seen": 247463936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671614844533601,
      "loss": 1.5181,
      "theoretical_loss": 4.24238178564821,
      "tokens_seen": 247595008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046714142427281845,
      "loss": 1.7096,
      "theoretical_loss": 4.242110106728022,
      "tokens_seen": 247726080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046712136409227686,
      "loss": 1.7045,
      "theoretical_loss": 4.241838611739748,
      "tokens_seen": 247857152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671013039117352,
      "loss": 1.4473,
      "theoretical_loss": 4.241567300461693,
      "tokens_seen": 247988224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670812437311936,
      "loss": 1.4892,
      "theoretical_loss": 4.241296172672547,
      "tokens_seen": 248119296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046706118355065195,
      "loss": 1.4327,
      "theoretical_loss": 4.241025228151383,
      "tokens_seen": 248250368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046704112337011037,
      "loss": 1.4395,
      "theoretical_loss": 4.240754466677659,
      "tokens_seen": 248381440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046702106318956873,
      "loss": 1.3767,
      "theoretical_loss": 4.240483888031212,
      "tokens_seen": 248512512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670010030090271,
      "loss": 1.6805,
      "theoretical_loss": 4.240213491992261,
      "tokens_seen": 248643584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046698094282848545,
      "loss": 1.7161,
      "theoretical_loss": 4.239943278341404,
      "tokens_seen": 248774656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669608826479438,
      "loss": 1.6768,
      "theoretical_loss": 4.239673246859619,
      "tokens_seen": 248905728
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.40518084168434143,
      "objective/train/avg_token_score": 0.00906757079064846,
      "objective/train/docs_used": 152583,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6137572526931763,
      "objective/train/original_loss": 3.96043062210083,
      "objective/train/theoretical_loss": 4.239403397328261,
      "objective/train/tokens_used": 249036800,
      "objective/train/unlikelihood_loss": 0.250114768743515,
      "theoretical_loss": 4.239403397328261,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046694082246740223,
      "loss": 1.6961,
      "theoretical_loss": 4.239403397328261,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669207622868606,
      "loss": 1.5261,
      "theoretical_loss": 4.239133729529064,
      "tokens_seen": 249167872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046690070210631896,
      "loss": 1.7208,
      "theoretical_loss": 4.2388642432441355,
      "tokens_seen": 249298944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668806419257773,
      "loss": 1.4695,
      "theoretical_loss": 4.238594938255963,
      "tokens_seen": 249430016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046686058174523574,
      "loss": 1.5757,
      "theoretical_loss": 4.238325814347404,
      "tokens_seen": 249561088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668405215646941,
      "loss": 1.5495,
      "theoretical_loss": 4.238056871301695,
      "tokens_seen": 249692160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046682046138415246,
      "loss": 1.4589,
      "theoretical_loss": 4.237788108902441,
      "tokens_seen": 249823232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668004012036108,
      "loss": 1.5079,
      "theoretical_loss": 4.237519526933622,
      "tokens_seen": 249954304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667803410230692,
      "loss": 1.6051,
      "theoretical_loss": 4.237251125179588,
      "tokens_seen": 250085376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667602808425276,
      "loss": 1.6881,
      "theoretical_loss": 4.23698290342506,
      "tokens_seen": 250216448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046674022066198596,
      "loss": 1.4733,
      "theoretical_loss": 4.23671486145513,
      "tokens_seen": 250347520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667201604814443,
      "loss": 1.4943,
      "theoretical_loss": 4.236446999055257,
      "tokens_seen": 250478592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667001003009027,
      "loss": 1.5532,
      "theoretical_loss": 4.2361793160112695,
      "tokens_seen": 250609664
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.38782989978790283,
      "objective/train/avg_token_score": 0.009073655121028423,
      "objective/train/docs_used": 153242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.569201111793518,
      "objective/train/original_loss": 4.112845420837402,
      "objective/train/theoretical_loss": 4.236045541680905,
      "objective/train/tokens_used": 250675200,
      "objective/train/unlikelihood_loss": 0.14907237887382507,
      "theoretical_loss": 4.236045541680905,
      "tokens_seen": 250675200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666800401203611,
      "loss": 1.5928,
      "theoretical_loss": 4.235911812109363,
      "tokens_seen": 250740736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046665997993981947,
      "loss": 1.5379,
      "theoretical_loss": 4.235644487136098,
      "tokens_seen": 250871808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046663991975927783,
      "loss": 1.5322,
      "theoretical_loss": 4.235377340878404,
      "tokens_seen": 251002880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666198595787362,
      "loss": 1.4665,
      "theoretical_loss": 4.235110373123572,
      "tokens_seen": 251133952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046659979939819455,
      "loss": 1.4523,
      "theoretical_loss": 4.2348435836592575,
      "tokens_seen": 251265024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046657973921765297,
      "loss": 1.6582,
      "theoretical_loss": 4.234576972273481,
      "tokens_seen": 251396096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046655967903711133,
      "loss": 1.662,
      "theoretical_loss": 4.234310538754624,
      "tokens_seen": 251527168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665396188565697,
      "loss": 1.4291,
      "theoretical_loss": 4.234044282891429,
      "tokens_seen": 251658240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046651955867602806,
      "loss": 1.4925,
      "theoretical_loss": 4.233778204473002,
      "tokens_seen": 251789312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664994984954865,
      "loss": 1.6234,
      "theoretical_loss": 4.233512303288807,
      "tokens_seen": 251920384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046647943831494484,
      "loss": 1.5104,
      "theoretical_loss": 4.233246579128666,
      "tokens_seen": 252051456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664593781344032,
      "loss": 1.5029,
      "theoretical_loss": 4.232981031782761,
      "tokens_seen": 252182528
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.3658357858657837,
      "objective/train/avg_token_score": 0.005103480536490679,
      "objective/train/docs_used": 154565,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2179532051086426,
      "objective/train/original_loss": 3.7866077423095703,
      "objective/train/theoretical_loss": 4.232715661041632,
      "objective/train/tokens_used": 252313600,
      "objective/train/unlikelihood_loss": 0.20572546124458313,
      "theoretical_loss": 4.232715661041632,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046643931795386156,
      "loss": 1.6638,
      "theoretical_loss": 4.232715661041632,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664192577733199,
      "loss": 1.4777,
      "theoretical_loss": 4.232450466696174,
      "tokens_seen": 252444672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046639919759277834,
      "loss": 1.4621,
      "theoretical_loss": 4.23218544853764,
      "tokens_seen": 252575744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663791374122367,
      "loss": 1.815,
      "theoretical_loss": 4.231920606357638,
      "tokens_seen": 252706816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046635907723169506,
      "loss": 1.5611,
      "theoretical_loss": 4.231655939948127,
      "tokens_seen": 252837888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663390170511534,
      "loss": 1.6306,
      "theoretical_loss": 4.231391449101425,
      "tokens_seen": 252968960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046631895687061184,
      "loss": 1.5308,
      "theoretical_loss": 4.231127133610198,
      "tokens_seen": 253100032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662988966900702,
      "loss": 1.4716,
      "theoretical_loss": 4.230862993267468,
      "tokens_seen": 253231104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046627883650952857,
      "loss": 1.5696,
      "theoretical_loss": 4.230599027866606,
      "tokens_seen": 253362176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046625877632898693,
      "loss": 1.6201,
      "theoretical_loss": 4.2303352372013325,
      "tokens_seen": 253493248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046623871614844535,
      "loss": 1.5034,
      "theoretical_loss": 4.230071621065721,
      "tokens_seen": 253624320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662186559679037,
      "loss": 1.422,
      "theoretical_loss": 4.229808179254192,
      "tokens_seen": 253755392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661985957873621,
      "loss": 1.49,
      "theoretical_loss": 4.229544911561513,
      "tokens_seen": 253886464
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.5327468514442444,
      "objective/train/avg_token_score": 0.0074698301032185555,
      "objective/train/docs_used": 156106,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.954049825668335,
      "objective/train/original_loss": 3.7289023399353027,
      "objective/train/theoretical_loss": 4.229413342945703,
      "objective/train/tokens_used": 253952000,
      "objective/train/unlikelihood_loss": 0.149134561419487,
      "theoretical_loss": 4.229413342945703,
      "tokens_seen": 253952000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661785356068205,
      "loss": 1.7898,
      "theoretical_loss": 4.229281817782801,
      "tokens_seen": 254017536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046615847542627885,
      "loss": 1.6864,
      "theoretical_loss": 4.229018897713519,
      "tokens_seen": 254148608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046613841524573727,
      "loss": 1.3615,
      "theoretical_loss": 4.228756151149475,
      "tokens_seen": 254279680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046611835506519563,
      "loss": 1.615,
      "theoretical_loss": 4.228493577886824,
      "tokens_seen": 254410752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466098294884654,
      "loss": 1.5096,
      "theoretical_loss": 4.228231177722063,
      "tokens_seen": 254541824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046607823470411235,
      "loss": 1.4544,
      "theoretical_loss": 4.227968950452035,
      "tokens_seen": 254672896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046605817452357077,
      "loss": 1.661,
      "theoretical_loss": 4.227706895873924,
      "tokens_seen": 254803968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046603811434302913,
      "loss": 1.4282,
      "theoretical_loss": 4.227445013785257,
      "tokens_seen": 254935040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660180541624875,
      "loss": 1.6551,
      "theoretical_loss": 4.227183303983901,
      "tokens_seen": 255066112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046599799398194586,
      "loss": 1.4951,
      "theoretical_loss": 4.226921766268067,
      "tokens_seen": 255197184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659779338014042,
      "loss": 1.449,
      "theoretical_loss": 4.226660400436302,
      "tokens_seen": 255328256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046595787362086263,
      "loss": 1.5032,
      "theoretical_loss": 4.226399206287493,
      "tokens_seen": 255459328
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.4268695116043091,
      "objective/train/avg_token_score": 0.012001312337815762,
      "objective/train/docs_used": 156630,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8062608242034912,
      "objective/train/original_loss": 4.282796382904053,
      "objective/train/theoretical_loss": 4.226138183620867,
      "objective/train/tokens_used": 255590400,
      "objective/train/unlikelihood_loss": 0.11538192629814148,
      "theoretical_loss": 4.226138183620867,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465937813440321,
      "loss": 1.5151,
      "theoretical_loss": 4.226138183620867,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046591775325977936,
      "loss": 1.6603,
      "theoretical_loss": 4.225877332235987,
      "tokens_seen": 255721472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658976930792377,
      "loss": 1.4763,
      "theoretical_loss": 4.225616651932753,
      "tokens_seen": 255852544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046587763289869614,
      "loss": 1.6467,
      "theoretical_loss": 4.225356142511402,
      "tokens_seen": 255983616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658575727181545,
      "loss": 1.5337,
      "theoretical_loss": 4.225095803772507,
      "tokens_seen": 256114688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046583751253761286,
      "loss": 1.5934,
      "theoretical_loss": 4.224835635516973,
      "tokens_seen": 256245760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658174523570712,
      "loss": 1.4188,
      "theoretical_loss": 4.224575637546041,
      "tokens_seen": 256376832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657973921765296,
      "loss": 1.3615,
      "theoretical_loss": 4.2243158096612845,
      "tokens_seen": 256507904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465777331995988,
      "loss": 1.7877,
      "theoretical_loss": 4.22405615166461,
      "tokens_seen": 256638976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046575727181544636,
      "loss": 1.652,
      "theoretical_loss": 4.223796663358255,
      "tokens_seen": 256770048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657372116349047,
      "loss": 1.5669,
      "theoretical_loss": 4.223537344544788,
      "tokens_seen": 256901120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657171514543631,
      "loss": 1.5028,
      "theoretical_loss": 4.22327819502711,
      "tokens_seen": 257032192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656970912738215,
      "loss": 1.5612,
      "theoretical_loss": 4.223019214608446,
      "tokens_seen": 257163264
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.35031768679618835,
      "objective/train/avg_token_score": 0.06552369147539139,
      "objective/train/docs_used": 157258,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4199466705322266,
      "objective/train/original_loss": 4.115721225738525,
      "objective/train/theoretical_loss": 4.222889787749845,
      "objective/train/tokens_used": 257228800,
      "objective/train/unlikelihood_loss": 0.19470366835594177,
      "theoretical_loss": 4.222889787749845,
      "tokens_seen": 257228800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046567703109327987,
      "loss": 1.5506,
      "theoretical_loss": 4.222760403092358,
      "tokens_seen": 257294336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046565697091273823,
      "loss": 1.5487,
      "theoretical_loss": 4.222501760282729,
      "tokens_seen": 257425408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656369107321966,
      "loss": 1.5823,
      "theoretical_loss": 4.222243285983772,
      "tokens_seen": 257556480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046561685055165495,
      "loss": 1.6221,
      "theoretical_loss": 4.221984980000029,
      "tokens_seen": 257687552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046559679037111337,
      "loss": 1.7549,
      "theoretical_loss": 4.221726842136364,
      "tokens_seen": 257818624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046557673019057173,
      "loss": 1.7467,
      "theoretical_loss": 4.221468872197967,
      "tokens_seen": 257949696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655566700100301,
      "loss": 1.6515,
      "theoretical_loss": 4.221211069990357,
      "tokens_seen": 258080768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046553660982948846,
      "loss": 1.5346,
      "theoretical_loss": 4.22095343531937,
      "tokens_seen": 258211840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655165496489469,
      "loss": 1.6984,
      "theoretical_loss": 4.220695967991171,
      "tokens_seen": 258342912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046549648946840524,
      "loss": 1.5019,
      "theoretical_loss": 4.220438667812244,
      "tokens_seen": 258473984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654764292878636,
      "loss": 1.5388,
      "theoretical_loss": 4.220181534589398,
      "tokens_seen": 258605056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046545636910732196,
      "loss": 1.7965,
      "theoretical_loss": 4.219924568129759,
      "tokens_seen": 258736128
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.4152003824710846,
      "objective/train/avg_token_score": 0.016226625069975853,
      "objective/train/docs_used": 158510,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.525114893913269,
      "objective/train/original_loss": 4.261040210723877,
      "objective/train/theoretical_loss": 4.219667768240775,
      "objective/train/tokens_used": 258867200,
      "objective/train/unlikelihood_loss": 0.1502227634191513,
      "theoretical_loss": 4.219667768240775,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654363089267803,
      "loss": 1.6609,
      "theoretical_loss": 4.219667768240775,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046541624874623874,
      "loss": 1.5927,
      "theoretical_loss": 4.2194111347302155,
      "tokens_seen": 258998272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653961885656971,
      "loss": 1.5297,
      "theoretical_loss": 4.219154667406166,
      "tokens_seen": 259129344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046537612838515546,
      "loss": 1.5517,
      "theoretical_loss": 4.218898366077035,
      "tokens_seen": 259260416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653560682046138,
      "loss": 1.6273,
      "theoretical_loss": 4.218642230551541,
      "tokens_seen": 259391488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046533600802407224,
      "loss": 1.478,
      "theoretical_loss": 4.218386260638727,
      "tokens_seen": 259522560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653159478435306,
      "loss": 1.6475,
      "theoretical_loss": 4.218130456147948,
      "tokens_seen": 259653632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046529588766298897,
      "loss": 1.3824,
      "theoretical_loss": 4.217874816888877,
      "tokens_seen": 259784704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046527582748244733,
      "loss": 1.474,
      "theoretical_loss": 4.217619342671498,
      "tokens_seen": 259915776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046525576730190575,
      "loss": 1.5144,
      "theoretical_loss": 4.217364033306113,
      "tokens_seen": 260046848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652357071213641,
      "loss": 1.7336,
      "theoretical_loss": 4.217108888603337,
      "tokens_seen": 260177920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046521564694082247,
      "loss": 1.577,
      "theoretical_loss": 4.216853908374097,
      "tokens_seen": 260308992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046519558676028083,
      "loss": 1.5176,
      "theoretical_loss": 4.216599092429631,
      "tokens_seen": 260440064
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.37017351388931274,
      "objective/train/avg_token_score": 0.00918534118682146,
      "objective/train/docs_used": 159098,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4167944192886353,
      "objective/train/original_loss": 3.8629274368286133,
      "objective/train/theoretical_loss": 4.216471746005286,
      "objective/train/tokens_used": 260505600,
      "objective/train/unlikelihood_loss": 0.2083285003900528,
      "theoretical_loss": 4.216471746005286,
      "tokens_seen": 260505600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651755265797392,
      "loss": 1.5863,
      "theoretical_loss": 4.216344440581491,
      "tokens_seen": 260571136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651554663991976,
      "loss": 1.4607,
      "theoretical_loss": 4.2160899526415365,
      "tokens_seen": 260702208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465135406218656,
      "loss": 1.394,
      "theoretical_loss": 4.215835628421942,
      "tokens_seen": 260833280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046511534603811434,
      "loss": 1.5575,
      "theoretical_loss": 4.215581467735187,
      "tokens_seen": 260964352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650952858575727,
      "loss": 1.5175,
      "theoretical_loss": 4.215327470394062,
      "tokens_seen": 261095424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650752256770311,
      "loss": 1.5183,
      "theoretical_loss": 4.215073636211664,
      "tokens_seen": 261226496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650551654964895,
      "loss": 1.5299,
      "theoretical_loss": 4.214819965001401,
      "tokens_seen": 261357568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046503510531594784,
      "loss": 1.5135,
      "theoretical_loss": 4.214566456576984,
      "tokens_seen": 261488640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650150451354062,
      "loss": 1.4946,
      "theoretical_loss": 4.214313110752431,
      "tokens_seen": 261619712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046499498495486456,
      "loss": 1.4256,
      "theoretical_loss": 4.214059927342068,
      "tokens_seen": 261750784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464974924774323,
      "loss": 1.473,
      "theoretical_loss": 4.213806906160523,
      "tokens_seen": 261881856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046495486459378134,
      "loss": 1.606,
      "theoretical_loss": 4.213554047022729,
      "tokens_seen": 262012928
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.30541300773620605,
      "objective/train/avg_token_score": 0.026100359857082367,
      "objective/train/docs_used": 160106,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1186697483062744,
      "objective/train/original_loss": 4.202697277069092,
      "objective/train/theoretical_loss": 4.213301349743924,
      "objective/train/tokens_used": 262144000,
      "objective/train/unlikelihood_loss": 0.18503674864768982,
      "theoretical_loss": 4.213301349743924,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649348044132397,
      "loss": 1.6211,
      "theoretical_loss": 4.213301349743924,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046491474423269807,
      "loss": 1.5147,
      "theoretical_loss": 4.213048814139647,
      "tokens_seen": 262275072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648946840521565,
      "loss": 1.4764,
      "theoretical_loss": 4.21279644002574,
      "tokens_seen": 262406144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046487462387161485,
      "loss": 1.5803,
      "theoretical_loss": 4.212544227218347,
      "tokens_seen": 262537216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648545636910732,
      "loss": 1.4761,
      "theoretical_loss": 4.212292175533912,
      "tokens_seen": 262668288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046483450351053157,
      "loss": 1.6497,
      "theoretical_loss": 4.212040284789181,
      "tokens_seen": 262799360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046481444332998993,
      "loss": 1.4949,
      "theoretical_loss": 4.211788554801198,
      "tokens_seen": 262930432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046479438314944835,
      "loss": 1.6391,
      "theoretical_loss": 4.211536985387307,
      "tokens_seen": 263061504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647743229689067,
      "loss": 1.4038,
      "theoretical_loss": 4.2112855763651496,
      "tokens_seen": 263192576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647542627883651,
      "loss": 1.5316,
      "theoretical_loss": 4.211034327552666,
      "tokens_seen": 263323648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046473420260782344,
      "loss": 1.4309,
      "theoretical_loss": 4.210783238768093,
      "tokens_seen": 263454720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647141424272819,
      "loss": 1.5506,
      "theoretical_loss": 4.210532309829965,
      "tokens_seen": 263585792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046469408224674027,
      "loss": 1.5781,
      "theoretical_loss": 4.21028154055711,
      "tokens_seen": 263716864
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.32484114170074463,
      "objective/train/avg_token_score": 0.0183444544672966,
      "objective/train/docs_used": 160649,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2061073780059814,
      "objective/train/original_loss": 4.3018927574157715,
      "objective/train/theoretical_loss": 4.2101562157386265,
      "objective/train/tokens_used": 263782400,
      "objective/train/unlikelihood_loss": 0.16418404877185822,
      "theoretical_loss": 4.2101562157386265,
      "tokens_seen": 263782400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046467402206619863,
      "loss": 1.4389,
      "theoretical_loss": 4.210030930768655,
      "tokens_seen": 263847936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464653961885657,
      "loss": 1.445,
      "theoretical_loss": 4.209780480284017,
      "tokens_seen": 263979008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046463390170511535,
      "loss": 1.5965,
      "theoretical_loss": 4.209530188922911,
      "tokens_seen": 264110080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046461384152457377,
      "loss": 1.4428,
      "theoretical_loss": 4.209280056505342,
      "tokens_seen": 264241152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046459378134403213,
      "loss": 1.3129,
      "theoretical_loss": 4.209030082851612,
      "tokens_seen": 264372224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645737211634905,
      "loss": 1.498,
      "theoretical_loss": 4.20878026778231,
      "tokens_seen": 264503296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046455366098294886,
      "loss": 1.5519,
      "theoretical_loss": 4.208530611118321,
      "tokens_seen": 264634368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645336008024073,
      "loss": 1.5022,
      "theoretical_loss": 4.208281112680817,
      "tokens_seen": 264765440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046451354062186564,
      "loss": 1.5408,
      "theoretical_loss": 4.208031772291265,
      "tokens_seen": 264896512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464493480441324,
      "loss": 1.2914,
      "theoretical_loss": 4.2077825897714165,
      "tokens_seen": 265027584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046447342026078236,
      "loss": 1.4717,
      "theoretical_loss": 4.207533564943316,
      "tokens_seen": 265158656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644533600802407,
      "loss": 1.5337,
      "theoretical_loss": 4.2072846976292935,
      "tokens_seen": 265289728
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.37787145376205444,
      "objective/train/avg_token_score": 0.033646829426288605,
      "objective/train/docs_used": 161970,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4991329908370972,
      "objective/train/original_loss": 4.190028667449951,
      "objective/train/theoretical_loss": 4.20703598765197,
      "objective/train/tokens_used": 265420800,
      "objective/train/unlikelihood_loss": 0.1276427060365677,
      "theoretical_loss": 4.20703598765197,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046443329989969914,
      "loss": 1.5316,
      "theoretical_loss": 4.20703598765197,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644132397191575,
      "loss": 1.5589,
      "theoretical_loss": 4.206787434834251,
      "tokens_seen": 265551872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046439317953861586,
      "loss": 1.4447,
      "theoretical_loss": 4.206539038999329,
      "tokens_seen": 265682944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643731193580742,
      "loss": 1.411,
      "theoretical_loss": 4.206290799970685,
      "tokens_seen": 265814016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046435305917753264,
      "loss": 1.4428,
      "theoretical_loss": 4.206042717572082,
      "tokens_seen": 265945088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464332998996991,
      "loss": 1.4921,
      "theoretical_loss": 4.20579479162757,
      "tokens_seen": 266076160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046431293881644937,
      "loss": 1.487,
      "theoretical_loss": 4.205547021961482,
      "tokens_seen": 266207232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046429287863590773,
      "loss": 1.5737,
      "theoretical_loss": 4.205299408398435,
      "tokens_seen": 266338304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046427281845536615,
      "loss": 1.6461,
      "theoretical_loss": 4.20505195076333,
      "tokens_seen": 266469376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642527582748245,
      "loss": 1.7422,
      "theoretical_loss": 4.20480464888135,
      "tokens_seen": 266600448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046423269809428287,
      "loss": 1.4776,
      "theoretical_loss": 4.204557502577957,
      "tokens_seen": 266731520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046421263791374123,
      "loss": 1.63,
      "theoretical_loss": 4.2043105116789,
      "tokens_seen": 266862592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641925777331996,
      "loss": 1.7364,
      "theoretical_loss": 4.204063676010202,
      "tokens_seen": 266993664
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.494012713432312,
      "objective/train/avg_token_score": 0.00496962945908308,
      "objective/train/docs_used": 162606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7318894863128662,
      "objective/train/original_loss": 3.5257256031036377,
      "objective/train/theoretical_loss": 4.203940316332948,
      "objective/train/tokens_used": 267059200,
      "objective/train/unlikelihood_loss": 0.2108738273382187,
      "theoretical_loss": 4.203940316332948,
      "tokens_seen": 267059200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464172517552658,
      "loss": 1.3925,
      "theoretical_loss": 4.203816995398171,
      "tokens_seen": 267124736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641524573721164,
      "loss": 1.6141,
      "theoretical_loss": 4.203570469669392,
      "tokens_seen": 267255808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046413239719157474,
      "loss": 1.608,
      "theoretical_loss": 4.203324098650731,
      "tokens_seen": 267386880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641123370110331,
      "loss": 1.761,
      "theoretical_loss": 4.20307788216933,
      "tokens_seen": 267517952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640922768304915,
      "loss": 1.636,
      "theoretical_loss": 4.202831820052609,
      "tokens_seen": 267649024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640722166499499,
      "loss": 1.5856,
      "theoretical_loss": 4.202585912128266,
      "tokens_seen": 267780096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046405215646940824,
      "loss": 1.6837,
      "theoretical_loss": 4.202340158224277,
      "tokens_seen": 267911168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640320962888666,
      "loss": 1.6194,
      "theoretical_loss": 4.2020945581688895,
      "tokens_seen": 268042240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046401203610832496,
      "loss": 1.7856,
      "theoretical_loss": 4.201849111790631,
      "tokens_seen": 268173312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639919759277834,
      "loss": 1.6562,
      "theoretical_loss": 4.201603818918302,
      "tokens_seen": 268304384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046397191574724174,
      "loss": 1.5358,
      "theoretical_loss": 4.201358679380976,
      "tokens_seen": 268435456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639518555667001,
      "loss": 1.6419,
      "theoretical_loss": 4.201113693008002,
      "tokens_seen": 268566528
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.3462243378162384,
      "objective/train/avg_token_score": 0.018634682521224022,
      "objective/train/docs_used": 163909,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2885785102844238,
      "objective/train/original_loss": 4.379140853881836,
      "objective/train/theoretical_loss": 4.2008688596290025,
      "objective/train/tokens_used": 268697600,
      "objective/train/unlikelihood_loss": 0.12308195978403091,
      "theoretical_loss": 4.2008688596290025,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046393179538615847,
      "loss": 1.5429,
      "theoretical_loss": 4.2008688596290025,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639117352056169,
      "loss": 1.596,
      "theoretical_loss": 4.200624179073869,
      "tokens_seen": 268828672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046389167502507525,
      "loss": 1.6086,
      "theoretical_loss": 4.200379651172769,
      "tokens_seen": 268959744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638716148445336,
      "loss": 1.4534,
      "theoretical_loss": 4.200135275756139,
      "tokens_seen": 269090816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046385155466399197,
      "loss": 1.6596,
      "theoretical_loss": 4.199891052654689,
      "tokens_seen": 269221888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046383149448345033,
      "loss": 1.4265,
      "theoretical_loss": 4.199646981699395,
      "tokens_seen": 269352960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046381143430290875,
      "loss": 1.3837,
      "theoretical_loss": 4.199403062721506,
      "tokens_seen": 269484032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637913741223671,
      "loss": 1.5549,
      "theoretical_loss": 4.1991592955525405,
      "tokens_seen": 269615104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637713139418255,
      "loss": 1.5172,
      "theoretical_loss": 4.198915680024282,
      "tokens_seen": 269746176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046375125376128384,
      "loss": 1.5907,
      "theoretical_loss": 4.198672215968785,
      "tokens_seen": 269877248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046373119358074225,
      "loss": 1.5905,
      "theoretical_loss": 4.198428903218371,
      "tokens_seen": 270008320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637111334002006,
      "loss": 1.5348,
      "theoretical_loss": 4.198185741605628,
      "tokens_seen": 270139392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463691073219659,
      "loss": 1.4577,
      "theoretical_loss": 4.197942730963412,
      "tokens_seen": 270270464
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.5173509120941162,
      "objective/train/avg_token_score": 0.017678294330835342,
      "objective/train/docs_used": 164570,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0276877880096436,
      "objective/train/original_loss": 4.127265930175781,
      "objective/train/theoretical_loss": 4.19782128220409,
      "objective/train/tokens_used": 270336000,
      "objective/train/unlikelihood_loss": 0.1530839055776596,
      "theoretical_loss": 4.19782128220409,
      "tokens_seen": 270336000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046367101303911734,
      "loss": 1.426,
      "theoretical_loss": 4.19769987112484,
      "tokens_seen": 270401536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636509528585757,
      "loss": 1.4833,
      "theoretical_loss": 4.1974571619233,
      "tokens_seen": 270532608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636308926780341,
      "loss": 1.5496,
      "theoretical_loss": 4.1972146031924416,
      "tokens_seen": 270663680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636108324974925,
      "loss": 1.6277,
      "theoretical_loss": 4.196972194766179,
      "tokens_seen": 270794752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046359077231695084,
      "loss": 1.5659,
      "theoretical_loss": 4.1967299364786905,
      "tokens_seen": 270925824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635707121364092,
      "loss": 1.6559,
      "theoretical_loss": 4.1964878281644165,
      "tokens_seen": 271056896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635506519558676,
      "loss": 1.5899,
      "theoretical_loss": 4.196245869658061,
      "tokens_seen": 271187968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463530591775326,
      "loss": 1.6648,
      "theoretical_loss": 4.196004060794589,
      "tokens_seen": 271319040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046351053159478434,
      "loss": 1.5551,
      "theoretical_loss": 4.195762401409229,
      "tokens_seen": 271450112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634904714142427,
      "loss": 1.6318,
      "theoretical_loss": 4.195520891337466,
      "tokens_seen": 271581184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046347041123370107,
      "loss": 1.7221,
      "theoretical_loss": 4.19527953041505,
      "tokens_seen": 271712256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634503510531595,
      "loss": 1.3852,
      "theoretical_loss": 4.195038318477989,
      "tokens_seen": 271843328
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.5086143612861633,
      "objective/train/avg_token_score": 0.018605908378958702,
      "objective/train/docs_used": 165961,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.964569330215454,
      "objective/train/original_loss": 4.285537242889404,
      "objective/train/theoretical_loss": 4.194797255362549,
      "objective/train/tokens_used": 271974400,
      "objective/train/unlikelihood_loss": 0.10453181713819504,
      "theoretical_loss": 4.194797255362549,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046343029087261785,
      "loss": 1.6417,
      "theoretical_loss": 4.194797255362549,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634102306920762,
      "loss": 1.5556,
      "theoretical_loss": 4.194556340905256,
      "tokens_seen": 272105472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046339017051153457,
      "loss": 1.6563,
      "theoretical_loss": 4.194315574942896,
      "tokens_seen": 272236544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463370110330993,
      "loss": 1.6543,
      "theoretical_loss": 4.19407495731251,
      "tokens_seen": 272367616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046335005015045135,
      "loss": 1.608,
      "theoretical_loss": 4.193834487851396,
      "tokens_seen": 272498688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633299899699097,
      "loss": 1.4458,
      "theoretical_loss": 4.1935941663971095,
      "tokens_seen": 272629760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633099297893681,
      "loss": 1.5378,
      "theoretical_loss": 4.193353992787463,
      "tokens_seen": 272760832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046328986960882644,
      "loss": 1.3823,
      "theoretical_loss": 4.1931139668605235,
      "tokens_seen": 272891904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046326980942828485,
      "loss": 1.47,
      "theoretical_loss": 4.192874088454613,
      "tokens_seen": 273022976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632497492477432,
      "loss": 1.508,
      "theoretical_loss": 4.192634357408309,
      "tokens_seen": 273154048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632296890672016,
      "loss": 1.5608,
      "theoretical_loss": 4.192394773560441,
      "tokens_seen": 273285120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046320962888666,
      "loss": 1.5031,
      "theoretical_loss": 4.192155336750094,
      "tokens_seen": 273416192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631895687061184,
      "loss": 1.8175,
      "theoretical_loss": 4.191916046816605,
      "tokens_seen": 273547264
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.39100682735443115,
      "objective/train/avg_token_score": 0.018308745697140694,
      "objective/train/docs_used": 166474,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.588956356048584,
      "objective/train/original_loss": 4.311248779296875,
      "objective/train/theoretical_loss": 4.191796456878544,
      "objective/train/tokens_used": 273612800,
      "objective/train/unlikelihood_loss": 0.15257127583026886,
      "theoretical_loss": 4.191796456878544,
      "tokens_seen": 273612800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631695085255768,
      "loss": 1.6023,
      "theoretical_loss": 4.191676903599563,
      "tokens_seen": 273678336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046314944834503514,
      "loss": 1.4118,
      "theoretical_loss": 4.191437906938811,
      "tokens_seen": 273809408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631293881644935,
      "loss": 1.557,
      "theoretical_loss": 4.19119905667444,
      "tokens_seen": 273940480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631093279839519,
      "loss": 1.3291,
      "theoretical_loss": 4.190960352646796,
      "tokens_seen": 274071552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630892678034103,
      "loss": 1.3456,
      "theoretical_loss": 4.19072179469647,
      "tokens_seen": 274202624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046306920762286864,
      "loss": 1.3644,
      "theoretical_loss": 4.190483382664308,
      "tokens_seen": 274333696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463049147442327,
      "loss": 1.624,
      "theoretical_loss": 4.190245116391403,
      "tokens_seen": 274464768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046302908726178536,
      "loss": 1.5528,
      "theoretical_loss": 4.190006995719098,
      "tokens_seen": 274595840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630090270812438,
      "loss": 1.4449,
      "theoretical_loss": 4.189769020488981,
      "tokens_seen": 274726912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046298896690070214,
      "loss": 1.7029,
      "theoretical_loss": 4.189531190542893,
      "tokens_seen": 274857984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629689067201605,
      "loss": 1.5462,
      "theoretical_loss": 4.189293505722918,
      "tokens_seen": 274989056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046294884653961887,
      "loss": 1.5461,
      "theoretical_loss": 4.189055965871389,
      "tokens_seen": 275120128
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.48130497336387634,
      "objective/train/avg_token_score": 0.014956628903746605,
      "objective/train/docs_used": 167916,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6151095628738403,
      "objective/train/original_loss": 3.838587999343872,
      "objective/train/theoretical_loss": 4.188818570830883,
      "objective/train/tokens_used": 275251200,
      "objective/train/unlikelihood_loss": 0.14401310682296753,
      "theoretical_loss": 4.188818570830883,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629287863590773,
      "loss": 1.5669,
      "theoretical_loss": 4.188818570830883,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046290872617853565,
      "loss": 1.6388,
      "theoretical_loss": 4.188581320444228,
      "tokens_seen": 275382272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462888665997994,
      "loss": 1.64,
      "theoretical_loss": 4.1883442145544905,
      "tokens_seen": 275513344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046286860581745237,
      "loss": 1.3469,
      "theoretical_loss": 4.188107253004986,
      "tokens_seen": 275644416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046284854563691073,
      "loss": 1.6807,
      "theoretical_loss": 4.187870435639275,
      "tokens_seen": 275775488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046282848545636915,
      "loss": 1.4609,
      "theoretical_loss": 4.187633762301159,
      "tokens_seen": 275906560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628084252758275,
      "loss": 1.3392,
      "theoretical_loss": 4.187397232834683,
      "tokens_seen": 276037632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627883650952859,
      "loss": 1.6093,
      "theoretical_loss": 4.1871608470841375,
      "tokens_seen": 276168704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046276830491474424,
      "loss": 1.6585,
      "theoretical_loss": 4.1869246048940525,
      "tokens_seen": 276299776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046274824473420265,
      "loss": 1.6341,
      "theoretical_loss": 4.186688506109202,
      "tokens_seen": 276430848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462728184553661,
      "loss": 1.513,
      "theoretical_loss": 4.186452550574599,
      "tokens_seen": 276561920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627081243731194,
      "loss": 1.4577,
      "theoretical_loss": 4.186216738135501,
      "tokens_seen": 276692992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046268806419257774,
      "loss": 1.4344,
      "theoretical_loss": 4.185981068637401,
      "tokens_seen": 276824064
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.46810850501060486,
      "objective/train/avg_token_score": 0.0036265684757381678,
      "objective/train/docs_used": 168545,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7359020709991455,
      "objective/train/original_loss": 3.6703922748565674,
      "objective/train/theoretical_loss": 4.185863287443008,
      "objective/train/tokens_used": 276889600,
      "objective/train/unlikelihood_loss": 0.1548842489719391,
      "theoretical_loss": 4.185863287443008,
      "tokens_seen": 276889600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626680040120361,
      "loss": 1.3947,
      "theoretical_loss": 4.185745541926035,
      "tokens_seen": 276955136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626479438314945,
      "loss": 1.4142,
      "theoretical_loss": 4.1855101578473795,
      "tokens_seen": 277086208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626278836509529,
      "loss": 1.5381,
      "theoretical_loss": 4.185274916247646,
      "tokens_seen": 277217280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046260782347041124,
      "loss": 1.5502,
      "theoretical_loss": 4.185039816973289,
      "tokens_seen": 277348352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625877632898696,
      "loss": 1.489,
      "theoretical_loss": 4.184804859870997,
      "tokens_seen": 277479424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462567703109328,
      "loss": 1.4585,
      "theoretical_loss": 4.184570044787698,
      "tokens_seen": 277610496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625476429287864,
      "loss": 1.605,
      "theoretical_loss": 4.184335371570556,
      "tokens_seen": 277741568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046252758274824475,
      "loss": 1.5093,
      "theoretical_loss": 4.1841008400669715,
      "tokens_seen": 277872640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625075225677031,
      "loss": 1.6693,
      "theoretical_loss": 4.183866450124584,
      "tokens_seen": 278003712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046248746238716147,
      "loss": 1.5868,
      "theoretical_loss": 4.183632201591264,
      "tokens_seen": 278134784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624674022066199,
      "loss": 1.7203,
      "theoretical_loss": 4.1833980943151206,
      "tokens_seen": 278265856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046244734202607825,
      "loss": 1.6068,
      "theoretical_loss": 4.183164128144495,
      "tokens_seen": 278396928
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.46242669224739075,
      "objective/train/avg_token_score": 0.020400462672114372,
      "objective/train/docs_used": 169743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8247275352478027,
      "objective/train/original_loss": 4.489882946014404,
      "objective/train/theoretical_loss": 4.182930302927963,
      "objective/train/tokens_used": 278528000,
      "objective/train/unlikelihood_loss": 0.11430107802152634,
      "theoretical_loss": 4.182930302927963,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624272818455366,
      "loss": 1.5359,
      "theoretical_loss": 4.182930302927963,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462407221664995,
      "loss": 1.7889,
      "theoretical_loss": 4.182696618514337,
      "tokens_seen": 278659072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623871614844534,
      "loss": 1.5107,
      "theoretical_loss": 4.1824630747526585,
      "tokens_seen": 278790144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046236710130391175,
      "loss": 1.6843,
      "theoretical_loss": 4.182229671492204,
      "tokens_seen": 278921216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623470411233701,
      "loss": 1.651,
      "theoretical_loss": 4.1819964085824815,
      "tokens_seen": 279052288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623269809428285,
      "loss": 1.5783,
      "theoretical_loss": 4.181763285873231,
      "tokens_seen": 279183360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046230692076228684,
      "loss": 1.5391,
      "theoretical_loss": 4.181530303214423,
      "tokens_seen": 279314432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046228686058174525,
      "loss": 1.4712,
      "theoretical_loss": 4.181297460456262,
      "tokens_seen": 279445504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622668004012036,
      "loss": 1.5429,
      "theoretical_loss": 4.181064757449178,
      "tokens_seen": 279576576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462246740220662,
      "loss": 1.6241,
      "theoretical_loss": 4.180832194043836,
      "tokens_seen": 279707648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046222668004012034,
      "loss": 1.6594,
      "theoretical_loss": 4.180599770091126,
      "tokens_seen": 279838720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046220661985957876,
      "loss": 1.546,
      "theoretical_loss": 4.18036748544217,
      "tokens_seen": 279969792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621865596790371,
      "loss": 1.4317,
      "theoretical_loss": 4.1801353399483165,
      "tokens_seen": 280100864
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.34316959977149963,
      "objective/train/avg_token_score": 0.019939497113227844,
      "objective/train/docs_used": 170343,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2674180269241333,
      "objective/train/original_loss": 4.336735725402832,
      "objective/train/theoretical_loss": 4.180019319338163,
      "objective/train/tokens_used": 280166400,
      "objective/train/unlikelihood_loss": 0.17246992886066437,
      "theoretical_loss": 4.180019319338163,
      "tokens_seen": 280166400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621664994984955,
      "loss": 1.6044,
      "theoretical_loss": 4.179903333461144,
      "tokens_seen": 280231936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046214643931795384,
      "loss": 1.6516,
      "theoretical_loss": 4.179671465832458,
      "tokens_seen": 280363008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621263791374122,
      "loss": 1.5465,
      "theoretical_loss": 4.17943973691429,
      "tokens_seen": 280494080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004621063189568706,
      "loss": 1.5127,
      "theoretical_loss": 4.179208146558899,
      "tokens_seen": 280625152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000462086258776329,
      "loss": 1.5414,
      "theoretical_loss": 4.178976694618772,
      "tokens_seen": 280756224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046206619859578735,
      "loss": 1.5515,
      "theoretical_loss": 4.178745380946619,
      "tokens_seen": 280887296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620461384152457,
      "loss": 1.4725,
      "theoretical_loss": 4.178514205395376,
      "tokens_seen": 281018368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620260782347041,
      "loss": 1.4888,
      "theoretical_loss": 4.178283167818206,
      "tokens_seen": 281149440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620060180541625,
      "loss": 1.4546,
      "theoretical_loss": 4.178052268068494,
      "tokens_seen": 281280512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046198595787362085,
      "loss": 1.4797,
      "theoretical_loss": 4.17782150599985,
      "tokens_seen": 281411584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619658976930792,
      "loss": 1.4201,
      "theoretical_loss": 4.1775908814661085,
      "tokens_seen": 281542656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046194583751253763,
      "loss": 1.3238,
      "theoretical_loss": 4.177360394321325,
      "tokens_seen": 281673728
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.38899070024490356,
      "objective/train/avg_token_score": 0.016084667295217514,
      "objective/train/docs_used": 171008,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.55722177028656,
      "objective/train/original_loss": 4.193042278289795,
      "objective/train/theoretical_loss": 4.17713004441978,
      "objective/train/tokens_used": 281804800,
      "objective/train/unlikelihood_loss": 0.173121839761734,
      "theoretical_loss": 4.17713004441978,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461925777331996,
      "loss": 1.4682,
      "theoretical_loss": 4.17713004441978,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046190571715145435,
      "loss": 1.4808,
      "theoretical_loss": 4.176899831615974,
      "tokens_seen": 281935872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618856569709127,
      "loss": 1.6193,
      "theoretical_loss": 4.176669755764632,
      "tokens_seen": 282066944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618655967903711,
      "loss": 1.5912,
      "theoretical_loss": 4.176439816720697,
      "tokens_seen": 282198016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618455366098295,
      "loss": 1.5359,
      "theoretical_loss": 4.176210014339335,
      "tokens_seen": 282329088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046182547642928786,
      "loss": 1.6488,
      "theoretical_loss": 4.1759803484759335,
      "tokens_seen": 282460160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618054162487462,
      "loss": 1.354,
      "theoretical_loss": 4.175750818986098,
      "tokens_seen": 282591232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617853560682046,
      "loss": 1.4048,
      "theoretical_loss": 4.1755214257256545,
      "tokens_seen": 282722304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461765295887663,
      "loss": 1.4483,
      "theoretical_loss": 4.175292168550648,
      "tokens_seen": 282853376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046174523570712136,
      "loss": 1.4696,
      "theoretical_loss": 4.175063047317342,
      "tokens_seen": 282984448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617251755265797,
      "loss": 1.5138,
      "theoretical_loss": 4.174834061882218,
      "tokens_seen": 283115520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046170511534603814,
      "loss": 1.5536,
      "theoretical_loss": 4.174605212101977,
      "tokens_seen": 283246592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616850551654965,
      "loss": 1.3895,
      "theoretical_loss": 4.174376497833537,
      "tokens_seen": 283377664
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.3909457325935364,
      "objective/train/avg_token_score": 0.02921941690146923,
      "objective/train/docs_used": 172282,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4512357711791992,
      "objective/train/original_loss": 4.0204291343688965,
      "objective/train/theoretical_loss": 4.174262191471587,
      "objective/train/tokens_used": 283443200,
      "objective/train/unlikelihood_loss": 0.15194293856620789,
      "theoretical_loss": 4.174262191471587,
      "tokens_seen": 283443200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616649949849549,
      "loss": 1.681,
      "theoretical_loss": 4.17414791893403,
      "tokens_seen": 283508736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616449348044133,
      "loss": 1.4106,
      "theoretical_loss": 4.173919475260808,
      "tokens_seen": 283639808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046162487462387164,
      "loss": 1.3723,
      "theoretical_loss": 4.173691166671439,
      "tokens_seen": 283770880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046160481444333,
      "loss": 1.5836,
      "theoretical_loss": 4.173462993023706,
      "tokens_seen": 283901952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615847542627884,
      "loss": 1.6088,
      "theoretical_loss": 4.173234954175605,
      "tokens_seen": 284033024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615646940822468,
      "loss": 1.6808,
      "theoretical_loss": 4.173007049985352,
      "tokens_seen": 284164096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046154463390170515,
      "loss": 1.6575,
      "theoretical_loss": 4.172779280311372,
      "tokens_seen": 284295168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615245737211635,
      "loss": 1.7206,
      "theoretical_loss": 4.172551645012307,
      "tokens_seen": 284426240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046150451354062187,
      "loss": 1.5265,
      "theoretical_loss": 4.172324143947012,
      "tokens_seen": 284557312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614844533600803,
      "loss": 1.4038,
      "theoretical_loss": 4.1720967769745565,
      "tokens_seen": 284688384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046146439317953865,
      "loss": 1.5643,
      "theoretical_loss": 4.1718695439542195,
      "tokens_seen": 284819456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461444332998997,
      "loss": 1.6387,
      "theoretical_loss": 4.171642444745497,
      "tokens_seen": 284950528
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.48081621527671814,
      "objective/train/avg_token_score": 0.016248999163508415,
      "objective/train/docs_used": 172985,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8190662860870361,
      "objective/train/original_loss": 3.9090092182159424,
      "objective/train/theoretical_loss": 4.1714154792080915,
      "objective/train/tokens_used": 285081600,
      "objective/train/unlikelihood_loss": 0.19899842143058777,
      "theoretical_loss": 4.1714154792080915,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614242728184554,
      "loss": 1.4715,
      "theoretical_loss": 4.1714154792080915,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614042126379138,
      "loss": 1.5394,
      "theoretical_loss": 4.171188647201921,
      "tokens_seen": 285212672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046138415245737215,
      "loss": 1.7036,
      "theoretical_loss": 4.170961948587115,
      "tokens_seen": 285343744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613640922768305,
      "loss": 1.5484,
      "theoretical_loss": 4.1707353832240095,
      "tokens_seen": 285474816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613440320962889,
      "loss": 1.5526,
      "theoretical_loss": 4.170508950973154,
      "tokens_seen": 285605888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046132397191574724,
      "loss": 1.5348,
      "theoretical_loss": 4.170282651695308,
      "tokens_seen": 285736960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046130391173520566,
      "loss": 1.5794,
      "theoretical_loss": 4.170056485251439,
      "tokens_seen": 285868032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461283851554664,
      "loss": 1.4694,
      "theoretical_loss": 4.169830451502724,
      "tokens_seen": 285999104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612637913741224,
      "loss": 1.383,
      "theoretical_loss": 4.1696045503105506,
      "tokens_seen": 286130176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046124373119358074,
      "loss": 1.5947,
      "theoretical_loss": 4.169378781536509,
      "tokens_seen": 286261248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046122367101303916,
      "loss": 1.6029,
      "theoretical_loss": 4.169153145042405,
      "tokens_seen": 286392320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612036108324975,
      "loss": 1.4988,
      "theoretical_loss": 4.168927640690246,
      "tokens_seen": 286523392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611835506519559,
      "loss": 1.4123,
      "theoretical_loss": 4.168702268342248,
      "tokens_seen": 286654464
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.4003543555736542,
      "objective/train/avg_token_score": 0.02369835413992405,
      "objective/train/docs_used": 174327,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5158040523529053,
      "objective/train/original_loss": 4.125824451446533,
      "objective/train/theoretical_loss": 4.168589631626808,
      "objective/train/tokens_used": 286720000,
      "objective/train/unlikelihood_loss": 0.19976606965065002,
      "theoretical_loss": 4.168589631626808,
      "tokens_seen": 286720000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046116349047141425,
      "loss": 1.4636,
      "theoretical_loss": 4.168477027860833,
      "tokens_seen": 286785536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611434302908726,
      "loss": 1.6322,
      "theoretical_loss": 4.168251919108632,
      "tokens_seen": 286916608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461123370110331,
      "loss": 1.4629,
      "theoretical_loss": 4.168026941948478,
      "tokens_seen": 287047680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611033099297894,
      "loss": 1.4635,
      "theoretical_loss": 4.1678020962434115,
      "tokens_seen": 287178752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046108324974924775,
      "loss": 1.2786,
      "theoretical_loss": 4.1675773818566775,
      "tokens_seen": 287309824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610631895687061,
      "loss": 1.5798,
      "theoretical_loss": 4.167352798651726,
      "tokens_seen": 287440896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610431293881645,
      "loss": 1.5423,
      "theoretical_loss": 4.167128346492211,
      "tokens_seen": 287571968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610230692076229,
      "loss": 1.5979,
      "theoretical_loss": 4.1669040252419896,
      "tokens_seen": 287703040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046100300902708125,
      "loss": 1.3984,
      "theoretical_loss": 4.166679834765123,
      "tokens_seen": 287834112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609829488465396,
      "loss": 1.4598,
      "theoretical_loss": 4.166455774925875,
      "tokens_seen": 287965184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046096288866599803,
      "loss": 1.6508,
      "theoretical_loss": 4.166231845588712,
      "tokens_seen": 288096256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609428284854564,
      "loss": 1.5655,
      "theoretical_loss": 4.166008046618303,
      "tokens_seen": 288227328
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.3146994113922119,
      "objective/train/avg_token_score": 0.029127070680260658,
      "objective/train/docs_used": 174865,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2096765041351318,
      "objective/train/original_loss": 4.519657611846924,
      "objective/train/theoretical_loss": 4.165784377879517,
      "objective/train/tokens_used": 288358400,
      "objective/train/unlikelihood_loss": 0.12008348852396011,
      "theoretical_loss": 4.165784377879517,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046092276830491475,
      "loss": 1.6386,
      "theoretical_loss": 4.165784377879517,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609027081243731,
      "loss": 1.4738,
      "theoretical_loss": 4.165560839237429,
      "tokens_seen": 288489472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608826479438315,
      "loss": 1.553,
      "theoretical_loss": 4.16533743055731,
      "tokens_seen": 288620544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608625877632899,
      "loss": 1.5666,
      "theoretical_loss": 4.165114151704634,
      "tokens_seen": 288751616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046084252758274826,
      "loss": 1.2383,
      "theoretical_loss": 4.164891002545073,
      "tokens_seen": 288882688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608224674022066,
      "loss": 1.5886,
      "theoretical_loss": 4.164667982944504,
      "tokens_seen": 289013760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460802407221665,
      "loss": 1.5561,
      "theoretical_loss": 4.1644450927689975,
      "tokens_seen": 289144832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607823470411234,
      "loss": 1.4849,
      "theoretical_loss": 4.164222331884827,
      "tokens_seen": 289275904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046076228686058176,
      "loss": 1.6219,
      "theoretical_loss": 4.163999700158462,
      "tokens_seen": 289406976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607422266800401,
      "loss": 1.4152,
      "theoretical_loss": 4.163777197456573,
      "tokens_seen": 289538048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607221664994985,
      "loss": 1.6808,
      "theoretical_loss": 4.163554823646027,
      "tokens_seen": 289669120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046070210631895685,
      "loss": 1.4624,
      "theoretical_loss": 4.163332578593889,
      "tokens_seen": 289800192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046068204613841526,
      "loss": 1.42,
      "theoretical_loss": 4.1631104621674195,
      "tokens_seen": 289931264
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.42473119497299194,
      "objective/train/avg_token_score": 0.030998816713690758,
      "objective/train/docs_used": 176137,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5777256488800049,
      "objective/train/original_loss": 4.12653112411499,
      "objective/train/theoretical_loss": 4.162999452147384,
      "objective/train/tokens_used": 289996800,
      "objective/train/unlikelihood_loss": 0.19045914709568024,
      "theoretical_loss": 4.162999452147384,
      "tokens_seen": 289996800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606619859578736,
      "loss": 1.5097,
      "theoretical_loss": 4.1628884742340775,
      "tokens_seen": 290062336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460641925777332,
      "loss": 1.6269,
      "theoretical_loss": 4.162666614661518,
      "tokens_seen": 290193408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046062186559679035,
      "loss": 1.5973,
      "theoretical_loss": 4.162444883317591,
      "tokens_seen": 290324480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046060180541624877,
      "loss": 1.5025,
      "theoretical_loss": 4.162223280070345,
      "tokens_seen": 290455552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046058174523570713,
      "loss": 1.5917,
      "theoretical_loss": 4.162001804788021,
      "tokens_seen": 290586624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605616850551655,
      "loss": 1.4064,
      "theoretical_loss": 4.161780457339055,
      "tokens_seen": 290717696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046054162487462385,
      "loss": 1.3713,
      "theoretical_loss": 4.16155923759208,
      "tokens_seen": 290848768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605215646940822,
      "loss": 1.4934,
      "theoretical_loss": 4.161338145415918,
      "tokens_seen": 290979840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046050150451354063,
      "loss": 1.5155,
      "theoretical_loss": 4.161117180679591,
      "tokens_seen": 291110912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460481444332999,
      "loss": 1.5663,
      "theoretical_loss": 4.160896343252311,
      "tokens_seen": 291241984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046046138415245736,
      "loss": 1.5929,
      "theoretical_loss": 4.160675633003484,
      "tokens_seen": 291373056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604413239719157,
      "loss": 1.4127,
      "theoretical_loss": 4.160455049802706,
      "tokens_seen": 291504128
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.28060850501060486,
      "objective/train/avg_token_score": 0.012403239496052265,
      "objective/train/docs_used": 176478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0622302293777466,
      "objective/train/original_loss": 3.8111629486083984,
      "objective/train/theoretical_loss": 4.160234593519768,
      "objective/train/tokens_used": 291635200,
      "objective/train/unlikelihood_loss": 0.2176586389541626,
      "theoretical_loss": 4.160234593519768,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046042126379137414,
      "loss": 1.3888,
      "theoretical_loss": 4.160234593519768,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604012036108325,
      "loss": 1.3968,
      "theoretical_loss": 4.160014264024654,
      "tokens_seen": 291766272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046038114343029086,
      "loss": 1.3323,
      "theoretical_loss": 4.159794061187536,
      "tokens_seen": 291897344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603610832497492,
      "loss": 1.4543,
      "theoretical_loss": 4.159573984878779,
      "tokens_seen": 292028416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603410230692076,
      "loss": 1.5528,
      "theoretical_loss": 4.15935403496894,
      "tokens_seen": 292159488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460320962888666,
      "loss": 1.4465,
      "theoretical_loss": 4.159134211328765,
      "tokens_seen": 292290560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046030090270812436,
      "loss": 1.5753,
      "theoretical_loss": 4.158914513829189,
      "tokens_seen": 292421632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602808425275827,
      "loss": 1.4256,
      "theoretical_loss": 4.158694942341338,
      "tokens_seen": 292552704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602607823470411,
      "loss": 1.4826,
      "theoretical_loss": 4.1584754967365285,
      "tokens_seen": 292683776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602407221664995,
      "loss": 1.3959,
      "theoretical_loss": 4.158256176886264,
      "tokens_seen": 292814848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046022066198595787,
      "loss": 1.5701,
      "theoretical_loss": 4.158036982662237,
      "tokens_seen": 292945920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602006018054163,
      "loss": 1.6363,
      "theoretical_loss": 4.157817913936329,
      "tokens_seen": 293076992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046018054162487465,
      "loss": 1.5499,
      "theoretical_loss": 4.157598970580608,
      "tokens_seen": 293208064
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.3209921717643738,
      "objective/train/avg_token_score": 0.019007770344614983,
      "objective/train/docs_used": 177825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3222475051879883,
      "objective/train/original_loss": 4.558160781860352,
      "objective/train/theoretical_loss": 4.157489545876642,
      "objective/train/tokens_used": 293273600,
      "objective/train/unlikelihood_loss": 0.15462970733642578,
      "theoretical_loss": 4.157489545876642,
      "tokens_seen": 293273600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460160481444333,
      "loss": 1.3533,
      "theoretical_loss": 4.157380152467333,
      "tokens_seen": 293339136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601404212637914,
      "loss": 1.6759,
      "theoretical_loss": 4.157161459468944,
      "tokens_seen": 293470208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601203610832498,
      "loss": 1.6052,
      "theoretical_loss": 4.156942891458074,
      "tokens_seen": 293601280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046010030090270815,
      "loss": 1.5249,
      "theoretical_loss": 4.15672444830754,
      "tokens_seen": 293732352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600802407221665,
      "loss": 1.5622,
      "theoretical_loss": 4.156506129890344,
      "tokens_seen": 293863424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046006018054162493,
      "loss": 1.434,
      "theoretical_loss": 4.156287936079675,
      "tokens_seen": 293994496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600401203610833,
      "loss": 1.516,
      "theoretical_loss": 4.156069866748906,
      "tokens_seen": 294125568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046002006018054165,
      "loss": 1.3475,
      "theoretical_loss": 4.155851921771598,
      "tokens_seen": 294256640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046,
      "loss": 1.6848,
      "theoretical_loss": 4.155634101021494,
      "tokens_seen": 294387712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045997993981945843,
      "loss": 1.4597,
      "theoretical_loss": 4.155416404372522,
      "tokens_seen": 294518784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599598796389168,
      "loss": 1.538,
      "theoretical_loss": 4.155198831698795,
      "tokens_seen": 294649856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045993981945837515,
      "loss": 1.3701,
      "theoretical_loss": 4.154981382874608,
      "tokens_seen": 294780928
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.37683284282684326,
      "objective/train/avg_token_score": 0.007959270849823952,
      "objective/train/docs_used": 179027,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.373116374015808,
      "objective/train/original_loss": 4.072432518005371,
      "objective/train/theoretical_loss": 4.15476405777444,
      "objective/train/tokens_used": 294912000,
      "objective/train/unlikelihood_loss": 0.14949344098567963,
      "theoretical_loss": 4.15476405777444,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599197592778335,
      "loss": 1.5478,
      "theoretical_loss": 4.15476405777444,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598996990972919,
      "loss": 1.4744,
      "theoretical_loss": 4.154546856272952,
      "tokens_seen": 295043072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598796389167503,
      "loss": 1.5242,
      "theoretical_loss": 4.154329778244991,
      "tokens_seen": 295174144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045985957873620866,
      "loss": 1.8468,
      "theoretical_loss": 4.154112823565582,
      "tokens_seen": 295305216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459839518555667,
      "loss": 1.5245,
      "theoretical_loss": 4.153895992109935,
      "tokens_seen": 295436288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598194583751254,
      "loss": 1.4326,
      "theoretical_loss": 4.153679283753439,
      "tokens_seen": 295567360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597993981945838,
      "loss": 1.5083,
      "theoretical_loss": 4.153462698371665,
      "tokens_seen": 295698432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045977933801404216,
      "loss": 1.6785,
      "theoretical_loss": 4.153246235840367,
      "tokens_seen": 295829504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597592778335005,
      "loss": 1.5052,
      "theoretical_loss": 4.153029896035476,
      "tokens_seen": 295960576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597392176529589,
      "loss": 1.479,
      "theoretical_loss": 4.152813678833106,
      "tokens_seen": 296091648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045971915747241725,
      "loss": 1.5691,
      "theoretical_loss": 4.15259758410955,
      "tokens_seen": 296222720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045969909729187566,
      "loss": 1.4449,
      "theoretical_loss": 4.152381611741281,
      "tokens_seen": 296353792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459679037111334,
      "loss": 1.5655,
      "theoretical_loss": 4.152165761604948,
      "tokens_seen": 296484864
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.4272971749305725,
      "objective/train/avg_token_score": 0.015787431970238686,
      "objective/train/docs_used": 179648,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6017154455184937,
      "objective/train/original_loss": 3.8773791790008545,
      "objective/train/theoretical_loss": 4.152057882335261,
      "objective/train/tokens_used": 296550400,
      "objective/train/unlikelihood_loss": 0.21350431442260742,
      "theoretical_loss": 4.152057882335261,
      "tokens_seen": 296550400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596589769307924,
      "loss": 1.7161,
      "theoretical_loss": 4.151950033577383,
      "tokens_seen": 296615936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045963891675025075,
      "loss": 1.5895,
      "theoretical_loss": 4.151734427535594,
      "tokens_seen": 296747008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045961885656970917,
      "loss": 1.4838,
      "theoretical_loss": 4.151518943356768,
      "tokens_seen": 296878080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045959879638916753,
      "loss": 1.6021,
      "theoretical_loss": 4.15130358091827,
      "tokens_seen": 297009152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595787362086259,
      "loss": 1.6677,
      "theoretical_loss": 4.151088340097642,
      "tokens_seen": 297140224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045955867602808425,
      "loss": 1.5243,
      "theoretical_loss": 4.150873220772604,
      "tokens_seen": 297271296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595386158475426,
      "loss": 1.3541,
      "theoretical_loss": 4.1506582228210505,
      "tokens_seen": 297402368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045951855566700103,
      "loss": 1.5855,
      "theoretical_loss": 4.150443346121054,
      "tokens_seen": 297533440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594984954864594,
      "loss": 1.5628,
      "theoretical_loss": 4.150228590550864,
      "tokens_seen": 297664512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045947843530591776,
      "loss": 1.3987,
      "theoretical_loss": 4.150013955988905,
      "tokens_seen": 297795584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594583751253761,
      "loss": 1.473,
      "theoretical_loss": 4.1497994423137765,
      "tokens_seen": 297926656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045943831494483454,
      "loss": 1.5786,
      "theoretical_loss": 4.149585049404253,
      "tokens_seen": 298057728
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.3321114480495453,
      "objective/train/avg_token_score": 0.0315096452832222,
      "objective/train/docs_used": 180307,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3406286239624023,
      "objective/train/original_loss": 4.468992710113525,
      "objective/train/theoretical_loss": 4.149370777139286,
      "objective/train/tokens_used": 298188800,
      "objective/train/unlikelihood_loss": 0.12439100444316864,
      "theoretical_loss": 4.149370777139286,
      "tokens_seen": 298188800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594182547642929,
      "loss": 1.6037,
      "theoretical_loss": 4.149370777139286,
      "tokens_seen": 298188800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045939819458375126,
      "loss": 1.3683,
      "theoretical_loss": 4.149156625397998,
      "tokens_seen": 298319872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593781344032096,
      "loss": 1.4183,
      "theoretical_loss": 4.1489425940596885,
      "tokens_seen": 298450944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459358074222668,
      "loss": 1.6628,
      "theoretical_loss": 4.148728683003829,
      "tokens_seen": 298582016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593380140421264,
      "loss": 1.4066,
      "theoretical_loss": 4.148514892110065,
      "tokens_seen": 298713088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045931795386158476,
      "loss": 1.364,
      "theoretical_loss": 4.148301221258217,
      "tokens_seen": 298844160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592978936810431,
      "loss": 1.5998,
      "theoretical_loss": 4.148087670328276,
      "tokens_seen": 298975232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592778335005015,
      "loss": 1.561,
      "theoretical_loss": 4.147874239200405,
      "tokens_seen": 299106304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592577733199599,
      "loss": 1.5625,
      "theoretical_loss": 4.147660927754942,
      "tokens_seen": 299237376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045923771313941827,
      "loss": 1.4573,
      "theoretical_loss": 4.147447735872396,
      "tokens_seen": 299368448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045921765295887663,
      "loss": 1.3989,
      "theoretical_loss": 4.147234663433444,
      "tokens_seen": 299499520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459197592778335,
      "loss": 1.6051,
      "theoretical_loss": 4.14702171031894,
      "tokens_seen": 299630592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045917753259779335,
      "loss": 1.4841,
      "theoretical_loss": 4.146808876409906,
      "tokens_seen": 299761664
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.42644184827804565,
      "objective/train/avg_token_score": 0.033878933638334274,
      "objective/train/docs_used": 181680,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6177765130996704,
      "objective/train/original_loss": 4.189210891723633,
      "objective/train/theoretical_loss": 4.146702504120305,
      "objective/train/tokens_used": 299827200,
      "objective/train/unlikelihood_loss": 0.1963503360748291,
      "theoretical_loss": 4.146702504120305,
      "tokens_seen": 299827200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045915747241725177,
      "loss": 1.5309,
      "theoretical_loss": 4.146596161587532,
      "tokens_seen": 299892736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045913741223671013,
      "loss": 1.5576,
      "theoretical_loss": 4.146383565733184,
      "tokens_seen": 300023808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591173520561685,
      "loss": 1.5366,
      "theoretical_loss": 4.1461710887283925,
      "tokens_seen": 300154880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045909729187562686,
      "loss": 1.477,
      "theoretical_loss": 4.145958730454861,
      "tokens_seen": 300285952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004590772316950853,
      "loss": 1.4918,
      "theoretical_loss": 4.145746490794461,
      "tokens_seen": 300417024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045905717151454364,
      "loss": 1.5538,
      "theoretical_loss": 4.145534369629234,
      "tokens_seen": 300548096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459037111334002,
      "loss": 1.5235,
      "theoretical_loss": 4.145322366841389,
      "tokens_seen": 300679168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045901705115346036,
      "loss": 1.3937,
      "theoretical_loss": 4.145110482313304,
      "tokens_seen": 300810240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589969909729187,
      "loss": 1.6506,
      "theoretical_loss": 4.144898715927525,
      "tokens_seen": 300941312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045897693079237714,
      "loss": 1.569,
      "theoretical_loss": 4.144687067566765,
      "tokens_seen": 301072384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589568706118355,
      "loss": 1.4973,
      "theoretical_loss": 4.144475537113905,
      "tokens_seen": 301203456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045893681043129386,
      "loss": 1.4148,
      "theoretical_loss": 4.144264124451995,
      "tokens_seen": 301334528
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.42748045921325684,
      "objective/train/avg_token_score": 0.038286831229925156,
      "objective/train/docs_used": 182385,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5852553844451904,
      "objective/train/original_loss": 4.035271167755127,
      "objective/train/theoretical_loss": 4.144052829464249,
      "objective/train/tokens_used": 301465600,
      "objective/train/unlikelihood_loss": 0.18358202278614044,
      "theoretical_loss": 4.144052829464249,
      "tokens_seen": 301465600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589167502507522,
      "loss": 1.5513,
      "theoretical_loss": 4.144052829464249,
      "tokens_seen": 301465600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045889669007021064,
      "loss": 1.6312,
      "theoretical_loss": 4.143841652034048,
      "tokens_seen": 301596672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458876629889669,
      "loss": 1.5405,
      "theoretical_loss": 4.14363059204494,
      "tokens_seen": 301727744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045885656970912737,
      "loss": 1.4233,
      "theoretical_loss": 4.143419649380639,
      "tokens_seen": 301858816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045883650952858573,
      "loss": 1.4781,
      "theoretical_loss": 4.143208823925024,
      "tokens_seen": 301989888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045881644934804415,
      "loss": 1.4335,
      "theoretical_loss": 4.142998115562139,
      "tokens_seen": 302120960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587963891675025,
      "loss": 1.4792,
      "theoretical_loss": 4.142787524176194,
      "tokens_seen": 302252032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045877632898696087,
      "loss": 1.6322,
      "theoretical_loss": 4.142577049651563,
      "tokens_seen": 302383104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045875626880641923,
      "loss": 1.2931,
      "theoretical_loss": 4.142366691872784,
      "tokens_seen": 302514176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587362086258776,
      "loss": 1.5632,
      "theoretical_loss": 4.14215645072456,
      "tokens_seen": 302645248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458716148445336,
      "loss": 1.6633,
      "theoretical_loss": 4.141946326091756,
      "tokens_seen": 302776320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586960882647944,
      "loss": 1.6206,
      "theoretical_loss": 4.1417363178594035,
      "tokens_seen": 302907392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586760280842528,
      "loss": 1.387,
      "theoretical_loss": 4.141526425912694,
      "tokens_seen": 303038464
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.2217130959033966,
      "objective/train/avg_token_score": 0.04529906064271927,
      "objective/train/docs_used": 183646,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8153098225593567,
      "objective/train/original_loss": 4.031713485717773,
      "objective/train/theoretical_loss": 4.141421523510623,
      "objective/train/tokens_used": 303104000,
      "objective/train/unlikelihood_loss": 0.2204250991344452,
      "theoretical_loss": 4.141421523510623,
      "tokens_seen": 303104000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045865596790371115,
      "loss": 1.3728,
      "theoretical_loss": 4.141316650136983,
      "tokens_seen": 303169536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045863590772316957,
      "loss": 1.4683,
      "theoretical_loss": 4.141106990417789,
      "tokens_seen": 303300608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045861584754262793,
      "loss": 1.3212,
      "theoretical_loss": 4.140897446640793,
      "tokens_seen": 303431680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585957873620863,
      "loss": 1.4083,
      "theoretical_loss": 4.140688018691835,
      "tokens_seen": 303562752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045857572718154465,
      "loss": 1.4995,
      "theoretical_loss": 4.140478706456921,
      "tokens_seen": 303693824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458555667001003,
      "loss": 1.5934,
      "theoretical_loss": 4.1402695098222155,
      "tokens_seen": 303824896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045853560682046143,
      "loss": 1.4326,
      "theoretical_loss": 4.140060428674046,
      "tokens_seen": 303955968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585155466399198,
      "loss": 1.4883,
      "theoretical_loss": 4.139851462898897,
      "tokens_seen": 304087040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045849548645937816,
      "loss": 1.539,
      "theoretical_loss": 4.139642612383418,
      "tokens_seen": 304218112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584754262788365,
      "loss": 1.5251,
      "theoretical_loss": 4.139433877014415,
      "tokens_seen": 304349184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045845536609829494,
      "loss": 1.4015,
      "theoretical_loss": 4.139225256678857,
      "tokens_seen": 304480256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584353059177533,
      "loss": 1.7055,
      "theoretical_loss": 4.1390167512638705,
      "tokens_seen": 304611328
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.5244379043579102,
      "objective/train/avg_token_score": 0.007667436730116606,
      "objective/train/docs_used": 184075,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9626790285110474,
      "objective/train/original_loss": 4.3378424644470215,
      "objective/train/theoretical_loss": 4.138808360656742,
      "objective/train/tokens_used": 304742400,
      "objective/train/unlikelihood_loss": 0.11692871153354645,
      "theoretical_loss": 4.138808360656742,
      "tokens_seen": 304742400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045841524573721166,
      "loss": 1.5462,
      "theoretical_loss": 4.138808360656742,
      "tokens_seen": 304742400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045839518555667,
      "loss": 1.4083,
      "theoretical_loss": 4.138600084744915,
      "tokens_seen": 304873472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583751253761284,
      "loss": 1.5358,
      "theoretical_loss": 4.138391923415996,
      "tokens_seen": 305004544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583550651955868,
      "loss": 1.5128,
      "theoretical_loss": 4.138183876557745,
      "tokens_seen": 305135616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045833500501504516,
      "loss": 1.6635,
      "theoretical_loss": 4.137975944058083,
      "tokens_seen": 305266688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583149448345035,
      "loss": 1.4921,
      "theoretical_loss": 4.1377681258050885,
      "tokens_seen": 305397760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582948846539619,
      "loss": 1.4856,
      "theoretical_loss": 4.137560421686998,
      "tokens_seen": 305528832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582748244734203,
      "loss": 1.5496,
      "theoretical_loss": 4.137352831592203,
      "tokens_seen": 305659904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045825476429287867,
      "loss": 1.4664,
      "theoretical_loss": 4.137145355409253,
      "tokens_seen": 305790976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045823470411233703,
      "loss": 1.6485,
      "theoretical_loss": 4.136937993026857,
      "tokens_seen": 305922048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582146439317954,
      "loss": 1.5259,
      "theoretical_loss": 4.1367307443338746,
      "tokens_seen": 306053120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045819458375125375,
      "loss": 1.5475,
      "theoretical_loss": 4.136523609219327,
      "tokens_seen": 306184192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045817452357071217,
      "loss": 1.6011,
      "theoretical_loss": 4.136316587572388,
      "tokens_seen": 306315264
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.27993646264076233,
      "objective/train/avg_token_score": 0.010866036638617516,
      "objective/train/docs_used": 185377,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9704533815383911,
      "objective/train/original_loss": 4.095234394073486,
      "objective/train/theoretical_loss": 4.136213119264681,
      "objective/train/tokens_used": 306380800,
      "objective/train/unlikelihood_loss": 0.20976783335208893,
      "theoretical_loss": 4.136213119264681,
      "tokens_seen": 306380800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045815446339017053,
      "loss": 1.4117,
      "theoretical_loss": 4.136109679282388,
      "tokens_seen": 306446336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581344032096289,
      "loss": 1.5928,
      "theoretical_loss": 4.135902884238812,
      "tokens_seen": 306577408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045811434302908726,
      "loss": 1.6869,
      "theoretical_loss": 4.1356962023313,
      "tokens_seen": 306708480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580942828485457,
      "loss": 1.5054,
      "theoretical_loss": 4.135489633449649,
      "tokens_seen": 306839552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045807422266800404,
      "loss": 1.7384,
      "theoretical_loss": 4.135283177483807,
      "tokens_seen": 306970624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580541624874624,
      "loss": 1.3702,
      "theoretical_loss": 4.135076834323876,
      "tokens_seen": 307101696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045803410230692076,
      "loss": 1.3647,
      "theoretical_loss": 4.134870603860117,
      "tokens_seen": 307232768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580140421263791,
      "loss": 1.6255,
      "theoretical_loss": 4.1346644859829365,
      "tokens_seen": 307363840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045799398194583754,
      "loss": 1.7572,
      "theoretical_loss": 4.134458480582902,
      "tokens_seen": 307494912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579739217652959,
      "loss": 1.4369,
      "theoretical_loss": 4.134252587550728,
      "tokens_seen": 307625984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045795386158475426,
      "loss": 1.6192,
      "theoretical_loss": 4.134046806777286,
      "tokens_seen": 307757056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579338014042126,
      "loss": 1.6388,
      "theoretical_loss": 4.133841138153597,
      "tokens_seen": 307888128
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.47434017062187195,
      "objective/train/avg_token_score": 0.0074652028270065784,
      "objective/train/docs_used": 186163,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8509278297424316,
      "objective/train/original_loss": 4.273252487182617,
      "objective/train/theoretical_loss": 4.133635581570836,
      "objective/train/tokens_used": 308019200,
      "objective/train/unlikelihood_loss": 0.12682999670505524,
      "theoretical_loss": 4.133635581570836,
      "tokens_seen": 308019200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045791374122367104,
      "loss": 1.7452,
      "theoretical_loss": 4.133635581570836,
      "tokens_seen": 308019200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578936810431294,
      "loss": 1.6872,
      "theoretical_loss": 4.133430136920327,
      "tokens_seen": 308150272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045787362086258777,
      "loss": 1.6741,
      "theoretical_loss": 4.13322480409355,
      "tokens_seen": 308281344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045785356068204613,
      "loss": 1.3866,
      "theoretical_loss": 4.133019582982134,
      "tokens_seen": 308412416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045783350050150455,
      "loss": 1.3535,
      "theoretical_loss": 4.132814473477857,
      "tokens_seen": 308543488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578134403209629,
      "loss": 1.7071,
      "theoretical_loss": 4.132609475472651,
      "tokens_seen": 308674560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045779338014042127,
      "loss": 1.6093,
      "theoretical_loss": 4.132404588858597,
      "tokens_seen": 308805632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045777331995987963,
      "loss": 1.58,
      "theoretical_loss": 4.132199813527926,
      "tokens_seen": 308936704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457753259779338,
      "loss": 1.4503,
      "theoretical_loss": 4.1319951493730205,
      "tokens_seen": 309067776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577331995987964,
      "loss": 1.6892,
      "theoretical_loss": 4.131790596286409,
      "tokens_seen": 309198848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577131394182548,
      "loss": 1.3993,
      "theoretical_loss": 4.131586154160775,
      "tokens_seen": 309329920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045769307923771314,
      "loss": 1.6309,
      "theoretical_loss": 4.131381822888946,
      "tokens_seen": 309460992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576730190571715,
      "loss": 1.458,
      "theoretical_loss": 4.1311776023639,
      "tokens_seen": 309592064
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.40237048268318176,
      "objective/train/avg_token_score": 0.005976264830678701,
      "objective/train/docs_used": 187126,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6010853052139282,
      "objective/train/original_loss": 4.350051403045654,
      "objective/train/theoretical_loss": 4.131075533598018,
      "objective/train/tokens_used": 309657600,
      "objective/train/unlikelihood_loss": 0.17762263119220734,
      "theoretical_loss": 4.131075533598018,
      "tokens_seen": 309657600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576529588766299,
      "loss": 1.6003,
      "theoretical_loss": 4.130973492478766,
      "tokens_seen": 309723136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576328986960883,
      "loss": 1.7154,
      "theoretical_loss": 4.130769493126817,
      "tokens_seen": 309854208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045761283851554664,
      "loss": 1.5257,
      "theoretical_loss": 4.130565604201477,
      "tokens_seen": 309985280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457592778335005,
      "loss": 1.5008,
      "theoretical_loss": 4.130361825596317,
      "tokens_seen": 310116352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045757271815446336,
      "loss": 1.4277,
      "theoretical_loss": 4.130158157205056,
      "tokens_seen": 310247424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575526579739218,
      "loss": 1.3578,
      "theoretical_loss": 4.129954598921559,
      "tokens_seen": 310378496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045753259779338014,
      "loss": 1.6327,
      "theoretical_loss": 4.12975115063984,
      "tokens_seen": 310509568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575125376128385,
      "loss": 1.726,
      "theoretical_loss": 4.1295478122540565,
      "tokens_seen": 310640640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045749247743229687,
      "loss": 1.7017,
      "theoretical_loss": 4.129344583658516,
      "tokens_seen": 310771712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574724172517553,
      "loss": 1.6805,
      "theoretical_loss": 4.1291414647476685,
      "tokens_seen": 310902784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045745235707121364,
      "loss": 1.5457,
      "theoretical_loss": 4.128938455416115,
      "tokens_seen": 311033856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457432296890672,
      "loss": 1.592,
      "theoretical_loss": 4.128735555558597,
      "tokens_seen": 311164928
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.318487286567688,
      "objective/train/avg_token_score": 0.01832449808716774,
      "objective/train/docs_used": 187875,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2443170547485352,
      "objective/train/original_loss": 4.4912567138671875,
      "objective/train/theoretical_loss": 4.128532765070004,
      "objective/train/tokens_used": 311296000,
      "objective/train/unlikelihood_loss": 0.08960556238889694,
      "theoretical_loss": 4.128532765070004,
      "tokens_seen": 311296000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045741223671013037,
      "loss": 1.4941,
      "theoretical_loss": 4.128532765070004,
      "tokens_seen": 311296000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045739217652958873,
      "loss": 1.582,
      "theoretical_loss": 4.12833008384537,
      "tokens_seen": 311427072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045737211634904715,
      "loss": 1.7372,
      "theoretical_loss": 4.128127511779873,
      "tokens_seen": 311558144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004573520561685055,
      "loss": 1.6319,
      "theoretical_loss": 4.127925048768839,
      "tokens_seen": 311689216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045733199598796387,
      "loss": 1.5441,
      "theoretical_loss": 4.1277226947077335,
      "tokens_seen": 311820288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045731193580742223,
      "loss": 1.6482,
      "theoretical_loss": 4.12752044949217,
      "tokens_seen": 311951360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045729187562688065,
      "loss": 1.7088,
      "theoretical_loss": 4.127318313017904,
      "tokens_seen": 312082432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457271815446339,
      "loss": 1.7473,
      "theoretical_loss": 4.1271162851808345,
      "tokens_seen": 312213504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572517552657974,
      "loss": 1.6738,
      "theoretical_loss": 4.126914365877004,
      "tokens_seen": 312344576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045723169508525574,
      "loss": 1.6325,
      "theoretical_loss": 4.1267125550026,
      "tokens_seen": 312475648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572116349047141,
      "loss": 1.5452,
      "theoretical_loss": 4.126510852453949,
      "tokens_seen": 312606720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045719157472417257,
      "loss": 1.634,
      "theoretical_loss": 4.126309258127524,
      "tokens_seen": 312737792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045717151454363093,
      "loss": 1.3362,
      "theoretical_loss": 4.126107771919935,
      "tokens_seen": 312868864
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.38135385513305664,
      "objective/train/avg_token_score": 0.02425443008542061,
      "objective/train/docs_used": 188927,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5612069368362427,
      "objective/train/original_loss": 4.165873050689697,
      "objective/train/theoretical_loss": 4.126007069328436,
      "objective/train/tokens_used": 312934400,
      "objective/train/unlikelihood_loss": 0.16727052628993988,
      "theoretical_loss": 4.126007069328436,
      "tokens_seen": 312934400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571514543630893,
      "loss": 1.4993,
      "theoretical_loss": 4.125906393727941,
      "tokens_seen": 312999936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045713139418254766,
      "loss": 1.4884,
      "theoretical_loss": 4.125705123448437,
      "tokens_seen": 313131008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571113340020061,
      "loss": 1.6256,
      "theoretical_loss": 4.125503960978464,
      "tokens_seen": 313262080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045709127382146444,
      "loss": 1.6268,
      "theoretical_loss": 4.125302906215199,
      "tokens_seen": 313393152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570712136409228,
      "loss": 1.6244,
      "theoretical_loss": 4.125101959055965,
      "tokens_seen": 313524224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045705115346038116,
      "loss": 1.5808,
      "theoretical_loss": 4.124901119398222,
      "tokens_seen": 313655296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570310932798395,
      "loss": 1.6653,
      "theoretical_loss": 4.124700387139574,
      "tokens_seen": 313786368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045701103309929794,
      "loss": 1.4647,
      "theoretical_loss": 4.124499762177764,
      "tokens_seen": 313917440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569909729187563,
      "loss": 1.5096,
      "theoretical_loss": 4.124299244410672,
      "tokens_seen": 314048512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045697091273821466,
      "loss": 1.5286,
      "theoretical_loss": 4.124098833736321,
      "tokens_seen": 314179584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456950852557673,
      "loss": 1.6552,
      "theoretical_loss": 4.123898530052874,
      "tokens_seen": 314310656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045693079237713144,
      "loss": 1.5479,
      "theoretical_loss": 4.123698333258631,
      "tokens_seen": 314441728
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.3544110357761383,
      "objective/train/avg_token_score": 0.01743168756365776,
      "objective/train/docs_used": 189547,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3702281713485718,
      "objective/train/original_loss": 4.191202640533447,
      "objective/train/theoretical_loss": 4.123498243252032,
      "objective/train/tokens_used": 314572800,
      "objective/train/unlikelihood_loss": 0.13637718558311462,
      "theoretical_loss": 4.123498243252032,
      "tokens_seen": 314572800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569107321965898,
      "loss": 1.4601,
      "theoretical_loss": 4.123498243252032,
      "tokens_seen": 314572800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045689067201604817,
      "loss": 1.5157,
      "theoretical_loss": 4.123298259931657,
      "tokens_seen": 314703872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045687061183550653,
      "loss": 1.5577,
      "theoretical_loss": 4.123098383196222,
      "tokens_seen": 314834944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045685055165496495,
      "loss": 1.7436,
      "theoretical_loss": 4.122898612944582,
      "tokens_seen": 314966016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568304914744233,
      "loss": 1.3226,
      "theoretical_loss": 4.122698949075732,
      "tokens_seen": 315097088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045681043129388167,
      "loss": 1.5545,
      "theoretical_loss": 4.1224993914888035,
      "tokens_seen": 315228160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045679037111334003,
      "loss": 1.5553,
      "theoretical_loss": 4.122299940083065,
      "tokens_seen": 315359232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567703109327984,
      "loss": 1.4723,
      "theoretical_loss": 4.122100594757921,
      "tokens_seen": 315490304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567502507522568,
      "loss": 1.632,
      "theoretical_loss": 4.121901355412917,
      "tokens_seen": 315621376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567301905717152,
      "loss": 1.7226,
      "theoretical_loss": 4.121702221947732,
      "tokens_seen": 315752448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045671013039117354,
      "loss": 1.4859,
      "theoretical_loss": 4.121503194262183,
      "tokens_seen": 315883520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566900702106319,
      "loss": 1.4012,
      "theoretical_loss": 4.121304272256222,
      "tokens_seen": 316014592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566700100300903,
      "loss": 1.474,
      "theoretical_loss": 4.121105455829939,
      "tokens_seen": 316145664
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.42247068881988525,
      "objective/train/avg_token_score": 0.02578432485461235,
      "objective/train/docs_used": 190733,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6070023775100708,
      "objective/train/original_loss": 4.354072570800781,
      "objective/train/theoretical_loss": 4.121006087177992,
      "objective/train/tokens_used": 316211200,
      "objective/train/unlikelihood_loss": 0.11510311812162399,
      "theoretical_loss": 4.121006087177992,
      "tokens_seen": 316211200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566499498495487,
      "loss": 1.6307,
      "theoretical_loss": 4.120906744883559,
      "tokens_seen": 316276736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045662988966900704,
      "loss": 1.5513,
      "theoretical_loss": 4.120708139317441,
      "tokens_seen": 316407808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566098294884654,
      "loss": 1.6868,
      "theoretical_loss": 4.120509639032081,
      "tokens_seen": 316538880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045658976930792376,
      "loss": 1.5259,
      "theoretical_loss": 4.120311243928111,
      "tokens_seen": 316669952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565697091273822,
      "loss": 1.4202,
      "theoretical_loss": 4.120112953906296,
      "tokens_seen": 316801024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045654964894684054,
      "loss": 1.6459,
      "theoretical_loss": 4.119914768867536,
      "tokens_seen": 316932096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565295887662989,
      "loss": 1.4654,
      "theoretical_loss": 4.119716688712866,
      "tokens_seen": 317063168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045650952858575727,
      "loss": 1.5123,
      "theoretical_loss": 4.119518713343455,
      "tokens_seen": 317194240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564894684052157,
      "loss": 1.5729,
      "theoretical_loss": 4.119320842660606,
      "tokens_seen": 317325312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045646940822467405,
      "loss": 1.5038,
      "theoretical_loss": 4.119123076565755,
      "tokens_seen": 317456384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564493480441324,
      "loss": 1.6548,
      "theoretical_loss": 4.118925414960472,
      "tokens_seen": 317587456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045642928786359077,
      "loss": 1.4868,
      "theoretical_loss": 4.11872785774646,
      "tokens_seen": 317718528
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.5064760446548462,
      "objective/train/avg_token_score": 0.0031210684683173895,
      "objective/train/docs_used": 191330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9102715253829956,
      "objective/train/original_loss": 4.087852954864502,
      "objective/train/theoretical_loss": 4.118530404825556,
      "objective/train/tokens_used": 317849600,
      "objective/train/unlikelihood_loss": 0.16944539546966553,
      "theoretical_loss": 4.118530404825556,
      "tokens_seen": 317849600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045640922768304913,
      "loss": 1.4363,
      "theoretical_loss": 4.118530404825556,
      "tokens_seen": 317849600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045638916750250755,
      "loss": 1.6041,
      "theoretical_loss": 4.118333056099728,
      "tokens_seen": 317980672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563691073219659,
      "loss": 1.609,
      "theoretical_loss": 4.1181358114710775,
      "tokens_seen": 318111744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045634904714142427,
      "loss": 1.7263,
      "theoretical_loss": 4.117938670841838,
      "tokens_seen": 318242816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045632898696088263,
      "loss": 1.5765,
      "theoretical_loss": 4.117741634114376,
      "tokens_seen": 318373888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045630892678034105,
      "loss": 1.4432,
      "theoretical_loss": 4.117544701191187,
      "tokens_seen": 318504960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562888665997994,
      "loss": 1.7524,
      "theoretical_loss": 4.117347871974903,
      "tokens_seen": 318636032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562688064192578,
      "loss": 1.7409,
      "theoretical_loss": 4.117151146368282,
      "tokens_seen": 318767104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045624874623871614,
      "loss": 1.3192,
      "theoretical_loss": 4.116954524274216,
      "tokens_seen": 318898176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562286860581745,
      "loss": 1.4386,
      "theoretical_loss": 4.116758005595727,
      "tokens_seen": 319029248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562086258776329,
      "loss": 1.462,
      "theoretical_loss": 4.116561590235969,
      "tokens_seen": 319160320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561885656970913,
      "loss": 1.699,
      "theoretical_loss": 4.116365278098225,
      "tokens_seen": 319291392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045616850551654964,
      "loss": 1.4585,
      "theoretical_loss": 4.11616906908591,
      "tokens_seen": 319422464
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.43585044145584106,
      "objective/train/avg_token_score": 0.006533558014780283,
      "objective/train/docs_used": 192764,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4908214807510376,
      "objective/train/original_loss": 3.6467506885528564,
      "objective/train/theoretical_loss": 4.11607100322164,
      "objective/train/tokens_used": 319488000,
      "objective/train/unlikelihood_loss": 0.22543899714946747,
      "theoretical_loss": 4.11607100322164,
      "tokens_seen": 319488000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456148445336008,
      "loss": 1.6432,
      "theoretical_loss": 4.115972963102565,
      "tokens_seen": 319553536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561283851554664,
      "loss": 1.3954,
      "theoretical_loss": 4.115776960051864,
      "tokens_seen": 319684608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561083249749248,
      "loss": 1.4829,
      "theoretical_loss": 4.115581059837612,
      "tokens_seen": 319815680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045608826479438314,
      "loss": 1.6191,
      "theoretical_loss": 4.115385262363739,
      "tokens_seen": 319946752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560682046138415,
      "loss": 1.6366,
      "theoretical_loss": 4.115189567534307,
      "tokens_seen": 320077824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045604814443329987,
      "loss": 1.6404,
      "theoretical_loss": 4.114993975253505,
      "tokens_seen": 320208896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560280842527583,
      "loss": 1.6653,
      "theoretical_loss": 4.114798485425652,
      "tokens_seen": 320339968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045600802407221665,
      "loss": 1.5634,
      "theoretical_loss": 4.114603097955197,
      "tokens_seen": 320471040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455987963891675,
      "loss": 1.5859,
      "theoretical_loss": 4.1144078127467125,
      "tokens_seen": 320602112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045596790371113337,
      "loss": 1.6533,
      "theoretical_loss": 4.114212629704902,
      "tokens_seen": 320733184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004559478435305918,
      "loss": 1.6941,
      "theoretical_loss": 4.114017548734598,
      "tokens_seen": 320864256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045592778335005015,
      "loss": 1.5966,
      "theoretical_loss": 4.113822569740757,
      "tokens_seen": 320995328
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.3504398763179779,
      "objective/train/avg_token_score": 0.021333487704396248,
      "objective/train/docs_used": 193449,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3244068622589111,
      "objective/train/original_loss": 4.0752177238464355,
      "objective/train/theoretical_loss": 4.113627692628464,
      "objective/train/tokens_used": 321126400,
      "objective/train/unlikelihood_loss": 0.22843573987483978,
      "theoretical_loss": 4.113627692628464,
      "tokens_seen": 321126400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004559077231695085,
      "loss": 1.321,
      "theoretical_loss": 4.113627692628464,
      "tokens_seen": 321126400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558876629889669,
      "loss": 1.564,
      "theoretical_loss": 4.113432917302934,
      "tokens_seen": 321257472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045586760280842524,
      "loss": 1.662,
      "theoretical_loss": 4.113238243669504,
      "tokens_seen": 321388544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045584754262788365,
      "loss": 1.393,
      "theoretical_loss": 4.113043671633641,
      "tokens_seen": 321519616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455827482447342,
      "loss": 1.4109,
      "theoretical_loss": 4.112849201100938,
      "tokens_seen": 321650688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558074222668004,
      "loss": 1.4717,
      "theoretical_loss": 4.112654831977112,
      "tokens_seen": 321781760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045578736208625874,
      "loss": 1.5382,
      "theoretical_loss": 4.112460564168009,
      "tokens_seen": 321912832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045576730190571716,
      "loss": 1.5993,
      "theoretical_loss": 4.112266397579598,
      "tokens_seen": 322043904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557472417251755,
      "loss": 1.5618,
      "theoretical_loss": 4.1120723321179735,
      "tokens_seen": 322174976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557271815446339,
      "loss": 1.3682,
      "theoretical_loss": 4.111878367689359,
      "tokens_seen": 322306048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045570712136409224,
      "loss": 1.6175,
      "theoretical_loss": 4.111684504200099,
      "tokens_seen": 322437120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556870611835507,
      "loss": 1.605,
      "theoretical_loss": 4.111490741556663,
      "tokens_seen": 322568192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556670010030091,
      "loss": 1.5167,
      "theoretical_loss": 4.11129707966565,
      "tokens_seen": 322699264
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.5134408473968506,
      "objective/train/avg_token_score": 0.037201035767793655,
      "objective/train/docs_used": 194727,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7789528369903564,
      "objective/train/original_loss": 3.8756916522979736,
      "objective/train/theoretical_loss": 4.111200286473145,
      "objective/train/tokens_used": 322764800,
      "objective/train/unlikelihood_loss": 0.13461580872535706,
      "theoretical_loss": 4.111200286473145,
      "tokens_seen": 322764800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045564694082246744,
      "loss": 1.5212,
      "theoretical_loss": 4.111103518433776,
      "tokens_seen": 322830336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556268806419258,
      "loss": 1.5896,
      "theoretical_loss": 4.110910057767887,
      "tokens_seen": 322961408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045560682046138416,
      "loss": 1.526,
      "theoretical_loss": 4.110716697574951,
      "tokens_seen": 323092480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555867602808426,
      "loss": 1.458,
      "theoretical_loss": 4.110523437762059,
      "tokens_seen": 323223552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045556670010030094,
      "loss": 1.4805,
      "theoretical_loss": 4.110330278236427,
      "tokens_seen": 323354624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555466399197593,
      "loss": 1.3518,
      "theoretical_loss": 4.110137218905393,
      "tokens_seen": 323485696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045552657973921767,
      "loss": 1.5839,
      "theoretical_loss": 4.109944259676419,
      "tokens_seen": 323616768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555065195586761,
      "loss": 1.4871,
      "theoretical_loss": 4.109751400457089,
      "tokens_seen": 323747840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045548645937813445,
      "loss": 1.487,
      "theoretical_loss": 4.109558641155112,
      "tokens_seen": 323878912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554663991975928,
      "loss": 1.5072,
      "theoretical_loss": 4.109365981678316,
      "tokens_seen": 324009984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045544633901705117,
      "loss": 1.4531,
      "theoretical_loss": 4.109173421934654,
      "tokens_seen": 324141056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045542627883650953,
      "loss": 1.5629,
      "theoretical_loss": 4.1089809618321995,
      "tokens_seen": 324272128
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.5034213066101074,
      "objective/train/avg_token_score": 0.00968634057790041,
      "objective/train/docs_used": 195463,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7959951162338257,
      "objective/train/original_loss": 3.7963671684265137,
      "objective/train/theoretical_loss": 4.108788601279149,
      "objective/train/tokens_used": 324403200,
      "objective/train/unlikelihood_loss": 0.1474306285381317,
      "theoretical_loss": 4.108788601279149,
      "tokens_seen": 324403200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045540621865596795,
      "loss": 1.5169,
      "theoretical_loss": 4.108788601279149,
      "tokens_seen": 324403200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553861584754263,
      "loss": 1.7446,
      "theoretical_loss": 4.108596340183819,
      "tokens_seen": 324534272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553660982948847,
      "loss": 1.6527,
      "theoretical_loss": 4.108404178454651,
      "tokens_seen": 324665344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045534603811434304,
      "loss": 1.5472,
      "theoretical_loss": 4.108212116000203,
      "tokens_seen": 324796416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045532597793380145,
      "loss": 1.6305,
      "theoretical_loss": 4.108020152729157,
      "tokens_seen": 324927488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553059177532598,
      "loss": 1.6545,
      "theoretical_loss": 4.107828288550314,
      "tokens_seen": 325058560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552858575727182,
      "loss": 1.4823,
      "theoretical_loss": 4.107636523372598,
      "tokens_seen": 325189632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045526579739217654,
      "loss": 1.4818,
      "theoretical_loss": 4.107444857105052,
      "tokens_seen": 325320704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552457372116349,
      "loss": 1.6236,
      "theoretical_loss": 4.107253289656838,
      "tokens_seen": 325451776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552256770310933,
      "loss": 1.4742,
      "theoretical_loss": 4.10706182093724,
      "tokens_seen": 325582848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552056168505517,
      "loss": 1.5021,
      "theoretical_loss": 4.106870450855661,
      "tokens_seen": 325713920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045518555667001004,
      "loss": 1.5357,
      "theoretical_loss": 4.106679179321622,
      "tokens_seen": 325844992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551654964894684,
      "loss": 1.7192,
      "theoretical_loss": 4.106488006244767,
      "tokens_seen": 325976064
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.49883919954299927,
      "objective/train/avg_token_score": 0.02902386523783207,
      "objective/train/docs_used": 196731,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7118357419967651,
      "objective/train/original_loss": 3.6418302059173584,
      "objective/train/theoretical_loss": 4.106392456599577,
      "objective/train/tokens_used": 326041600,
      "objective/train/unlikelihood_loss": 0.22379085421562195,
      "theoretical_loss": 4.106392456599577,
      "tokens_seen": 326041600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551454363089268,
      "loss": 1.5264,
      "theoretical_loss": 4.106296931534854,
      "tokens_seen": 326107136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551253761283852,
      "loss": 1.5085,
      "theoretical_loss": 4.106105955101766,
      "tokens_seen": 326238208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045510531594784354,
      "loss": 1.375,
      "theoretical_loss": 4.105915076855499,
      "tokens_seen": 326369280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550852557673019,
      "loss": 1.5322,
      "theoretical_loss": 4.105724296706172,
      "tokens_seen": 326500352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045506519558676027,
      "loss": 1.6069,
      "theoretical_loss": 4.1055336145640196,
      "tokens_seen": 326631424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550451354062187,
      "loss": 1.6159,
      "theoretical_loss": 4.105343030339395,
      "tokens_seen": 326762496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045502507522567705,
      "loss": 1.5891,
      "theoretical_loss": 4.10515254394277,
      "tokens_seen": 326893568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550050150451354,
      "loss": 1.5234,
      "theoretical_loss": 4.104962155284734,
      "tokens_seen": 327024640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045498495486459377,
      "loss": 1.5651,
      "theoretical_loss": 4.104771864275993,
      "tokens_seen": 327155712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549648946840522,
      "loss": 1.407,
      "theoretical_loss": 4.104581670827372,
      "tokens_seen": 327286784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045494483450351055,
      "loss": 1.6212,
      "theoretical_loss": 4.104391574849812,
      "tokens_seen": 327417856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549247743229689,
      "loss": 1.5187,
      "theoretical_loss": 4.104201576254369,
      "tokens_seen": 327548928
    },
    {
      "debugging/length": 505.3636363636364,
      "debugging/num_segments": 11,
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.5214442610740662,
      "objective/train/avg_token_score": 0.01770769990980625,
      "objective/train/docs_used": 197513,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0139143466949463,
      "objective/train/original_loss": 4.308063983917236,
      "objective/train/theoretical_loss": 4.10401167495222,
      "objective/train/tokens_used": 327680000,
      "objective/train/unlikelihood_loss": 0.12178285419940948,
      "theoretical_loss": 4.10401167495222,
      "tokens_seen": 327680000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549047141424273,
      "loss": 1.6542,
      "theoretical_loss": 4.10401167495222,
      "tokens_seen": 327680000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045488465396188564,
      "loss": 1.6045,
      "theoretical_loss": 4.103821870854656,
      "tokens_seen": 327811072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045486459378134405,
      "loss": 1.5895,
      "theoretical_loss": 4.103632163873083,
      "tokens_seen": 327942144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548445336008024,
      "loss": 1.6359,
      "theoretical_loss": 4.103442553919026,
      "tokens_seen": 328073216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548244734202608,
      "loss": 1.6449,
      "theoretical_loss": 4.103253040904124,
      "tokens_seen": 328204288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045480441323971914,
      "loss": 1.5452,
      "theoretical_loss": 4.103063624740133,
      "tokens_seen": 328335360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045478435305917756,
      "loss": 1.4296,
      "theoretical_loss": 4.102874305338923,
      "tokens_seen": 328466432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547642928786359,
      "loss": 1.532,
      "theoretical_loss": 4.10268508261248,
      "tokens_seen": 328597504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547442326980943,
      "loss": 1.6441,
      "theoretical_loss": 4.1024959564729055,
      "tokens_seen": 328728576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045472417251755264,
      "loss": 1.5003,
      "theoretical_loss": 4.102306926832417,
      "tokens_seen": 328859648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454704112337011,
      "loss": 1.5285,
      "theoretical_loss": 4.1021179936033425,
      "tokens_seen": 328990720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546840521564694,
      "loss": 1.5756,
      "theoretical_loss": 4.10192915669813,
      "tokens_seen": 329121792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546639919759278,
      "loss": 1.5778,
      "theoretical_loss": 4.101740416029338,
      "tokens_seen": 329252864
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.48374876379966736,
      "objective/train/avg_token_score": 0.012546883895993233,
      "objective/train/docs_used": 198138,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9386160373687744,
      "objective/train/original_loss": 4.203801155090332,
      "objective/train/theoretical_loss": 4.101646081756305,
      "objective/train/tokens_used": 329318400,
      "objective/train/unlikelihood_loss": 0.1514439433813095,
      "theoretical_loss": 4.101646081756305,
      "tokens_seen": 329318400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045464393179538615,
      "loss": 1.4654,
      "theoretical_loss": 4.101551771509641,
      "tokens_seen": 329383936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546238716148445,
      "loss": 1.6519,
      "theoretical_loss": 4.101363223051826,
      "tokens_seen": 329515008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546038114343029,
      "loss": 1.4715,
      "theoretical_loss": 4.101174770568795,
      "tokens_seen": 329646080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545837512537613,
      "loss": 1.4744,
      "theoretical_loss": 4.100986413973564,
      "tokens_seen": 329777152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045456369107321965,
      "loss": 1.7026,
      "theoretical_loss": 4.1007981531792606,
      "tokens_seen": 329908224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454543630892678,
      "loss": 1.4468,
      "theoretical_loss": 4.1006099880991265,
      "tokens_seen": 330039296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045452357071213643,
      "loss": 1.5817,
      "theoretical_loss": 4.100421918646517,
      "tokens_seen": 330170368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545035105315948,
      "loss": 1.5614,
      "theoretical_loss": 4.100233944734899,
      "tokens_seen": 330301440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045448345035105315,
      "loss": 1.6208,
      "theoretical_loss": 4.100046066277853,
      "tokens_seen": 330432512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544633901705115,
      "loss": 1.5886,
      "theoretical_loss": 4.0998582831890715,
      "tokens_seen": 330563584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544433299899699,
      "loss": 1.5603,
      "theoretical_loss": 4.09967059538236,
      "tokens_seen": 330694656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544232698094283,
      "loss": 1.5664,
      "theoretical_loss": 4.099483002771633,
      "tokens_seen": 330825728
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.34042033553123474,
      "objective/train/avg_token_score": 0.018601443618535995,
      "objective/train/docs_used": 199416,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2680652141571045,
      "objective/train/original_loss": 3.715749740600586,
      "objective/train/theoretical_loss": 4.099295505270921,
      "objective/train/tokens_used": 330956800,
      "objective/train/unlikelihood_loss": 0.23622927069664001,
      "theoretical_loss": 4.099295505270921,
      "tokens_seen": 330956800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045440320962888666,
      "loss": 1.3597,
      "theoretical_loss": 4.099295505270921,
      "tokens_seen": 330956800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454383149448345,
      "loss": 1.4881,
      "theoretical_loss": 4.099108102794363,
      "tokens_seen": 331087872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543630892678034,
      "loss": 1.5775,
      "theoretical_loss": 4.098920795256213,
      "tokens_seen": 331218944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543430290872618,
      "loss": 1.6801,
      "theoretical_loss": 4.098733582570831,
      "tokens_seen": 331350016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045432296890672016,
      "loss": 1.6044,
      "theoretical_loss": 4.098546464652693,
      "tokens_seen": 331481088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543029087261785,
      "loss": 1.5239,
      "theoretical_loss": 4.098359441416383,
      "tokens_seen": 331612160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542828485456369,
      "loss": 1.7179,
      "theoretical_loss": 4.098172512776597,
      "tokens_seen": 331743232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045426278836509525,
      "loss": 1.5754,
      "theoretical_loss": 4.097985678648142,
      "tokens_seen": 331874304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045424272818455366,
      "loss": 1.5853,
      "theoretical_loss": 4.097798938945933,
      "tokens_seen": 332005376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454222668004012,
      "loss": 1.4991,
      "theoretical_loss": 4.097612293584998,
      "tokens_seen": 332136448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045420260782347044,
      "loss": 1.389,
      "theoretical_loss": 4.097425742480472,
      "tokens_seen": 332267520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541825476429288,
      "loss": 1.7117,
      "theoretical_loss": 4.0972392855476025,
      "tokens_seen": 332398592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541624874623872,
      "loss": 1.5459,
      "theoretical_loss": 4.097052922701746,
      "tokens_seen": 332529664
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.3013196587562561,
      "objective/train/avg_token_score": 0.018458493053913116,
      "objective/train/docs_used": 200524,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0405497550964355,
      "objective/train/original_loss": 3.868088960647583,
      "objective/train/theoretical_loss": 4.0969597765350265,
      "objective/train/tokens_used": 332595200,
      "objective/train/unlikelihood_loss": 0.16207247972488403,
      "theoretical_loss": 4.0969597765350265,
      "tokens_seen": 332595200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541424272818456,
      "loss": 1.367,
      "theoretical_loss": 4.096866653858368,
      "tokens_seen": 332660736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045412236710130395,
      "loss": 1.617,
      "theoretical_loss": 4.096680478933042,
      "tokens_seen": 332791808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541023069207623,
      "loss": 1.5846,
      "theoretical_loss": 4.09649439784145,
      "tokens_seen": 332922880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045408224674022067,
      "loss": 1.6318,
      "theoretical_loss": 4.09630841049939,
      "tokens_seen": 333053952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540621865596791,
      "loss": 1.655,
      "theoretical_loss": 4.096122516822757,
      "tokens_seen": 333185024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045404212637913745,
      "loss": 1.4591,
      "theoretical_loss": 4.095936716727564,
      "tokens_seen": 333316096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540220661985958,
      "loss": 1.4523,
      "theoretical_loss": 4.095751010129929,
      "tokens_seen": 333447168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045400200601805417,
      "loss": 1.6392,
      "theoretical_loss": 4.0955653969460775,
      "tokens_seen": 333578240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539819458375126,
      "loss": 1.5289,
      "theoretical_loss": 4.095379877092343,
      "tokens_seen": 333709312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045396188565697095,
      "loss": 1.7503,
      "theoretical_loss": 4.0951944504851685,
      "tokens_seen": 333840384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539418254764293,
      "loss": 1.4979,
      "theoretical_loss": 4.095009117041102,
      "tokens_seen": 333971456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539217652958877,
      "loss": 1.7049,
      "theoretical_loss": 4.094823876676802,
      "tokens_seen": 334102528
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.5921309590339661,
      "objective/train/avg_token_score": 0.006544371135532856,
      "objective/train/docs_used": 201196,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.2652223110198975,
      "objective/train/original_loss": 4.068779468536377,
      "objective/train/theoretical_loss": 4.094638729309031,
      "objective/train/tokens_used": 334233600,
      "objective/train/unlikelihood_loss": 0.11625197529792786,
      "theoretical_loss": 4.094638729309031,
      "tokens_seen": 334233600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045390170511534604,
      "loss": 1.4098,
      "theoretical_loss": 4.094638729309031,
      "tokens_seen": 334233600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045388164493480445,
      "loss": 1.6587,
      "theoretical_loss": 4.09445367485466,
      "tokens_seen": 334364672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538615847542628,
      "loss": 1.5036,
      "theoretical_loss": 4.094268713230667,
      "tokens_seen": 334495744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538415245737212,
      "loss": 1.4698,
      "theoretical_loss": 4.094083844354137,
      "tokens_seen": 334626816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045382146439317954,
      "loss": 1.4929,
      "theoretical_loss": 4.093899068142262,
      "tokens_seen": 334757888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045380140421263796,
      "loss": 1.6098,
      "theoretical_loss": 4.093714384512337,
      "tokens_seen": 334888960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537813440320963,
      "loss": 1.6043,
      "theoretical_loss": 4.093529793381768,
      "tokens_seen": 335020032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537612838515547,
      "loss": 1.6377,
      "theoretical_loss": 4.093345294668063,
      "tokens_seen": 335151104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045374122367101304,
      "loss": 1.7649,
      "theoretical_loss": 4.093160888288837,
      "tokens_seen": 335282176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537211634904714,
      "loss": 1.5487,
      "theoretical_loss": 4.0929765741618125,
      "tokens_seen": 335413248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537011033099298,
      "loss": 1.519,
      "theoretical_loss": 4.092792352204814,
      "tokens_seen": 335544320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536810431293882,
      "loss": 1.5532,
      "theoretical_loss": 4.092608222335774,
      "tokens_seen": 335675392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045366098294884655,
      "loss": 1.5037,
      "theoretical_loss": 4.09242418447273,
      "tokens_seen": 335806464
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.4757453501224518,
      "objective/train/avg_token_score": 0.009879210032522678,
      "objective/train/docs_used": 202460,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8038827180862427,
      "objective/train/original_loss": 4.1414971351623535,
      "objective/train/theoretical_loss": 4.092332200017871,
      "objective/train/tokens_used": 335872000,
      "objective/train/unlikelihood_loss": 0.12569139897823334,
      "theoretical_loss": 4.092332200017871,
      "tokens_seen": 335872000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536409227683049,
      "loss": 1.7663,
      "theoretical_loss": 4.092240238533822,
      "tokens_seen": 335937536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536208625877633,
      "loss": 1.5856,
      "theoretical_loss": 4.092056384437297,
      "tokens_seen": 336068608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536008024072217,
      "loss": 1.63,
      "theoretical_loss": 4.091872622101506,
      "tokens_seen": 336199680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045358074222668005,
      "loss": 1.6248,
      "theoretical_loss": 4.091688951444904,
      "tokens_seen": 336330752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535606820461384,
      "loss": 1.5825,
      "theoretical_loss": 4.091505372386051,
      "tokens_seen": 336461824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045354062186559683,
      "loss": 1.6858,
      "theoretical_loss": 4.091321884843609,
      "tokens_seen": 336592896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535205616850552,
      "loss": 1.5953,
      "theoretical_loss": 4.0911384887363464,
      "tokens_seen": 336723968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045350050150451355,
      "loss": 1.6498,
      "theoretical_loss": 4.0909551839831355,
      "tokens_seen": 336855040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534804413239719,
      "loss": 1.5834,
      "theoretical_loss": 4.090771970502948,
      "tokens_seen": 336986112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534603811434303,
      "loss": 1.5166,
      "theoretical_loss": 4.090588848214865,
      "tokens_seen": 337117184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534403209628887,
      "loss": 1.5447,
      "theoretical_loss": 4.090405817038065,
      "tokens_seen": 337248256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045342026078234706,
      "loss": 1.4795,
      "theoretical_loss": 4.090222876891834,
      "tokens_seen": 337379328
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.4949902296066284,
      "objective/train/avg_token_score": 0.009650841355323792,
      "objective/train/docs_used": 203033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7750602960586548,
      "objective/train/original_loss": 3.9335618019104004,
      "objective/train/theoretical_loss": 4.090040027695556,
      "objective/train/tokens_used": 337510400,
      "objective/train/unlikelihood_loss": 0.13872487843036652,
      "theoretical_loss": 4.090040027695556,
      "tokens_seen": 337510400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534002006018054,
      "loss": 1.712,
      "theoretical_loss": 4.090040027695556,
      "tokens_seen": 337510400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533801404212638,
      "loss": 1.874,
      "theoretical_loss": 4.089857269368725,
      "tokens_seen": 337641472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533600802407222,
      "loss": 1.5283,
      "theoretical_loss": 4.089674601830929,
      "tokens_seen": 337772544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045334002006018056,
      "loss": 1.6553,
      "theoretical_loss": 4.089492025001864,
      "tokens_seen": 337903616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533199598796389,
      "loss": 1.5234,
      "theoretical_loss": 4.089309538801327,
      "tokens_seen": 338034688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532998996990973,
      "loss": 1.5957,
      "theoretical_loss": 4.0891271431492155,
      "tokens_seen": 338165760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045327983951855565,
      "loss": 1.7097,
      "theoretical_loss": 4.08894483796553,
      "tokens_seen": 338296832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045325977933801406,
      "loss": 1.6363,
      "theoretical_loss": 4.088762623170373,
      "tokens_seen": 338427904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532397191574724,
      "loss": 1.6686,
      "theoretical_loss": 4.088580498683948,
      "tokens_seen": 338558976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532196589769308,
      "loss": 1.6399,
      "theoretical_loss": 4.088398464426559,
      "tokens_seen": 338690048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045319959879638915,
      "loss": 1.6365,
      "theoretical_loss": 4.088216520318612,
      "tokens_seen": 338821120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045317953861584757,
      "loss": 1.5666,
      "theoretical_loss": 4.088034666280614,
      "tokens_seen": 338952192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045315947843530593,
      "loss": 1.6203,
      "theoretical_loss": 4.087852902233173,
      "tokens_seen": 339083264
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.4335288405418396,
      "objective/train/avg_token_score": 0.013492316007614136,
      "objective/train/docs_used": 204232,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5629843473434448,
      "objective/train/original_loss": 3.536388874053955,
      "objective/train/theoretical_loss": 4.087762053931129,
      "objective/train/tokens_used": 339148800,
      "objective/train/unlikelihood_loss": 0.25973203778266907,
      "theoretical_loss": 4.087762053931129,
      "tokens_seen": 339148800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531394182547643,
      "loss": 1.6724,
      "theoretical_loss": 4.087671228096997,
      "tokens_seen": 339214336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045311935807422265,
      "loss": 1.5495,
      "theoretical_loss": 4.087489643792894,
      "tokens_seen": 339345408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453099297893681,
      "loss": 1.6182,
      "theoretical_loss": 4.087308149241774,
      "tokens_seen": 339476480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045307923771313943,
      "loss": 1.6192,
      "theoretical_loss": 4.087126744364646,
      "tokens_seen": 339607552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530591775325978,
      "loss": 1.5072,
      "theoretical_loss": 4.086945429082618,
      "tokens_seen": 339738624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045303911735205616,
      "loss": 1.6224,
      "theoretical_loss": 4.086764203316902,
      "tokens_seen": 339869696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530190571715145,
      "loss": 1.5255,
      "theoretical_loss": 4.086583066988802,
      "tokens_seen": 340000768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045299899699097294,
      "loss": 1.5888,
      "theoretical_loss": 4.0864020200197295,
      "tokens_seen": 340131840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529789368104313,
      "loss": 1.6588,
      "theoretical_loss": 4.086221062331192,
      "tokens_seen": 340262912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045295887662988966,
      "loss": 1.6366,
      "theoretical_loss": 4.086040193844794,
      "tokens_seen": 340393984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452938816449348,
      "loss": 1.4848,
      "theoretical_loss": 4.0858594144822415,
      "tokens_seen": 340525056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529187562688064,
      "loss": 1.5903,
      "theoretical_loss": 4.085678724165341,
      "tokens_seen": 340656128
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.32972875237464905,
      "objective/train/avg_token_score": 0.018639761954545975,
      "objective/train/docs_used": 205032,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0958448648452759,
      "objective/train/original_loss": 3.859755754470825,
      "objective/train/theoretical_loss": 4.085498122815992,
      "objective/train/tokens_used": 340787200,
      "objective/train/unlikelihood_loss": 0.2070046067237854,
      "theoretical_loss": 4.085498122815992,
      "tokens_seen": 340787200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528986960882648,
      "loss": 1.5739,
      "theoretical_loss": 4.085498122815992,
      "tokens_seen": 340787200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045287863590772316,
      "loss": 1.6326,
      "theoretical_loss": 4.085317610356199,
      "tokens_seen": 340918272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528585757271815,
      "loss": 1.4208,
      "theoretical_loss": 4.08513718670806,
      "tokens_seen": 341049344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528385155466399,
      "loss": 1.5432,
      "theoretical_loss": 4.084956851793773,
      "tokens_seen": 341180416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528184553660983,
      "loss": 1.5042,
      "theoretical_loss": 4.0847766055356365,
      "tokens_seen": 341311488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045279839518555667,
      "loss": 1.6324,
      "theoretical_loss": 4.0845964478560415,
      "tokens_seen": 341442560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045277833500501503,
      "loss": 1.4951,
      "theoretical_loss": 4.0844163786774805,
      "tokens_seen": 341573632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004527582748244734,
      "loss": 1.6173,
      "theoretical_loss": 4.084236397922544,
      "tokens_seen": 341704704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045273821464393175,
      "loss": 1.5795,
      "theoretical_loss": 4.0840565055139155,
      "tokens_seen": 341835776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045271815446339017,
      "loss": 1.5769,
      "theoretical_loss": 4.083876701374381,
      "tokens_seen": 341966848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526980942828486,
      "loss": 1.505,
      "theoretical_loss": 4.08369698542682,
      "tokens_seen": 342097920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045267803410230695,
      "loss": 1.5581,
      "theoretical_loss": 4.083517357594212,
      "tokens_seen": 342228992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526579739217653,
      "loss": 1.3931,
      "theoretical_loss": 4.083337817799629,
      "tokens_seen": 342360064
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.339259535074234,
      "objective/train/avg_token_score": 0.027388932183384895,
      "objective/train/docs_used": 206109,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2726820707321167,
      "objective/train/original_loss": 4.560004234313965,
      "objective/train/theoretical_loss": 4.083248080892584,
      "objective/train/tokens_used": 342425600,
      "objective/train/unlikelihood_loss": 0.11201095581054688,
      "theoretical_loss": 4.083248080892584,
      "tokens_seen": 342425600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526379137412237,
      "loss": 1.5391,
      "theoretical_loss": 4.083158365966243,
      "tokens_seen": 342491136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526178535606821,
      "loss": 1.6968,
      "theoretical_loss": 4.082979002017321,
      "tokens_seen": 342622208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045259779338014045,
      "loss": 1.6573,
      "theoretical_loss": 4.082799725876228,
      "tokens_seen": 342753280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525777331995988,
      "loss": 1.6508,
      "theoretical_loss": 4.082620537466422,
      "tokens_seen": 342884352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045255767301905723,
      "loss": 1.6432,
      "theoretical_loss": 4.082441436711459,
      "tokens_seen": 343015424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525376128385156,
      "loss": 1.5502,
      "theoretical_loss": 4.082262423534992,
      "tokens_seen": 343146496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045251755265797395,
      "loss": 1.4503,
      "theoretical_loss": 4.082083497860766,
      "tokens_seen": 343277568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524974924774323,
      "loss": 1.5342,
      "theoretical_loss": 4.081904659612624,
      "tokens_seen": 343408640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524774322968907,
      "loss": 1.3966,
      "theoretical_loss": 4.081725908714505,
      "tokens_seen": 343539712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524573721163491,
      "loss": 1.4216,
      "theoretical_loss": 4.08154724509044,
      "tokens_seen": 343670784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045243731193580746,
      "loss": 1.3198,
      "theoretical_loss": 4.08136866866456,
      "tokens_seen": 343801856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524172517552658,
      "loss": 1.5836,
      "theoretical_loss": 4.0811901793610845,
      "tokens_seen": 343932928
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.34237536787986755,
      "objective/train/avg_token_score": 0.0033272351138293743,
      "objective/train/docs_used": 206760,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2112702131271362,
      "objective/train/original_loss": 3.6624834537506104,
      "objective/train/theoretical_loss": 4.081011777104333,
      "objective/train/tokens_used": 344064000,
      "objective/train/unlikelihood_loss": 0.18153946101665497,
      "theoretical_loss": 4.081011777104333,
      "tokens_seen": 344064000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523971915747242,
      "loss": 1.4097,
      "theoretical_loss": 4.081011777104333,
      "tokens_seen": 344064000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523771313941826,
      "loss": 1.7856,
      "theoretical_loss": 4.080833461818718,
      "tokens_seen": 344195072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045235707121364096,
      "loss": 1.5064,
      "theoretical_loss": 4.080655233428745,
      "tokens_seen": 344326144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523370110330993,
      "loss": 1.5193,
      "theoretical_loss": 4.080477091859015,
      "tokens_seen": 344457216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523169508525577,
      "loss": 1.5518,
      "theoretical_loss": 4.080299037034224,
      "tokens_seen": 344588288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045229689067201605,
      "loss": 1.6474,
      "theoretical_loss": 4.0801210688791585,
      "tokens_seen": 344719360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045227683049147446,
      "loss": 1.5025,
      "theoretical_loss": 4.079943187318703,
      "tokens_seen": 344850432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522567703109328,
      "loss": 1.6411,
      "theoretical_loss": 4.0797653922778325,
      "tokens_seen": 344981504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522367101303912,
      "loss": 1.6179,
      "theoretical_loss": 4.079587683681617,
      "tokens_seen": 345112576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045221664994984955,
      "loss": 1.4868,
      "theoretical_loss": 4.07941006145522,
      "tokens_seen": 345243648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045219658976930797,
      "loss": 1.357,
      "theoretical_loss": 4.0792325255238975,
      "tokens_seen": 345374720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045217652958876633,
      "loss": 1.4853,
      "theoretical_loss": 4.079055075812997,
      "tokens_seen": 345505792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521564694082247,
      "loss": 1.6133,
      "theoretical_loss": 4.078877712247963,
      "tokens_seen": 345636864
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.1969086080789566,
      "objective/train/avg_token_score": 0.01607353612780571,
      "objective/train/docs_used": 207336,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7453041076660156,
      "objective/train/original_loss": 4.304450988769531,
      "objective/train/theoretical_loss": 4.078789062746872,
      "objective/train/tokens_used": 345702400,
      "objective/train/unlikelihood_loss": 0.1982918530702591,
      "theoretical_loss": 4.078789062746872,
      "tokens_seen": 345702400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045213640922768305,
      "loss": 1.3566,
      "theoretical_loss": 4.07870043475433,
      "tokens_seen": 345767936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521163490471414,
      "loss": 1.4831,
      "theoretical_loss": 4.078523243257723,
      "tokens_seen": 345899008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045209628886659983,
      "loss": 1.6821,
      "theoretical_loss": 4.078346137683862,
      "tokens_seen": 346030080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520762286860582,
      "loss": 1.5421,
      "theoretical_loss": 4.078169117958562,
      "tokens_seen": 346161152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045205616850551656,
      "loss": 1.4795,
      "theoretical_loss": 4.077992184007725,
      "tokens_seen": 346292224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520361083249749,
      "loss": 1.6154,
      "theoretical_loss": 4.077815335757347,
      "tokens_seen": 346423296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045201604814443334,
      "loss": 1.5112,
      "theoretical_loss": 4.077638573133515,
      "tokens_seen": 346554368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519959879638917,
      "loss": 1.5462,
      "theoretical_loss": 4.0774618960624105,
      "tokens_seen": 346685440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045197592778335006,
      "loss": 1.6378,
      "theoretical_loss": 4.077285304470304,
      "tokens_seen": 346816512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519558676028084,
      "loss": 1.5589,
      "theoretical_loss": 4.077108798283557,
      "tokens_seen": 346947584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519358074222668,
      "loss": 1.3784,
      "theoretical_loss": 4.076932377428625,
      "tokens_seen": 347078656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519157472417252,
      "loss": 1.4946,
      "theoretical_loss": 4.076756041832049,
      "tokens_seen": 347209728
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.3781769275665283,
      "objective/train/avg_token_score": 0.007616470567882061,
      "objective/train/docs_used": 208796,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4440606832504272,
      "objective/train/original_loss": 3.906726360321045,
      "objective/train/theoretical_loss": 4.076579791420469,
      "objective/train/tokens_used": 347340800,
      "objective/train/unlikelihood_loss": 0.1508537232875824,
      "theoretical_loss": 4.076579791420469,
      "tokens_seen": 347340800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045189568706118356,
      "loss": 1.6182,
      "theoretical_loss": 4.076579791420469,
      "tokens_seen": 347340800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518756268806419,
      "loss": 1.5178,
      "theoretical_loss": 4.07640362612061,
      "tokens_seen": 347471872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518555667001003,
      "loss": 1.529,
      "theoretical_loss": 4.076227545859288,
      "tokens_seen": 347602944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518355065195587,
      "loss": 1.6502,
      "theoretical_loss": 4.076051550563411,
      "tokens_seen": 347734016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045181544633901707,
      "loss": 1.3819,
      "theoretical_loss": 4.075875640159978,
      "tokens_seen": 347865088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045179538615847543,
      "loss": 1.5274,
      "theoretical_loss": 4.075699814576076,
      "tokens_seen": 347996160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004517753259779338,
      "loss": 1.5282,
      "theoretical_loss": 4.075524073738883,
      "tokens_seen": 348127232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045175526579739215,
      "loss": 1.5711,
      "theoretical_loss": 4.075348417575669,
      "tokens_seen": 348258304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045173520561685057,
      "loss": 1.6626,
      "theoretical_loss": 4.075172846013789,
      "tokens_seen": 348389376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045171514543630893,
      "loss": 1.4903,
      "theoretical_loss": 4.074997358980694,
      "tokens_seen": 348520448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516950852557673,
      "loss": 1.6022,
      "theoretical_loss": 4.074821956403917,
      "tokens_seen": 348651520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045167502507522566,
      "loss": 1.5102,
      "theoretical_loss": 4.074646638211087,
      "tokens_seen": 348782592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045165496489468407,
      "loss": 1.6114,
      "theoretical_loss": 4.074471404329918,
      "tokens_seen": 348913664
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.5323802828788757,
      "objective/train/avg_token_score": 0.004500185139477253,
      "objective/train/docs_used": 209597,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1040139198303223,
      "objective/train/original_loss": 4.173861980438232,
      "objective/train/theoretical_loss": 4.074383818983643,
      "objective/train/tokens_used": 348979200,
      "objective/train/unlikelihood_loss": 0.11532780528068542,
      "theoretical_loss": 4.074383818983643,
      "tokens_seen": 348979200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045163490471414244,
      "loss": 1.7397,
      "theoretical_loss": 4.074296254688216,
      "tokens_seen": 349044736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516148445336008,
      "loss": 1.5713,
      "theoretical_loss": 4.074121189213873,
      "tokens_seen": 349175808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045159478435305916,
      "loss": 1.4462,
      "theoretical_loss": 4.073946207834872,
      "tokens_seen": 349306880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515747241725175,
      "loss": 1.5493,
      "theoretical_loss": 4.073771310479284,
      "tokens_seen": 349437952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045155466399197594,
      "loss": 1.4127,
      "theoretical_loss": 4.073596497075267,
      "tokens_seen": 349569024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515346038114343,
      "loss": 1.3747,
      "theoretical_loss": 4.07342176755107,
      "tokens_seen": 349700096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045151454363089266,
      "loss": 1.4241,
      "theoretical_loss": 4.073247121835028,
      "tokens_seen": 349831168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451494483450351,
      "loss": 1.5795,
      "theoretical_loss": 4.073072559855565,
      "tokens_seen": 349962240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045147442326980944,
      "loss": 1.5063,
      "theoretical_loss": 4.072898081541193,
      "tokens_seen": 350093312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514543630892678,
      "loss": 1.4867,
      "theoretical_loss": 4.0727236868205114,
      "tokens_seen": 350224384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045143430290872617,
      "loss": 1.4653,
      "theoretical_loss": 4.072549375622207,
      "tokens_seen": 350355456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045141424272818453,
      "loss": 1.5842,
      "theoretical_loss": 4.072375147875054,
      "tokens_seen": 350486528
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.3748167157173157,
      "objective/train/avg_token_score": 0.026631662622094154,
      "objective/train/docs_used": 210243,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2623140811920166,
      "objective/train/original_loss": 4.159633636474609,
      "objective/train/theoretical_loss": 4.0722010035079155,
      "objective/train/tokens_used": 350617600,
      "objective/train/unlikelihood_loss": 0.13227179646492004,
      "theoretical_loss": 4.0722010035079155,
      "tokens_seen": 350617600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045139418254764294,
      "loss": 1.4965,
      "theoretical_loss": 4.0722010035079155,
      "tokens_seen": 350617600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513741223671013,
      "loss": 1.4648,
      "theoretical_loss": 4.07202694244974,
      "tokens_seen": 350748672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045135406218655967,
      "loss": 1.2912,
      "theoretical_loss": 4.0718529646295645,
      "tokens_seen": 350879744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045133400200601803,
      "loss": 1.3967,
      "theoretical_loss": 4.071679069976511,
      "tokens_seen": 351010816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513139418254764,
      "loss": 1.4605,
      "theoretical_loss": 4.07150525841979,
      "tokens_seen": 351141888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512938816449348,
      "loss": 1.4834,
      "theoretical_loss": 4.071331529888698,
      "tokens_seen": 351272960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045127382146439317,
      "loss": 1.4192,
      "theoretical_loss": 4.071157884312619,
      "tokens_seen": 351404032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045125376128385153,
      "loss": 1.5059,
      "theoretical_loss": 4.070984321621021,
      "tokens_seen": 351535104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512337011033099,
      "loss": 1.5377,
      "theoretical_loss": 4.0708108417434605,
      "tokens_seen": 351666176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512136409227683,
      "loss": 1.4409,
      "theoretical_loss": 4.07063744460958,
      "tokens_seen": 351797248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045119358074222673,
      "loss": 1.4988,
      "theoretical_loss": 4.070464130149106,
      "tokens_seen": 351928320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511735205616851,
      "loss": 1.5584,
      "theoretical_loss": 4.070290898291852,
      "tokens_seen": 352059392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045115346038114345,
      "loss": 1.5342,
      "theoretical_loss": 4.07011774896772,
      "tokens_seen": 352190464
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.5329301357269287,
      "objective/train/avg_token_score": 0.007351990323513746,
      "objective/train/docs_used": 211560,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9991000890731812,
      "objective/train/original_loss": 4.122071743011475,
      "objective/train/theoretical_loss": 4.07003120523369,
      "objective/train/tokens_used": 352256000,
      "objective/train/unlikelihood_loss": 0.16980284452438354,
      "theoretical_loss": 4.07003120523369,
      "tokens_seen": 352256000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511334002006018,
      "loss": 1.599,
      "theoretical_loss": 4.069944682106691,
      "tokens_seen": 352321536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045111334002006023,
      "loss": 1.4121,
      "theoretical_loss": 4.069771697638838,
      "tokens_seen": 352452608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510932798395186,
      "loss": 1.5508,
      "theoretical_loss": 4.069598795494315,
      "tokens_seen": 352583680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045107321965897696,
      "loss": 1.3157,
      "theoretical_loss": 4.069425975603364,
      "tokens_seen": 352714752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510531594784353,
      "loss": 1.4493,
      "theoretical_loss": 4.0692532378963095,
      "tokens_seen": 352845824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045103309929789374,
      "loss": 1.4906,
      "theoretical_loss": 4.069080582303563,
      "tokens_seen": 352976896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510130391173521,
      "loss": 1.5558,
      "theoretical_loss": 4.068908008755619,
      "tokens_seen": 353107968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045099297893681046,
      "loss": 1.4319,
      "theoretical_loss": 4.0687355171830575,
      "tokens_seen": 353239040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509729187562688,
      "loss": 1.6799,
      "theoretical_loss": 4.068563107516544,
      "tokens_seen": 353370112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509528585757272,
      "loss": 1.5795,
      "theoretical_loss": 4.068390779686826,
      "tokens_seen": 353501184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509327983951856,
      "loss": 1.2278,
      "theoretical_loss": 4.068218533624737,
      "tokens_seen": 353632256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045091273821464396,
      "loss": 1.5487,
      "theoretical_loss": 4.068046369261194,
      "tokens_seen": 353763328
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.3893572688102722,
      "objective/train/avg_token_score": 0.027582136914134026,
      "objective/train/docs_used": 212189,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.257407784461975,
      "objective/train/original_loss": 3.744044065475464,
      "objective/train/theoretical_loss": 4.067874286527197,
      "objective/train/tokens_used": 353894400,
      "objective/train/unlikelihood_loss": 0.16484889388084412,
      "theoretical_loss": 4.067874286527197,
      "tokens_seen": 353894400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508926780341023,
      "loss": 1.5564,
      "theoretical_loss": 4.067874286527197,
      "tokens_seen": 353894400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508726178535607,
      "loss": 1.4902,
      "theoretical_loss": 4.067702285353833,
      "tokens_seen": 354025472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508525576730191,
      "loss": 1.4888,
      "theoretical_loss": 4.067530365672267,
      "tokens_seen": 354156544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045083249749247747,
      "loss": 1.4045,
      "theoretical_loss": 4.067358527413753,
      "tokens_seen": 354287616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045081243731193583,
      "loss": 1.5815,
      "theoretical_loss": 4.067186770509627,
      "tokens_seen": 354418688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507923771313942,
      "loss": 1.4217,
      "theoretical_loss": 4.067015094891306,
      "tokens_seen": 354549760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045077231695085255,
      "loss": 1.5362,
      "theoretical_loss": 4.066843500490292,
      "tokens_seen": 354680832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045075225677031097,
      "loss": 1.4491,
      "theoretical_loss": 4.066671987238171,
      "tokens_seen": 354811904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045073219658976933,
      "loss": 1.3298,
      "theoretical_loss": 4.066500555066609,
      "tokens_seen": 354942976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507121364092277,
      "loss": 1.4452,
      "theoretical_loss": 4.066329203907358,
      "tokens_seen": 355074048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045069207622868606,
      "loss": 1.6151,
      "theoretical_loss": 4.066157933692249,
      "tokens_seen": 355205120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506720160481445,
      "loss": 1.4978,
      "theoretical_loss": 4.0659867443532,
      "tokens_seen": 355336192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045065195586760284,
      "loss": 1.6458,
      "theoretical_loss": 4.065815635822206,
      "tokens_seen": 355467264
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.5819892287254333,
      "objective/train/avg_token_score": 0.017979789525270462,
      "objective/train/docs_used": 213521,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.2102198600769043,
      "objective/train/original_loss": 4.2459540367126465,
      "objective/train/theoretical_loss": 4.065730111838502,
      "objective/train/tokens_used": 355532800,
      "objective/train/unlikelihood_loss": 0.09294069558382034,
      "theoretical_loss": 4.065730111838502,
      "tokens_seen": 355532800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506318956870612,
      "loss": 1.6467,
      "theoretical_loss": 4.0656446080313495,
      "tokens_seen": 355598336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045061183550651956,
      "loss": 1.509,
      "theoretical_loss": 4.065473660912791,
      "tokens_seen": 355729408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505917753259779,
      "loss": 1.3671,
      "theoretical_loss": 4.065302794398775,
      "tokens_seen": 355860480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045057171514543634,
      "loss": 1.3185,
      "theoretical_loss": 4.065132008421628,
      "tokens_seen": 355991552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505516549648947,
      "loss": 1.4506,
      "theoretical_loss": 4.0649613029137575,
      "tokens_seen": 356122624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045053159478435306,
      "loss": 1.31,
      "theoretical_loss": 4.064790677807652,
      "tokens_seen": 356253696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505115346038114,
      "loss": 1.3887,
      "theoretical_loss": 4.064620133035883,
      "tokens_seen": 356384768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045049147442326984,
      "loss": 1.4987,
      "theoretical_loss": 4.064449668531103,
      "tokens_seen": 356515840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504714142427282,
      "loss": 1.6307,
      "theoretical_loss": 4.064279284226044,
      "tokens_seen": 356646912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045045135406218657,
      "loss": 1.4173,
      "theoretical_loss": 4.06410898005352,
      "tokens_seen": 356777984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045043129388164493,
      "loss": 1.7445,
      "theoretical_loss": 4.0639387559464275,
      "tokens_seen": 356909056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045041123370110334,
      "loss": 1.7331,
      "theoretical_loss": 4.063768611837742,
      "tokens_seen": 357040128
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.2592253088951111,
      "objective/train/avg_token_score": 0.012151106260716915,
      "objective/train/docs_used": 214188,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0094984769821167,
      "objective/train/original_loss": 4.309907913208008,
      "objective/train/theoretical_loss": 4.063598547660519,
      "objective/train/tokens_used": 357171200,
      "objective/train/unlikelihood_loss": 0.2334572970867157,
      "theoretical_loss": 4.063598547660519,
      "tokens_seen": 357171200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503911735205617,
      "loss": 1.4034,
      "theoretical_loss": 4.063598547660519,
      "tokens_seen": 357171200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045037111334002007,
      "loss": 1.5863,
      "theoretical_loss": 4.063428563347898,
      "tokens_seen": 357302272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045035105315947843,
      "loss": 1.4359,
      "theoretical_loss": 4.063258658833096,
      "tokens_seen": 357433344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503309929789368,
      "loss": 1.468,
      "theoretical_loss": 4.0630888340494105,
      "tokens_seen": 357564416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503109327983952,
      "loss": 1.4797,
      "theoretical_loss": 4.06291908893022,
      "tokens_seen": 357695488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045029087261785357,
      "loss": 1.5528,
      "theoretical_loss": 4.062749423408981,
      "tokens_seen": 357826560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045027081243731193,
      "loss": 1.4224,
      "theoretical_loss": 4.062579837419235,
      "tokens_seen": 357957632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502507522567703,
      "loss": 1.5149,
      "theoretical_loss": 4.062410330894597,
      "tokens_seen": 358088704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502306920762287,
      "loss": 1.5902,
      "theoretical_loss": 4.062240903768767,
      "tokens_seen": 358219776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502106318956871,
      "loss": 1.5077,
      "theoretical_loss": 4.062071555975519,
      "tokens_seen": 358350848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045019057171514544,
      "loss": 1.3566,
      "theoretical_loss": 4.061902287448714,
      "tokens_seen": 358481920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501705115346038,
      "loss": 1.5275,
      "theoretical_loss": 4.061733098122284,
      "tokens_seen": 358612992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045015045135406216,
      "loss": 1.525,
      "theoretical_loss": 4.0615639879302465,
      "tokens_seen": 358744064
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.5088587403297424,
      "objective/train/avg_token_score": 0.01710294932126999,
      "objective/train/docs_used": 215284,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8704804182052612,
      "objective/train/original_loss": 3.966881513595581,
      "objective/train/theoretical_loss": 4.061479462489026,
      "objective/train/tokens_used": 358809600,
      "objective/train/unlikelihood_loss": 0.14359432458877563,
      "theoretical_loss": 4.061479462489026,
      "tokens_seen": 358809600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501303911735206,
      "loss": 1.5061,
      "theoretical_loss": 4.061394956806694,
      "tokens_seen": 358875136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045011033099297894,
      "loss": 1.3324,
      "theoretical_loss": 4.0612260046858015,
      "tokens_seen": 359006208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500902708124373,
      "loss": 1.3371,
      "theoretical_loss": 4.06105713150182,
      "tokens_seen": 359137280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045007021063189567,
      "loss": 1.4478,
      "theoretical_loss": 4.0608883371890805,
      "tokens_seen": 359268352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500501504513541,
      "loss": 1.4508,
      "theoretical_loss": 4.06071962168199,
      "tokens_seen": 359399424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045003009027081244,
      "loss": 1.5374,
      "theoretical_loss": 4.060550984915039,
      "tokens_seen": 359530496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500100300902708,
      "loss": 1.3958,
      "theoretical_loss": 4.060382426822792,
      "tokens_seen": 359661568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044998996990972917,
      "loss": 1.4925,
      "theoretical_loss": 4.060213947339894,
      "tokens_seen": 359792640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044996990972918753,
      "loss": 1.4814,
      "theoretical_loss": 4.060045546401065,
      "tokens_seen": 359923712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044994984954864595,
      "loss": 1.5105,
      "theoretical_loss": 4.0598772239411085,
      "tokens_seen": 360054784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004499297893681043,
      "loss": 1.3412,
      "theoretical_loss": 4.0597089798949,
      "tokens_seen": 360185856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044990972918756267,
      "loss": 1.443,
      "theoretical_loss": 4.059540814197396,
      "tokens_seen": 360316928
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.3905791640281677,
      "objective/train/avg_token_score": 0.048172734677791595,
      "objective/train/docs_used": 215704,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4849482774734497,
      "objective/train/original_loss": 4.436104774475098,
      "objective/train/theoretical_loss": 4.05937272678363,
      "objective/train/tokens_used": 360448000,
      "objective/train/unlikelihood_loss": 0.11548055708408356,
      "theoretical_loss": 4.05937272678363,
      "tokens_seen": 360448000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044988966900702103,
      "loss": 1.4311,
      "theoretical_loss": 4.05937272678363,
      "tokens_seen": 360448000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044986960882647945,
      "loss": 1.552,
      "theoretical_loss": 4.059204717588712,
      "tokens_seen": 360579072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498495486459378,
      "loss": 1.4372,
      "theoretical_loss": 4.059036786547831,
      "tokens_seen": 360710144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498294884653962,
      "loss": 1.481,
      "theoretical_loss": 4.058868933596252,
      "tokens_seen": 360841216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044980942828485454,
      "loss": 1.598,
      "theoretical_loss": 4.058701158669318,
      "tokens_seen": 360972288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497893681043129,
      "loss": 1.4096,
      "theoretical_loss": 4.0585334617024476,
      "tokens_seen": 361103360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497693079237713,
      "loss": 1.5581,
      "theoretical_loss": 4.058365842631137,
      "tokens_seen": 361234432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497492477432297,
      "loss": 1.4585,
      "theoretical_loss": 4.058198301390959,
      "tokens_seen": 361365504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044972918756268804,
      "loss": 1.4782,
      "theoretical_loss": 4.058030837917565,
      "tokens_seen": 361496576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497091273821464,
      "loss": 1.2856,
      "theoretical_loss": 4.057863452146679,
      "tokens_seen": 361627648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496890672016049,
      "loss": 1.3921,
      "theoretical_loss": 4.057696144014106,
      "tokens_seen": 361758720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044966900702106324,
      "loss": 1.6007,
      "theoretical_loss": 4.057528913455723,
      "tokens_seen": 361889792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496489468405216,
      "loss": 1.6407,
      "theoretical_loss": 4.0573617604074865,
      "tokens_seen": 362020864
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.47562316060066223,
      "objective/train/avg_token_score": 0.012415323406457901,
      "objective/train/docs_used": 216313,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8667166233062744,
      "objective/train/original_loss": 4.416706085205078,
      "objective/train/theoretical_loss": 4.0572782129296785,
      "objective/train/tokens_used": 362086400,
      "objective/train/unlikelihood_loss": 0.12603291869163513,
      "theoretical_loss": 4.0572782129296785,
      "tokens_seen": 362086400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044962888665997996,
      "loss": 1.5666,
      "theoretical_loss": 4.057194684805426,
      "tokens_seen": 362151936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496088264794383,
      "loss": 1.4589,
      "theoretical_loss": 4.0570276865856485,
      "tokens_seen": 362283008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044958876629889674,
      "loss": 1.6308,
      "theoretical_loss": 4.056860765684338,
      "tokens_seen": 362414080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495687061183551,
      "loss": 1.5062,
      "theoretical_loss": 4.056693922037753,
      "tokens_seen": 362545152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044954864593781346,
      "loss": 1.3438,
      "theoretical_loss": 4.056527155582226,
      "tokens_seen": 362676224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495285857572718,
      "loss": 1.3656,
      "theoretical_loss": 4.056360466254168,
      "tokens_seen": 362807296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044950852557673024,
      "loss": 1.5156,
      "theoretical_loss": 4.056193853990063,
      "tokens_seen": 362938368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494884653961886,
      "loss": 1.5408,
      "theoretical_loss": 4.056027318726472,
      "tokens_seen": 363069440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044946840521564697,
      "loss": 1.6677,
      "theoretical_loss": 4.0558608604000295,
      "tokens_seen": 363200512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044944834503510533,
      "loss": 1.4249,
      "theoretical_loss": 4.055694478947445,
      "tokens_seen": 363331584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044942828485456375,
      "loss": 1.4203,
      "theoretical_loss": 4.055528174305505,
      "tokens_seen": 363462656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494082246740221,
      "loss": 1.5913,
      "theoretical_loss": 4.055361946411068,
      "tokens_seen": 363593728
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.37341153621673584,
      "objective/train/avg_token_score": 0.040493469685316086,
      "objective/train/docs_used": 217479,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2378941774368286,
      "objective/train/original_loss": 3.728134870529175,
      "objective/train/theoretical_loss": 4.055195795201069,
      "objective/train/tokens_used": 363724800,
      "objective/train/unlikelihood_loss": 0.2129121869802475,
      "theoretical_loss": 4.055195795201069,
      "tokens_seen": 363724800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044938816449348047,
      "loss": 1.6193,
      "theoretical_loss": 4.055195795201069,
      "tokens_seen": 363724800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044936810431293883,
      "loss": 1.465,
      "theoretical_loss": 4.0550297206125165,
      "tokens_seen": 363855872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493480441323972,
      "loss": 1.3898,
      "theoretical_loss": 4.054863722582494,
      "tokens_seen": 363986944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493279839518556,
      "loss": 1.3853,
      "theoretical_loss": 4.05469780104816,
      "tokens_seen": 364118016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044930792377131397,
      "loss": 1.5886,
      "theoretical_loss": 4.054531955946745,
      "tokens_seen": 364249088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044928786359077234,
      "loss": 1.6761,
      "theoretical_loss": 4.054366187215556,
      "tokens_seen": 364380160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492678034102307,
      "loss": 1.5347,
      "theoretical_loss": 4.054200494791972,
      "tokens_seen": 364511232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492477432296891,
      "loss": 1.3878,
      "theoretical_loss": 4.054034878613447,
      "tokens_seen": 364642304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492276830491475,
      "loss": 1.4606,
      "theoretical_loss": 4.053869338617508,
      "tokens_seen": 364773376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044920762286860584,
      "loss": 1.5716,
      "theoretical_loss": 4.053703874741757,
      "tokens_seen": 364904448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491875626880642,
      "loss": 1.5397,
      "theoretical_loss": 4.053538486923868,
      "tokens_seen": 365035520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044916750250752256,
      "loss": 1.4733,
      "theoretical_loss": 4.053373175101589,
      "tokens_seen": 365166592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449147442326981,
      "loss": 1.5254,
      "theoretical_loss": 4.05320793921274,
      "tokens_seen": 365297664
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.4299853444099426,
      "objective/train/avg_token_score": 0.02031981386244297,
      "objective/train/docs_used": 218231,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5544748306274414,
      "objective/train/original_loss": 4.155126571655273,
      "objective/train/theoretical_loss": 4.053125349723942,
      "objective/train/tokens_used": 365363200,
      "objective/train/unlikelihood_loss": 0.17956046760082245,
      "theoretical_loss": 4.053125349723942,
      "tokens_seen": 365363200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044912738214643934,
      "loss": 1.4319,
      "theoretical_loss": 4.053042779195217,
      "tokens_seen": 365428736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491073219658977,
      "loss": 1.4881,
      "theoretical_loss": 4.052877694986987,
      "tokens_seen": 365559808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044908726178535607,
      "loss": 1.4094,
      "theoretical_loss": 4.05271268652609,
      "tokens_seen": 365690880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490672016048145,
      "loss": 1.671,
      "theoretical_loss": 4.052547753750639,
      "tokens_seen": 365821952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044904714142427284,
      "loss": 1.4686,
      "theoretical_loss": 4.05238289659882,
      "tokens_seen": 365953024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490270812437312,
      "loss": 1.3926,
      "theoretical_loss": 4.052218115008892,
      "tokens_seen": 366084096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044900702106318957,
      "loss": 1.5302,
      "theoretical_loss": 4.0520534089191855,
      "tokens_seen": 366215168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044898696088264793,
      "loss": 1.6902,
      "theoretical_loss": 4.051888778268103,
      "tokens_seen": 366346240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044896690070210635,
      "loss": 1.5773,
      "theoretical_loss": 4.051724222994122,
      "tokens_seen": 366477312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004489468405215647,
      "loss": 1.4207,
      "theoretical_loss": 4.051559743035789,
      "tokens_seen": 366608384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044892678034102307,
      "loss": 1.513,
      "theoretical_loss": 4.051395338331723,
      "tokens_seen": 366739456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044890672016048143,
      "loss": 1.4494,
      "theoretical_loss": 4.051231008820618,
      "tokens_seen": 366870528
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.34115347266197205,
      "objective/train/avg_token_score": 0.10597312450408936,
      "objective/train/docs_used": 219467,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.169438123703003,
      "objective/train/original_loss": 5.253897190093994,
      "objective/train/theoretical_loss": 4.051066754441235,
      "objective/train/tokens_used": 367001600,
      "objective/train/unlikelihood_loss": 0.0806141048669815,
      "theoretical_loss": 4.051066754441235,
      "tokens_seen": 367001600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044888665997993985,
      "loss": 1.3879,
      "theoretical_loss": 4.051066754441235,
      "tokens_seen": 367001600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488665997993982,
      "loss": 1.6292,
      "theoretical_loss": 4.050902575132412,
      "tokens_seen": 367132672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488465396188566,
      "loss": 1.5422,
      "theoretical_loss": 4.050738470833054,
      "tokens_seen": 367263744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044882647943831494,
      "loss": 1.5571,
      "theoretical_loss": 4.05057444148214,
      "tokens_seen": 367394816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488064192577733,
      "loss": 1.5531,
      "theoretical_loss": 4.05041048701872,
      "tokens_seen": 367525888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487863590772317,
      "loss": 1.4186,
      "theoretical_loss": 4.050246607381913,
      "tokens_seen": 367656960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487662988966901,
      "loss": 1.6269,
      "theoretical_loss": 4.050082802510915,
      "tokens_seen": 367788032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044874623871614844,
      "loss": 1.5642,
      "theoretical_loss": 4.049919072344987,
      "tokens_seen": 367919104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487261785356068,
      "loss": 1.5353,
      "theoretical_loss": 4.049755416823463,
      "tokens_seen": 368050176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487061183550652,
      "loss": 1.4776,
      "theoretical_loss": 4.0495918358857494,
      "tokens_seen": 368181248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486860581745236,
      "loss": 1.465,
      "theoretical_loss": 4.04942832947132,
      "tokens_seen": 368312320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044866599799398194,
      "loss": 1.5645,
      "theoretical_loss": 4.049264897519723,
      "tokens_seen": 368443392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486459378134403,
      "loss": 1.425,
      "theoretical_loss": 4.049101539970576,
      "tokens_seen": 368574464
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.47611191868782043,
      "objective/train/avg_token_score": 0.012634911574423313,
      "objective/train/docs_used": 220098,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.645254135131836,
      "objective/train/original_loss": 4.079015731811523,
      "objective/train/theoretical_loss": 4.04901988907807,
      "objective/train/tokens_used": 368640000,
      "objective/train/unlikelihood_loss": 0.1734001785516739,
      "theoretical_loss": 4.04901988907807,
      "tokens_seen": 368640000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044862587763289867,
      "loss": 1.7137,
      "theoretical_loss": 4.048938256763565,
      "tokens_seen": 368705536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486058174523571,
      "loss": 1.6932,
      "theoretical_loss": 4.048775047838447,
      "tokens_seen": 368836608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044858575727181545,
      "loss": 1.705,
      "theoretical_loss": 4.048611913135052,
      "tokens_seen": 368967680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485656970912738,
      "loss": 1.7297,
      "theoretical_loss": 4.048448852593277,
      "tokens_seen": 369098752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044854563691073217,
      "loss": 1.7156,
      "theoretical_loss": 4.0482858661530905,
      "tokens_seen": 369229824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485255767301906,
      "loss": 1.4436,
      "theoretical_loss": 4.04812295375453,
      "tokens_seen": 369360896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044850551654964895,
      "loss": 1.5983,
      "theoretical_loss": 4.047960115337702,
      "tokens_seen": 369491968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484854563691073,
      "loss": 1.5025,
      "theoretical_loss": 4.047797350842786,
      "tokens_seen": 369623040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484653961885657,
      "loss": 1.5012,
      "theoretical_loss": 4.0476346602100275,
      "tokens_seen": 369754112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044844533600802404,
      "loss": 1.5321,
      "theoretical_loss": 4.047472043379743,
      "tokens_seen": 369885184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044842527582748245,
      "loss": 1.3779,
      "theoretical_loss": 4.04730950029232,
      "tokens_seen": 370016256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484052156469408,
      "loss": 1.7209,
      "theoretical_loss": 4.04714703088821,
      "tokens_seen": 370147328
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.4204545319080353,
      "objective/train/avg_token_score": 0.013336265459656715,
      "objective/train/docs_used": 221122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5654873847961426,
      "objective/train/original_loss": 4.028966903686523,
      "objective/train/theoretical_loss": 4.04698463510794,
      "objective/train/tokens_used": 370278400,
      "objective/train/unlikelihood_loss": 0.19616243243217468,
      "theoretical_loss": 4.04698463510794,
      "tokens_seen": 370278400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483851554663992,
      "loss": 1.3742,
      "theoretical_loss": 4.04698463510794,
      "tokens_seen": 370278400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044836509528585754,
      "loss": 1.3853,
      "theoretical_loss": 4.046822312892102,
      "tokens_seen": 370409472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044834503510531596,
      "loss": 1.5133,
      "theoretical_loss": 4.046660064181359,
      "tokens_seen": 370540544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483249749247743,
      "loss": 1.4851,
      "theoretical_loss": 4.0464978889164405,
      "tokens_seen": 370671616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483049147442327,
      "loss": 1.5003,
      "theoretical_loss": 4.046335787038148,
      "tokens_seen": 370802688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044828485456369104,
      "loss": 1.6905,
      "theoretical_loss": 4.046173758487347,
      "tokens_seen": 370933760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482647943831494,
      "loss": 1.4056,
      "theoretical_loss": 4.046011803204976,
      "tokens_seen": 371064832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482447342026078,
      "loss": 1.4829,
      "theoretical_loss": 4.045849921132042,
      "tokens_seen": 371195904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482246740220662,
      "loss": 1.6773,
      "theoretical_loss": 4.045688112209615,
      "tokens_seen": 371326976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044820461384152455,
      "loss": 1.4661,
      "theoretical_loss": 4.04552637637884,
      "tokens_seen": 371458048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044818455366098296,
      "loss": 1.6595,
      "theoretical_loss": 4.045364713580923,
      "tokens_seen": 371589120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481644934804414,
      "loss": 1.6206,
      "theoretical_loss": 4.045203123757146,
      "tokens_seen": 371720192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044814443329989974,
      "loss": 1.6652,
      "theoretical_loss": 4.0450416068488515,
      "tokens_seen": 371851264
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.4794110357761383,
      "objective/train/avg_token_score": 0.004854011349380016,
      "objective/train/docs_used": 221664,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7908248901367188,
      "objective/train/original_loss": 3.7173187732696533,
      "objective/train/theoretical_loss": 4.044960875719701,
      "objective/train/tokens_used": 371916800,
      "objective/train/unlikelihood_loss": 0.17050793766975403,
      "theoretical_loss": 4.044960875719701,
      "tokens_seen": 371916800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481243731193581,
      "loss": 1.493,
      "theoretical_loss": 4.0448801627974555,
      "tokens_seen": 371982336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044810431293881647,
      "loss": 1.5649,
      "theoretical_loss": 4.044718791544438,
      "tokens_seen": 372113408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480842527582749,
      "loss": 1.7985,
      "theoretical_loss": 4.044557493031347,
      "tokens_seen": 372244480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044806419257773324,
      "loss": 1.5757,
      "theoretical_loss": 4.0443962671998,
      "tokens_seen": 372375552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480441323971916,
      "loss": 1.5035,
      "theoretical_loss": 4.04423511399148,
      "tokens_seen": 372506624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044802407221664997,
      "loss": 1.5995,
      "theoretical_loss": 4.044074033348139,
      "tokens_seen": 372637696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044800401203610833,
      "loss": 1.7926,
      "theoretical_loss": 4.0439130252115945,
      "tokens_seen": 372768768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044798395185556675,
      "loss": 1.7443,
      "theoretical_loss": 4.043752089523732,
      "tokens_seen": 372899840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004479638916750251,
      "loss": 1.6664,
      "theoretical_loss": 4.043591226226503,
      "tokens_seen": 373030912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044794383149448347,
      "loss": 1.55,
      "theoretical_loss": 4.043430435261927,
      "tokens_seen": 373161984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044792377131394183,
      "loss": 1.5638,
      "theoretical_loss": 4.043269716572091,
      "tokens_seen": 373293056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044790371113340025,
      "loss": 1.5155,
      "theoretical_loss": 4.043109070099146,
      "tokens_seen": 373424128
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.5913978219032288,
      "objective/train/avg_token_score": 0.00555779505521059,
      "objective/train/docs_used": 222906,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1588377952575684,
      "objective/train/original_loss": 3.770258665084839,
      "objective/train/theoretical_loss": 4.042948495785312,
      "objective/train/tokens_used": 373555200,
      "objective/train/unlikelihood_loss": 0.16115084290504456,
      "theoretical_loss": 4.042948495785312,
      "tokens_seen": 373555200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478836509528586,
      "loss": 1.5525,
      "theoretical_loss": 4.042948495785312,
      "tokens_seen": 373555200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447863590772317,
      "loss": 1.5479,
      "theoretical_loss": 4.042787993572875,
      "tokens_seen": 373686272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044784353059177534,
      "loss": 1.4116,
      "theoretical_loss": 4.042627563404187,
      "tokens_seen": 373817344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478234704112337,
      "loss": 1.5096,
      "theoretical_loss": 4.042467205221665,
      "tokens_seen": 373948416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478034102306921,
      "loss": 1.5331,
      "theoretical_loss": 4.042306918967795,
      "tokens_seen": 374079488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477833500501505,
      "loss": 1.6504,
      "theoretical_loss": 4.042146704585127,
      "tokens_seen": 374210560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044776328986960884,
      "loss": 1.6323,
      "theoretical_loss": 4.041986562016278,
      "tokens_seen": 374341632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477432296890672,
      "loss": 1.4255,
      "theoretical_loss": 4.0418264912039295,
      "tokens_seen": 374472704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477231695085256,
      "loss": 1.7105,
      "theoretical_loss": 4.041666492090831,
      "tokens_seen": 374603776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447703109327984,
      "loss": 1.4864,
      "theoretical_loss": 4.041506564619796,
      "tokens_seen": 374734848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044768304914744234,
      "loss": 1.5879,
      "theoretical_loss": 4.041346708733704,
      "tokens_seen": 374865920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476629889669007,
      "loss": 1.6726,
      "theoretical_loss": 4.041186924375499,
      "tokens_seen": 374996992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044764292878635907,
      "loss": 1.7443,
      "theoretical_loss": 4.041027211488194,
      "tokens_seen": 375128064
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.2611192464828491,
      "objective/train/avg_token_score": 0.02251238189637661,
      "objective/train/docs_used": 223164,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0515928268432617,
      "objective/train/original_loss": 4.479992389678955,
      "objective/train/theoretical_loss": 4.040947381828337,
      "objective/train/tokens_used": 375193600,
      "objective/train/unlikelihood_loss": 0.1949537694454193,
      "theoretical_loss": 4.040947381828337,
      "tokens_seen": 375193600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476228686058175,
      "loss": 1.406,
      "theoretical_loss": 4.040867570014862,
      "tokens_seen": 375259136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044760280842527585,
      "loss": 1.4454,
      "theoretical_loss": 4.040707999898647,
      "tokens_seen": 375390208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475827482447342,
      "loss": 1.7935,
      "theoretical_loss": 4.0405485010827515,
      "tokens_seen": 375521280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044756268806419257,
      "loss": 1.5507,
      "theoretical_loss": 4.040389073510449,
      "tokens_seen": 375652352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447542627883651,
      "loss": 1.3859,
      "theoretical_loss": 4.040229717125074,
      "tokens_seen": 375783424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044752256770310935,
      "loss": 1.6392,
      "theoretical_loss": 4.040070431870029,
      "tokens_seen": 375914496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475025075225677,
      "loss": 1.5432,
      "theoretical_loss": 4.039911217688777,
      "tokens_seen": 376045568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474824473420261,
      "loss": 1.4319,
      "theoretical_loss": 4.039752074524851,
      "tokens_seen": 376176640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044746238716148444,
      "loss": 1.5912,
      "theoretical_loss": 4.039593002321844,
      "tokens_seen": 376307712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044744232698094285,
      "loss": 1.5599,
      "theoretical_loss": 4.039434001023415,
      "tokens_seen": 376438784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474222668004012,
      "loss": 1.3618,
      "theoretical_loss": 4.039275070573288,
      "tokens_seen": 376569856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474022066198596,
      "loss": 1.5952,
      "theoretical_loss": 4.03911621091525,
      "tokens_seen": 376700928
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.4695136845111847,
      "objective/train/avg_token_score": 0.009801458567380905,
      "objective/train/docs_used": 224842,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.732696771621704,
      "objective/train/original_loss": 4.10999059677124,
      "objective/train/theoretical_loss": 4.038957421993153,
      "objective/train/tokens_used": 376832000,
      "objective/train/unlikelihood_loss": 0.14230914413928986,
      "theoretical_loss": 4.038957421993153,
      "tokens_seen": 376832000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044738214643931794,
      "loss": 1.7543,
      "theoretical_loss": 4.038957421993153,
      "tokens_seen": 376832000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044736208625877636,
      "loss": 1.4096,
      "theoretical_loss": 4.038798703750913,
      "tokens_seen": 376963072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473420260782347,
      "loss": 1.6097,
      "theoretical_loss": 4.038640056132509,
      "tokens_seen": 377094144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473219658976931,
      "loss": 1.6665,
      "theoretical_loss": 4.0384814790819865,
      "tokens_seen": 377225216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044730190571715144,
      "loss": 1.5166,
      "theoretical_loss": 4.038322972543451,
      "tokens_seen": 377356288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472818455366098,
      "loss": 1.7482,
      "theoretical_loss": 4.038164536461075,
      "tokens_seen": 377487360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472617853560682,
      "loss": 1.5796,
      "theoretical_loss": 4.038006170779091,
      "tokens_seen": 377618432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472417251755266,
      "loss": 1.4874,
      "theoretical_loss": 4.037847875441798,
      "tokens_seen": 377749504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044722166499498495,
      "loss": 1.6282,
      "theoretical_loss": 4.037689650393559,
      "tokens_seen": 377880576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472016048144433,
      "loss": 1.6695,
      "theoretical_loss": 4.037531495578797,
      "tokens_seen": 378011648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471815446339017,
      "loss": 1.4349,
      "theoretical_loss": 4.037373410942001,
      "tokens_seen": 378142720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471614844533601,
      "loss": 1.7403,
      "theoretical_loss": 4.037215396427721,
      "tokens_seen": 378273792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044714142427281845,
      "loss": 1.5023,
      "theoretical_loss": 4.037057451980571,
      "tokens_seen": 378404864
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.4757453501224518,
      "objective/train/avg_token_score": 0.021771633997559547,
      "objective/train/docs_used": 225570,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.804304599761963,
      "objective/train/original_loss": 4.102111339569092,
      "objective/train/theoretical_loss": 4.036978506014879,
      "objective/train/tokens_used": 378470400,
      "objective/train/unlikelihood_loss": 0.18979206681251526,
      "theoretical_loss": 4.036978506014879,
      "tokens_seen": 378470400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471213640922768,
      "loss": 1.7116,
      "theoretical_loss": 4.036899577545229,
      "tokens_seen": 378535936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044710130391173523,
      "loss": 1.6378,
      "theoretical_loss": 4.036741773066434,
      "tokens_seen": 378667008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470812437311936,
      "loss": 1.6524,
      "theoretical_loss": 4.036584038488988,
      "tokens_seen": 378798080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044706118355065195,
      "loss": 1.4253,
      "theoretical_loss": 4.036426373757759,
      "tokens_seen": 378929152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470411233701103,
      "loss": 1.6715,
      "theoretical_loss": 4.0362687788176705,
      "tokens_seen": 379060224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470210631895687,
      "loss": 1.8991,
      "theoretical_loss": 4.036111253613715,
      "tokens_seen": 379191296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470010030090271,
      "loss": 1.6512,
      "theoretical_loss": 4.0359537980909455,
      "tokens_seen": 379322368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044698094282848546,
      "loss": 1.4178,
      "theoretical_loss": 4.0357964121944745,
      "tokens_seen": 379453440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469608826479438,
      "loss": 1.5667,
      "theoretical_loss": 4.035639095869481,
      "tokens_seen": 379584512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469408224674022,
      "loss": 1.5906,
      "theoretical_loss": 4.0354818490612026,
      "tokens_seen": 379715584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469207622868606,
      "loss": 1.5448,
      "theoretical_loss": 4.035324671714942,
      "tokens_seen": 379846656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044690070210631896,
      "loss": 1.6204,
      "theoretical_loss": 4.03516756377606,
      "tokens_seen": 379977728
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.4055474102497101,
      "objective/train/avg_token_score": 0.01529501099139452,
      "objective/train/docs_used": 226798,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4889705181121826,
      "objective/train/original_loss": 3.754312515258789,
      "objective/train/theoretical_loss": 4.035010525189982,
      "objective/train/tokens_used": 380108800,
      "objective/train/unlikelihood_loss": 0.2511163651943207,
      "theoretical_loss": 4.035010525189982,
      "tokens_seen": 380108800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468806419257773,
      "loss": 1.5714,
      "theoretical_loss": 4.035010525189982,
      "tokens_seen": 380108800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468605817452357,
      "loss": 1.3369,
      "theoretical_loss": 4.0348535559021945,
      "tokens_seen": 380239872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044684052156469405,
      "loss": 1.5174,
      "theoretical_loss": 4.034696655858246,
      "tokens_seen": 380370944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044682046138415246,
      "loss": 1.607,
      "theoretical_loss": 4.034539825003746,
      "tokens_seen": 380502016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468004012036108,
      "loss": 1.5884,
      "theoretical_loss": 4.034383063284364,
      "tokens_seen": 380633088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467803410230692,
      "loss": 1.5842,
      "theoretical_loss": 4.0342263706458334,
      "tokens_seen": 380764160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044676028084252755,
      "loss": 1.6033,
      "theoretical_loss": 4.034069747033948,
      "tokens_seen": 380895232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044674022066198597,
      "loss": 1.5395,
      "theoretical_loss": 4.033913192394563,
      "tokens_seen": 381026304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044672016048144433,
      "loss": 1.7102,
      "theoretical_loss": 4.033756706673593,
      "tokens_seen": 381157376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467001003009027,
      "loss": 1.5634,
      "theoretical_loss": 4.033600289817015,
      "tokens_seen": 381288448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466800401203611,
      "loss": 1.4824,
      "theoretical_loss": 4.033443941770867,
      "tokens_seen": 381419520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044665997993981947,
      "loss": 1.6288,
      "theoretical_loss": 4.033287662481247,
      "tokens_seen": 381550592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466399197592779,
      "loss": 1.653,
      "theoretical_loss": 4.033131451894317,
      "tokens_seen": 381681664
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.4856427311897278,
      "objective/train/avg_token_score": 0.012365730479359627,
      "objective/train/docs_used": 227456,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6744704246520996,
      "objective/train/original_loss": 3.758188486099243,
      "objective/train/theoretical_loss": 4.03305337234755,
      "objective/train/tokens_used": 381747200,
      "objective/train/unlikelihood_loss": 0.1711551696062088,
      "theoretical_loss": 4.03305337234755,
      "tokens_seen": 381747200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044661985957873625,
      "loss": 1.5343,
      "theoretical_loss": 4.032975309956292,
      "tokens_seen": 381812736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465997993981946,
      "loss": 1.5904,
      "theoretical_loss": 4.032819236613457,
      "tokens_seen": 381943808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044657973921765297,
      "loss": 1.4637,
      "theoretical_loss": 4.03266323181215,
      "tokens_seen": 382074880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465596790371114,
      "loss": 1.5051,
      "theoretical_loss": 4.032507295498774,
      "tokens_seen": 382205952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044653961885656975,
      "loss": 1.6082,
      "theoretical_loss": 4.032351427619789,
      "tokens_seen": 382337024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465195586760281,
      "loss": 1.5642,
      "theoretical_loss": 4.032195628121718,
      "tokens_seen": 382468096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464994984954865,
      "loss": 1.5348,
      "theoretical_loss": 4.032039896951142,
      "tokens_seen": 382599168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044647943831494484,
      "loss": 1.6252,
      "theoretical_loss": 4.0318842340547025,
      "tokens_seen": 382730240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044645937813440325,
      "loss": 1.4619,
      "theoretical_loss": 4.031728639379103,
      "tokens_seen": 382861312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464393179538616,
      "loss": 1.4193,
      "theoretical_loss": 4.031573112871103,
      "tokens_seen": 382992384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044641925777332,
      "loss": 1.7358,
      "theoretical_loss": 4.031417654477526,
      "tokens_seen": 383123456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044639919759277834,
      "loss": 1.6106,
      "theoretical_loss": 4.03126226414525,
      "tokens_seen": 383254528
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.4290078282356262,
      "objective/train/avg_token_score": 0.0078100161626935005,
      "objective/train/docs_used": 228876,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6035910844802856,
      "objective/train/original_loss": 4.235310077667236,
      "objective/train/theoretical_loss": 4.031106941821218,
      "objective/train/tokens_used": 383385600,
      "objective/train/unlikelihood_loss": 0.13287082314491272,
      "theoretical_loss": 4.031106941821218,
      "tokens_seen": 383385600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044637913741223676,
      "loss": 1.7171,
      "theoretical_loss": 4.031106941821218,
      "tokens_seen": 383385600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463590772316951,
      "loss": 1.6425,
      "theoretical_loss": 4.030951687452429,
      "tokens_seen": 383516672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463390170511535,
      "loss": 1.5361,
      "theoretical_loss": 4.030796500985943,
      "tokens_seen": 383647744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044631895687061184,
      "loss": 1.6445,
      "theoretical_loss": 4.030641382368878,
      "tokens_seen": 383778816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462988966900702,
      "loss": 1.5612,
      "theoretical_loss": 4.030486331548413,
      "tokens_seen": 383909888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462788365095286,
      "loss": 1.649,
      "theoretical_loss": 4.030331348471784,
      "tokens_seen": 384040960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446258776328987,
      "loss": 1.4784,
      "theoretical_loss": 4.030176433086289,
      "tokens_seen": 384172032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044623871614844535,
      "loss": 1.4671,
      "theoretical_loss": 4.030021585339281,
      "tokens_seen": 384303104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462186559679037,
      "loss": 1.7199,
      "theoretical_loss": 4.029866805178175,
      "tokens_seen": 384434176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461985957873621,
      "loss": 1.4296,
      "theoretical_loss": 4.029712092550443,
      "tokens_seen": 384565248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461785356068205,
      "loss": 1.4691,
      "theoretical_loss": 4.029557447403618,
      "tokens_seen": 384696320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044615847542627885,
      "loss": 1.4171,
      "theoretical_loss": 4.029402869685288,
      "tokens_seen": 384827392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461384152457372,
      "loss": 1.4194,
      "theoretical_loss": 4.029248359343105,
      "tokens_seen": 384958464
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.589442789554596,
      "objective/train/avg_token_score": 0.008675182238221169,
      "objective/train/docs_used": 229454,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1350181102752686,
      "objective/train/original_loss": 3.9059112071990967,
      "objective/train/theoretical_loss": 4.029171129421724,
      "objective/train/tokens_used": 385024000,
      "objective/train/unlikelihood_loss": 0.1192452609539032,
      "theoretical_loss": 4.029171129421724,
      "tokens_seen": 385024000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044611835506519563,
      "loss": 1.8329,
      "theoretical_loss": 4.029093916324774,
      "tokens_seen": 385089536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446098294884654,
      "loss": 1.5253,
      "theoretical_loss": 4.02893954057806,
      "tokens_seen": 385220608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044607823470411235,
      "loss": 1.5488,
      "theoretical_loss": 4.0287852320507875,
      "tokens_seen": 385351680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460581745235707,
      "loss": 1.6576,
      "theoretical_loss": 4.028630990690838,
      "tokens_seen": 385482752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460381143430291,
      "loss": 1.5913,
      "theoretical_loss": 4.028476816446153,
      "tokens_seen": 385613824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460180541624875,
      "loss": 1.5977,
      "theoretical_loss": 4.028322709264729,
      "tokens_seen": 385744896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044599799398194586,
      "loss": 1.4783,
      "theoretical_loss": 4.028168669094623,
      "tokens_seen": 385875968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459779338014042,
      "loss": 1.3898,
      "theoretical_loss": 4.0280146958839484,
      "tokens_seen": 386007040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459578736208626,
      "loss": 1.3856,
      "theoretical_loss": 4.027860789580878,
      "tokens_seen": 386138112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445937813440321,
      "loss": 1.3978,
      "theoretical_loss": 4.0277069501336396,
      "tokens_seen": 386269184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044591775325977936,
      "loss": 1.4399,
      "theoretical_loss": 4.027553177490521,
      "tokens_seen": 386400256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458976930792377,
      "loss": 1.5377,
      "theoretical_loss": 4.027399471599867,
      "tokens_seen": 386531328
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.48521506786346436,
      "objective/train/avg_token_score": 0.010920585133135319,
      "objective/train/docs_used": 230121,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8535081148147583,
      "objective/train/original_loss": 4.084407329559326,
      "objective/train/theoretical_loss": 4.027245832410079,
      "objective/train/tokens_used": 386662400,
      "objective/train/unlikelihood_loss": 0.12329123914241791,
      "theoretical_loss": 4.027245832410079,
      "tokens_seen": 386662400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458776328986961,
      "loss": 1.6056,
      "theoretical_loss": 4.027245832410079,
      "tokens_seen": 386662400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044585757271815445,
      "loss": 1.4379,
      "theoretical_loss": 4.027092259869616,
      "tokens_seen": 386793472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044583751253761286,
      "loss": 1.6012,
      "theoretical_loss": 4.026938753926997,
      "tokens_seen": 386924544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458174523570712,
      "loss": 1.5904,
      "theoretical_loss": 4.026785314530792,
      "tokens_seen": 387055616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457973921765296,
      "loss": 1.5274,
      "theoretical_loss": 4.026631941629635,
      "tokens_seen": 387186688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044577733199598795,
      "loss": 1.6511,
      "theoretical_loss": 4.026478635172212,
      "tokens_seen": 387317760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044575727181544637,
      "loss": 1.4566,
      "theoretical_loss": 4.0263253951072695,
      "tokens_seen": 387448832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044573721163490473,
      "loss": 1.3738,
      "theoretical_loss": 4.026172221383607,
      "tokens_seen": 387579904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457171514543631,
      "loss": 1.3676,
      "theoretical_loss": 4.026019113950086,
      "tokens_seen": 387710976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044569709127382145,
      "loss": 1.5872,
      "theoretical_loss": 4.025866072755619,
      "tokens_seen": 387842048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456770310932798,
      "loss": 1.509,
      "theoretical_loss": 4.02571309774918,
      "tokens_seen": 387973120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044565697091273823,
      "loss": 1.4532,
      "theoretical_loss": 4.025560188879796,
      "tokens_seen": 388104192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456369107321966,
      "loss": 1.572,
      "theoretical_loss": 4.025407346096551,
      "tokens_seen": 388235264
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.3402370512485504,
      "objective/train/avg_token_score": 0.010277241468429565,
      "objective/train/docs_used": 231531,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1777230501174927,
      "objective/train/original_loss": 4.140812397003174,
      "objective/train/theoretical_loss": 4.025330949471336,
      "objective/train/tokens_used": 388300800,
      "objective/train/unlikelihood_loss": 0.12250532954931259,
      "theoretical_loss": 4.025330949471336,
      "tokens_seen": 388300800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044561685055165496,
      "loss": 1.5369,
      "theoretical_loss": 4.025254569348588,
      "tokens_seen": 388366336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455967903711133,
      "loss": 1.5416,
      "theoretical_loss": 4.025101858585104,
      "tokens_seen": 388497408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044557673019057173,
      "loss": 1.5914,
      "theoretical_loss": 4.0249492137553515,
      "tokens_seen": 388628480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455566700100301,
      "loss": 1.3606,
      "theoretical_loss": 4.0247966348086415,
      "tokens_seen": 388759552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044553660982948846,
      "loss": 1.5075,
      "theoretical_loss": 4.02464412169434,
      "tokens_seen": 388890624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455165496489468,
      "loss": 1.5384,
      "theoretical_loss": 4.024491674361868,
      "tokens_seen": 389021696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454964894684052,
      "loss": 1.4394,
      "theoretical_loss": 4.024339292760703,
      "tokens_seen": 389152768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454764292878636,
      "loss": 1.5151,
      "theoretical_loss": 4.0241869768403795,
      "tokens_seen": 389283840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044545636910732196,
      "loss": 1.6464,
      "theoretical_loss": 4.024034726550486,
      "tokens_seen": 389414912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454363089267803,
      "loss": 1.5365,
      "theoretical_loss": 4.023882541840667,
      "tokens_seen": 389545984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454162487462387,
      "loss": 1.4798,
      "theoretical_loss": 4.023730422660624,
      "tokens_seen": 389677056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453961885656971,
      "loss": 1.4354,
      "theoretical_loss": 4.023578368960113,
      "tokens_seen": 389808128
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.34365835785865784,
      "objective/train/avg_token_score": 0.01555408164858818,
      "objective/train/docs_used": 232298,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2046136856079102,
      "objective/train/original_loss": 3.8979644775390625,
      "objective/train/theoretical_loss": 4.023426380688943,
      "objective/train/tokens_used": 389939200,
      "objective/train/unlikelihood_loss": 0.1804252713918686,
      "theoretical_loss": 4.023426380688943,
      "tokens_seen": 389939200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044537612838515547,
      "loss": 1.3799,
      "theoretical_loss": 4.023426380688943,
      "tokens_seen": 389939200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044535606820461383,
      "loss": 1.6791,
      "theoretical_loss": 4.023274457796983,
      "tokens_seen": 390070272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453360080240722,
      "loss": 1.45,
      "theoretical_loss": 4.023122600234155,
      "tokens_seen": 390201344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044531594784353055,
      "loss": 1.5732,
      "theoretical_loss": 4.022970807950434,
      "tokens_seen": 390332416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044529588766298897,
      "loss": 1.5985,
      "theoretical_loss": 4.022819080895853,
      "tokens_seen": 390463488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044527582748244733,
      "loss": 1.5921,
      "theoretical_loss": 4.022667419020499,
      "tokens_seen": 390594560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452557673019057,
      "loss": 1.5946,
      "theoretical_loss": 4.022515822274515,
      "tokens_seen": 390725632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044523570712136406,
      "loss": 1.5798,
      "theoretical_loss": 4.022364290608098,
      "tokens_seen": 390856704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044521564694082247,
      "loss": 1.4749,
      "theoretical_loss": 4.022212823971498,
      "tokens_seen": 390987776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044519558676028083,
      "loss": 1.7443,
      "theoretical_loss": 4.022061422315023,
      "tokens_seen": 391118848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044517552657973925,
      "loss": 1.5852,
      "theoretical_loss": 4.021910085589032,
      "tokens_seen": 391249920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451554663991976,
      "loss": 1.6875,
      "theoretical_loss": 4.021758813743942,
      "tokens_seen": 391380992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044513540621865603,
      "loss": 1.4254,
      "theoretical_loss": 4.021607606730223,
      "tokens_seen": 391512064
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.3057795763015747,
      "objective/train/avg_token_score": 0.016242092475295067,
      "objective/train/docs_used": 233436,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1886520385742188,
      "objective/train/original_loss": 4.423647880554199,
      "objective/train/theoretical_loss": 4.0215320275196635,
      "objective/train/tokens_used": 391577600,
      "objective/train/unlikelihood_loss": 0.1317605823278427,
      "theoretical_loss": 4.0215320275196635,
      "tokens_seen": 391577600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451153460381144,
      "loss": 1.3848,
      "theoretical_loss": 4.0214564644983986,
      "tokens_seen": 391643136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044509528585757275,
      "loss": 1.7228,
      "theoretical_loss": 4.021305386999048,
      "tokens_seen": 391774208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450752256770311,
      "loss": 1.475,
      "theoretical_loss": 4.021154374182803,
      "tokens_seen": 391905280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450551654964895,
      "loss": 1.4708,
      "theoretical_loss": 4.021003426000353,
      "tokens_seen": 392036352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450351053159479,
      "loss": 1.7547,
      "theoretical_loss": 4.0208525424024355,
      "tokens_seen": 392167424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044501504513540626,
      "loss": 1.5899,
      "theoretical_loss": 4.020701723339848,
      "tokens_seen": 392298496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449949849548646,
      "loss": 1.4629,
      "theoretical_loss": 4.020550968763439,
      "tokens_seen": 392429568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444974924774323,
      "loss": 1.5781,
      "theoretical_loss": 4.020400278624112,
      "tokens_seen": 392560640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449548645937814,
      "loss": 1.3149,
      "theoretical_loss": 4.020249652872822,
      "tokens_seen": 392691712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044493480441323976,
      "loss": 1.7288,
      "theoretical_loss": 4.02009909146058,
      "tokens_seen": 392822784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449147442326981,
      "loss": 1.6467,
      "theoretical_loss": 4.01994859433845,
      "tokens_seen": 392953856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448946840521565,
      "loss": 1.7551,
      "theoretical_loss": 4.019798161457548,
      "tokens_seen": 393084928
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.4335899353027344,
      "objective/train/avg_token_score": 0.016622694209218025,
      "objective/train/docs_used": 234056,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6204777956008911,
      "objective/train/original_loss": 3.885645866394043,
      "objective/train/theoretical_loss": 4.019647792769048,
      "objective/train/tokens_used": 393216000,
      "objective/train/unlikelihood_loss": 0.16480211913585663,
      "theoretical_loss": 4.019647792769048,
      "tokens_seen": 393216000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044487462387161485,
      "loss": 1.5896,
      "theoretical_loss": 4.019647792769048,
      "tokens_seen": 393216000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044485456369107326,
      "loss": 1.5453,
      "theoretical_loss": 4.01949748822417,
      "tokens_seen": 393347072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448345035105316,
      "loss": 1.4782,
      "theoretical_loss": 4.019347247774195,
      "tokens_seen": 393478144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044481444332999,
      "loss": 1.4937,
      "theoretical_loss": 4.019197071370451,
      "tokens_seen": 393609216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044479438314944835,
      "loss": 1.438,
      "theoretical_loss": 4.019046958964324,
      "tokens_seen": 393740288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044477432296890677,
      "loss": 1.3938,
      "theoretical_loss": 4.01889691050725,
      "tokens_seen": 393871360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044475426278836513,
      "loss": 1.7109,
      "theoretical_loss": 4.018746925950719,
      "tokens_seen": 394002432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447342026078235,
      "loss": 1.3604,
      "theoretical_loss": 4.018597005246273,
      "tokens_seen": 394133504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044471414242728185,
      "loss": 1.6068,
      "theoretical_loss": 4.01844714834551,
      "tokens_seen": 394264576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446940822467402,
      "loss": 1.5331,
      "theoretical_loss": 4.018297355200076,
      "tokens_seen": 394395648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044467402206619863,
      "loss": 1.465,
      "theoretical_loss": 4.018147625761673,
      "tokens_seen": 394526720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444653961885657,
      "loss": 1.5658,
      "theoretical_loss": 4.017997959982055,
      "tokens_seen": 394657792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044463390170511536,
      "loss": 1.7217,
      "theoretical_loss": 4.017848357813028,
      "tokens_seen": 394788864
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.45723363757133484,
      "objective/train/avg_token_score": 0.017939608544111252,
      "objective/train/docs_used": 235431,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7509286403656006,
      "objective/train/original_loss": 4.263587474822998,
      "objective/train/theoretical_loss": 4.01777358056744,
      "objective/train/tokens_used": 394854400,
      "objective/train/unlikelihood_loss": 0.13543236255645752,
      "theoretical_loss": 4.01777358056744,
      "tokens_seen": 394854400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446138415245737,
      "loss": 1.635,
      "theoretical_loss": 4.017698819206451,
      "tokens_seen": 394919936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044459378134403214,
      "loss": 1.5588,
      "theoretical_loss": 4.017549344114236,
      "tokens_seen": 395051008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445737211634905,
      "loss": 1.5377,
      "theoretical_loss": 4.0173999324883445,
      "tokens_seen": 395182080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044455366098294886,
      "loss": 1.5516,
      "theoretical_loss": 4.017250584280794,
      "tokens_seen": 395313152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445336008024072,
      "loss": 1.5806,
      "theoretical_loss": 4.0171012994436515,
      "tokens_seen": 395444224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445135406218656,
      "loss": 1.574,
      "theoretical_loss": 4.016952077929037,
      "tokens_seen": 395575296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444493480441324,
      "loss": 1.7532,
      "theoretical_loss": 4.016802919689124,
      "tokens_seen": 395706368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044447342026078236,
      "loss": 1.4473,
      "theoretical_loss": 4.016653824676135,
      "tokens_seen": 395837440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444533600802407,
      "loss": 1.5449,
      "theoretical_loss": 4.016504792842346,
      "tokens_seen": 395968512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444332998996991,
      "loss": 1.7389,
      "theoretical_loss": 4.016355824140085,
      "tokens_seen": 396099584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444132397191575,
      "loss": 1.6257,
      "theoretical_loss": 4.016206918521731,
      "tokens_seen": 396230656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044439317953861587,
      "loss": 1.5358,
      "theoretical_loss": 4.016058075939716,
      "tokens_seen": 396361728
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.3450024425983429,
      "objective/train/avg_token_score": 0.024041682481765747,
      "objective/train/docs_used": 235840,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.240441083908081,
      "objective/train/original_loss": 3.7224650382995605,
      "objective/train/theoretical_loss": 4.015909296346521,
      "objective/train/tokens_used": 396492800,
      "objective/train/unlikelihood_loss": 0.21158063411712646,
      "theoretical_loss": 4.015909296346521,
      "tokens_seen": 396492800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044437311935807423,
      "loss": 1.5572,
      "theoretical_loss": 4.015909296346521,
      "tokens_seen": 396492800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443530591775326,
      "loss": 1.5529,
      "theoretical_loss": 4.015760579694683,
      "tokens_seen": 396623872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044433299899699095,
      "loss": 1.5808,
      "theoretical_loss": 4.015611925936785,
      "tokens_seen": 396754944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044431293881644937,
      "loss": 1.449,
      "theoretical_loss": 4.015463335025465,
      "tokens_seen": 396886016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044429287863590773,
      "loss": 1.3055,
      "theoretical_loss": 4.015314806913411,
      "tokens_seen": 397017088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442728184553661,
      "loss": 1.3535,
      "theoretical_loss": 4.015166341553362,
      "tokens_seen": 397148160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044425275827482446,
      "loss": 1.536,
      "theoretical_loss": 4.01501793889811,
      "tokens_seen": 397279232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044423269809428287,
      "loss": 1.4453,
      "theoretical_loss": 4.014869598900496,
      "tokens_seen": 397410304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044421263791374123,
      "loss": 1.6041,
      "theoretical_loss": 4.014721321513412,
      "tokens_seen": 397541376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441925777331996,
      "loss": 1.6268,
      "theoretical_loss": 4.014573106689803,
      "tokens_seen": 397672448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044417251755265796,
      "loss": 1.5815,
      "theoretical_loss": 4.014424954382663,
      "tokens_seen": 397803520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441524573721163,
      "loss": 1.4915,
      "theoretical_loss": 4.014276864545036,
      "tokens_seen": 397934592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044413239719157474,
      "loss": 1.5733,
      "theoretical_loss": 4.014128837130021,
      "tokens_seen": 398065664
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.42839688062667847,
      "objective/train/avg_token_score": 0.010799635201692581,
      "objective/train/docs_used": 236468,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.580099105834961,
      "objective/train/original_loss": 3.822235107421875,
      "objective/train/theoretical_loss": 4.0140548468163475,
      "objective/train/tokens_used": 398131200,
      "objective/train/unlikelihood_loss": 0.15448230504989624,
      "theoretical_loss": 4.0140548468163475,
      "tokens_seen": 398131200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441123370110331,
      "loss": 1.613,
      "theoretical_loss": 4.0139808720907615,
      "tokens_seen": 398196736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044409227683049146,
      "loss": 1.4741,
      "theoretical_loss": 4.013832969380456,
      "tokens_seen": 398327808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440722166499498,
      "loss": 1.4448,
      "theoretical_loss": 4.013685128952353,
      "tokens_seen": 398458880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044405215646940824,
      "loss": 1.3947,
      "theoretical_loss": 4.01353735075975,
      "tokens_seen": 398589952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440320962888666,
      "loss": 1.5525,
      "theoretical_loss": 4.013389634755995,
      "tokens_seen": 398721024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044401203610832497,
      "loss": 1.5298,
      "theoretical_loss": 4.013241980894488,
      "tokens_seen": 398852096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044399197592778333,
      "loss": 1.4135,
      "theoretical_loss": 4.013094389128677,
      "tokens_seen": 398983168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439719157472417,
      "loss": 1.5502,
      "theoretical_loss": 4.012946859412062,
      "tokens_seen": 399114240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439518555667001,
      "loss": 1.5869,
      "theoretical_loss": 4.012799391698191,
      "tokens_seen": 399245312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044393179538615847,
      "loss": 1.4827,
      "theoretical_loss": 4.012651985940666,
      "tokens_seen": 399376384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044391173520561683,
      "loss": 1.6088,
      "theoretical_loss": 4.012504642093133,
      "tokens_seen": 399507456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438916750250752,
      "loss": 1.4438,
      "theoretical_loss": 4.012357360109293,
      "tokens_seen": 399638528
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.4299853444099426,
      "objective/train/avg_token_score": 0.011655104346573353,
      "objective/train/docs_used": 237759,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6880359649658203,
      "objective/train/original_loss": 3.903928279876709,
      "objective/train/theoretical_loss": 4.012210139942894,
      "objective/train/tokens_used": 399769600,
      "objective/train/unlikelihood_loss": 0.17748136818408966,
      "theoretical_loss": 4.012210139942894,
      "tokens_seen": 399769600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438716148445336,
      "loss": 1.6091,
      "theoretical_loss": 4.012210139942894,
      "tokens_seen": 399769600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044385155466399197,
      "loss": 1.6209,
      "theoretical_loss": 4.012062981547736,
      "tokens_seen": 399900672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044383149448345033,
      "loss": 1.6198,
      "theoretical_loss": 4.0119158848776655,
      "tokens_seen": 400031744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438114343029087,
      "loss": 1.3378,
      "theoretical_loss": 4.011768849886582,
      "tokens_seen": 400162816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437913741223671,
      "loss": 1.6247,
      "theoretical_loss": 4.011621876528431,
      "tokens_seen": 400293888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437713139418255,
      "loss": 1.7281,
      "theoretical_loss": 4.011474964757211,
      "tokens_seen": 400424960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044375125376128384,
      "loss": 1.477,
      "theoretical_loss": 4.011328114526968,
      "tokens_seen": 400556032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437311935807422,
      "loss": 1.6307,
      "theoretical_loss": 4.011181325791797,
      "tokens_seen": 400687104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044371113340020056,
      "loss": 1.3736,
      "theoretical_loss": 4.011034598505843,
      "tokens_seen": 400818176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044369107321965903,
      "loss": 1.4352,
      "theoretical_loss": 4.010887932623299,
      "tokens_seen": 400949248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436710130391174,
      "loss": 1.6627,
      "theoretical_loss": 4.01074132809841,
      "tokens_seen": 401080320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044365095285857576,
      "loss": 1.6658,
      "theoretical_loss": 4.010594784885466,
      "tokens_seen": 401211392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436308926780341,
      "loss": 1.4219,
      "theoretical_loss": 4.01044830293881,
      "tokens_seen": 401342464
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.31763195991516113,
      "objective/train/avg_token_score": 0.008867236785590649,
      "objective/train/docs_used": 238664,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1373664140701294,
      "objective/train/original_loss": 3.9792494773864746,
      "objective/train/theoretical_loss": 4.0103750849260855,
      "objective/train/tokens_used": 401408000,
      "objective/train/unlikelihood_loss": 0.17361101508140564,
      "theoretical_loss": 4.0103750849260855,
      "tokens_seen": 401408000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044361083249749254,
      "loss": 1.4158,
      "theoretical_loss": 4.010301882212831,
      "tokens_seen": 401473536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435907723169509,
      "loss": 1.6034,
      "theoretical_loss": 4.010155522661968,
      "tokens_seen": 401604608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044357071213640926,
      "loss": 1.6155,
      "theoretical_loss": 4.010009224240708,
      "tokens_seen": 401735680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435506519558676,
      "loss": 1.7323,
      "theoretical_loss": 4.009862986903588,
      "tokens_seen": 401866752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443530591775326,
      "loss": 1.5855,
      "theoretical_loss": 4.009716810605193,
      "tokens_seen": 401997824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435105315947844,
      "loss": 1.481,
      "theoretical_loss": 4.009570695300155,
      "tokens_seen": 402128896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044349047141424276,
      "loss": 1.5404,
      "theoretical_loss": 4.009424640943157,
      "tokens_seen": 402259968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434704112337011,
      "loss": 1.5359,
      "theoretical_loss": 4.00927864748893,
      "tokens_seen": 402391040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434503510531595,
      "loss": 1.5011,
      "theoretical_loss": 4.009132714892251,
      "tokens_seen": 402522112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434302908726179,
      "loss": 1.6215,
      "theoretical_loss": 4.008986843107949,
      "tokens_seen": 402653184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044341023069207627,
      "loss": 1.5534,
      "theoretical_loss": 4.008841032090898,
      "tokens_seen": 402784256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044339017051153463,
      "loss": 1.4099,
      "theoretical_loss": 4.0086952817960215,
      "tokens_seen": 402915328
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.4825879633426666,
      "objective/train/avg_token_score": 0.010442715138196945,
      "objective/train/docs_used": 240054,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6706886291503906,
      "objective/train/original_loss": 3.7450501918792725,
      "objective/train/theoretical_loss": 4.008549592178291,
      "objective/train/tokens_used": 403046400,
      "objective/train/unlikelihood_loss": 0.15135115385055542,
      "theoretical_loss": 4.008549592178291,
      "tokens_seen": 403046400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443370110330993,
      "loss": 1.4657,
      "theoretical_loss": 4.008549592178291,
      "tokens_seen": 403046400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044335005015045135,
      "loss": 1.4444,
      "theoretical_loss": 4.008403963192725,
      "tokens_seen": 403177472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044332998996990977,
      "loss": 1.6658,
      "theoretical_loss": 4.008258394794392,
      "tokens_seen": 403308544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044330992978936813,
      "loss": 1.5749,
      "theoretical_loss": 4.008112886938409,
      "tokens_seen": 403439616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432898696088265,
      "loss": 1.4549,
      "theoretical_loss": 4.007967439579936,
      "tokens_seen": 403570688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044326980942828486,
      "loss": 1.3103,
      "theoretical_loss": 4.007822052674185,
      "tokens_seen": 403701760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044324974924774327,
      "loss": 1.6347,
      "theoretical_loss": 4.007676726176415,
      "tokens_seen": 403832832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044322968906720163,
      "loss": 1.3592,
      "theoretical_loss": 4.007531460041933,
      "tokens_seen": 403963904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044320962888666,
      "loss": 1.4303,
      "theoretical_loss": 4.00738625422609,
      "tokens_seen": 404094976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044318956870611836,
      "loss": 1.2863,
      "theoretical_loss": 4.0072411086842905,
      "tokens_seen": 404226048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431695085255767,
      "loss": 1.4716,
      "theoretical_loss": 4.007096023371981,
      "tokens_seen": 404357120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044314944834503514,
      "loss": 1.4291,
      "theoretical_loss": 4.006950998244658,
      "tokens_seen": 404488192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431293881644935,
      "loss": 1.5443,
      "theoretical_loss": 4.006806033257866,
      "tokens_seen": 404619264
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.4843597114086151,
      "objective/train/avg_token_score": 0.03304527327418327,
      "objective/train/docs_used": 240788,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6086971759796143,
      "objective/train/original_loss": 3.774994134902954,
      "objective/train/theoretical_loss": 4.006733573303289,
      "objective/train/tokens_used": 404684800,
      "objective/train/unlikelihood_loss": 0.15561363101005554,
      "theoretical_loss": 4.006733573303289,
      "tokens_seen": 404684800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044310932798395186,
      "loss": 1.4442,
      "theoretical_loss": 4.0066611283671945,
      "tokens_seen": 404750336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004430892678034102,
      "loss": 1.498,
      "theoretical_loss": 4.00651628352828,
      "tokens_seen": 404881408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044306920762286864,
      "loss": 1.5164,
      "theoretical_loss": 4.00637149869681,
      "tokens_seen": 405012480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443049147442327,
      "loss": 1.468,
      "theoretical_loss": 4.006226773828514,
      "tokens_seen": 405143552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044302908726178537,
      "loss": 1.517,
      "theoretical_loss": 4.0060821088791725,
      "tokens_seen": 405274624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044300902708124373,
      "loss": 1.6745,
      "theoretical_loss": 4.00593750380461,
      "tokens_seen": 405405696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429889669007021,
      "loss": 1.6433,
      "theoretical_loss": 4.005792958560699,
      "tokens_seen": 405536768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429689067201605,
      "loss": 1.5431,
      "theoretical_loss": 4.00564847310336,
      "tokens_seen": 405667840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044294884653961887,
      "loss": 1.4441,
      "theoretical_loss": 4.005504047388558,
      "tokens_seen": 405798912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044292878635907723,
      "loss": 1.4946,
      "theoretical_loss": 4.005359681372306,
      "tokens_seen": 405929984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429087261785356,
      "loss": 1.5999,
      "theoretical_loss": 4.005215375010662,
      "tokens_seen": 406061056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442888665997994,
      "loss": 1.5072,
      "theoretical_loss": 4.005071128259734,
      "tokens_seen": 406192128
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.388868510723114,
      "objective/train/avg_token_score": 0.012570223771035671,
      "objective/train/docs_used": 241796,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4023349285125732,
      "objective/train/original_loss": 3.9673821926116943,
      "objective/train/theoretical_loss": 4.004926941075674,
      "objective/train/tokens_used": 406323200,
      "objective/train/unlikelihood_loss": 0.16696441173553467,
      "theoretical_loss": 4.004926941075674,
      "tokens_seen": 406323200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044286860581745237,
      "loss": 1.536,
      "theoretical_loss": 4.004926941075674,
      "tokens_seen": 406323200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044284854563691073,
      "loss": 1.482,
      "theoretical_loss": 4.004782813414678,
      "tokens_seen": 406454272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428284854563691,
      "loss": 1.5833,
      "theoretical_loss": 4.0046387452329935,
      "tokens_seen": 406585344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428084252758275,
      "loss": 1.5516,
      "theoretical_loss": 4.00449473648691,
      "tokens_seen": 406716416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427883650952859,
      "loss": 1.5163,
      "theoretical_loss": 4.004350787132766,
      "tokens_seen": 406847488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044276830491474424,
      "loss": 1.4299,
      "theoretical_loss": 4.004206897126943,
      "tokens_seen": 406978560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427482447342026,
      "loss": 1.3877,
      "theoretical_loss": 4.004063066425874,
      "tokens_seen": 407109632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044272818455366096,
      "loss": 1.5866,
      "theoretical_loss": 4.003919294986032,
      "tokens_seen": 407240704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427081243731194,
      "loss": 1.3976,
      "theoretical_loss": 4.003775582763939,
      "tokens_seen": 407371776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044268806419257774,
      "loss": 1.4836,
      "theoretical_loss": 4.003631929716162,
      "tokens_seen": 407502848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426680040120361,
      "loss": 1.6081,
      "theoretical_loss": 4.0034883357993145,
      "tokens_seen": 407633920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044264794383149447,
      "loss": 1.5082,
      "theoretical_loss": 4.003344800970056,
      "tokens_seen": 407764992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426278836509529,
      "loss": 1.6963,
      "theoretical_loss": 4.00320132518509,
      "tokens_seen": 407896064
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.3863636255264282,
      "objective/train/avg_token_score": 0.017943046987056732,
      "objective/train/docs_used": 242638,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4344366788864136,
      "objective/train/original_loss": 4.310576915740967,
      "objective/train/theoretical_loss": 4.0031296094207,
      "objective/train/tokens_used": 407961600,
      "objective/train/unlikelihood_loss": 0.1840338259935379,
      "theoretical_loss": 4.0031296094207,
      "tokens_seen": 407961600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044260782347041124,
      "loss": 1.5529,
      "theoretical_loss": 4.003057908401168,
      "tokens_seen": 408027136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425877632898696,
      "loss": 1.5715,
      "theoretical_loss": 4.002914550575085,
      "tokens_seen": 408158208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044256770310932797,
      "loss": 1.5628,
      "theoretical_loss": 4.002771251663681,
      "tokens_seen": 408289280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044254764292878633,
      "loss": 1.4782,
      "theoretical_loss": 4.0026280116238455,
      "tokens_seen": 408420352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044252758274824475,
      "loss": 1.6979,
      "theoretical_loss": 4.0024848304125085,
      "tokens_seen": 408551424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425075225677031,
      "loss": 1.4791,
      "theoretical_loss": 4.002341707986648,
      "tokens_seen": 408682496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044248746238716147,
      "loss": 1.6751,
      "theoretical_loss": 4.002198644303288,
      "tokens_seen": 408813568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044246740220661983,
      "loss": 1.5135,
      "theoretical_loss": 4.002055639319494,
      "tokens_seen": 408944640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044244734202607825,
      "loss": 1.6368,
      "theoretical_loss": 4.001912692992379,
      "tokens_seen": 409075712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424272818455366,
      "loss": 1.5111,
      "theoretical_loss": 4.001769805279103,
      "tokens_seen": 409206784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442407221664995,
      "loss": 1.4949,
      "theoretical_loss": 4.001626976136868,
      "tokens_seen": 409337856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044238716148445334,
      "loss": 1.5723,
      "theoretical_loss": 4.001484205522922,
      "tokens_seen": 409468928
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.42625856399536133,
      "objective/train/avg_token_score": 0.018668072298169136,
      "objective/train/docs_used": 244086,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4982153177261353,
      "objective/train/original_loss": 3.7382147312164307,
      "objective/train/theoretical_loss": 4.001341493394558,
      "objective/train/tokens_used": 409600000,
      "objective/train/unlikelihood_loss": 0.17688016593456268,
      "theoretical_loss": 4.001341493394558,
      "tokens_seen": 409600000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423671013039117,
      "loss": 1.6406,
      "theoretical_loss": 4.001341493394558,
      "tokens_seen": 409600000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423470411233701,
      "loss": 1.59,
      "theoretical_loss": 4.001198839709113,
      "tokens_seen": 409731072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423269809428285,
      "loss": 1.439,
      "theoretical_loss": 4.001056244423971,
      "tokens_seen": 409862144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044230692076228684,
      "loss": 1.5345,
      "theoretical_loss": 4.000913707496556,
      "tokens_seen": 409993216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422868605817452,
      "loss": 1.5954,
      "theoretical_loss": 4.000771228884344,
      "tokens_seen": 410124288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422668004012036,
      "loss": 1.6115,
      "theoretical_loss": 4.000628808544848,
      "tokens_seen": 410255360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442246740220662,
      "loss": 1.4791,
      "theoretical_loss": 4.000486446435629,
      "tokens_seen": 410386432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044222668004012034,
      "loss": 1.4222,
      "theoretical_loss": 4.000344142514294,
      "tokens_seen": 410517504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422066198595787,
      "loss": 1.3933,
      "theoretical_loss": 4.000201896738492,
      "tokens_seen": 410648576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421865596790371,
      "loss": 1.6559,
      "theoretical_loss": 4.0000597090659165,
      "tokens_seen": 410779648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044216649949849554,
      "loss": 1.5433,
      "theoretical_loss": 3.9999175794543063,
      "tokens_seen": 410910720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421464393179539,
      "loss": 1.4679,
      "theoretical_loss": 3.999775507861443,
      "tokens_seen": 411041792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044212637913741226,
      "loss": 1.3713,
      "theoretical_loss": 3.9996334942451544,
      "tokens_seen": 411172864
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.39326736330986023,
      "objective/train/avg_token_score": 0.010529226623475552,
      "objective/train/docs_used": 244630,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.416571021080017,
      "objective/train/original_loss": 3.8771302700042725,
      "objective/train/theoretical_loss": 3.999562509165058,
      "objective/train/tokens_used": 411238400,
      "objective/train/unlikelihood_loss": 0.1982622891664505,
      "theoretical_loss": 3.999562509165058,
      "tokens_seen": 411238400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421063189568706,
      "loss": 1.6983,
      "theoretical_loss": 3.9994915385633103,
      "tokens_seen": 411303936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044208625877632904,
      "loss": 1.5469,
      "theoretical_loss": 3.9993496407738256,
      "tokens_seen": 411435008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420661985957874,
      "loss": 1.3864,
      "theoretical_loss": 3.999207800834659,
      "tokens_seen": 411566080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044204613841524577,
      "loss": 1.5093,
      "theoretical_loss": 3.999066018703814,
      "tokens_seen": 411697152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044202607823470413,
      "loss": 1.617,
      "theoretical_loss": 3.9989242943393357,
      "tokens_seen": 411828224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420060180541625,
      "loss": 1.5732,
      "theoretical_loss": 3.998782627699315,
      "tokens_seen": 411959296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004419859578736209,
      "loss": 1.657,
      "theoretical_loss": 3.9986410187418855,
      "tokens_seen": 412090368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044196589769307927,
      "loss": 1.5478,
      "theoretical_loss": 3.9984994674252254,
      "tokens_seen": 412221440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044194583751253763,
      "loss": 1.4761,
      "theoretical_loss": 3.9983579737075554,
      "tokens_seen": 412352512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000441925777331996,
      "loss": 1.4249,
      "theoretical_loss": 3.9982165375471403,
      "tokens_seen": 412483584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004419057171514544,
      "loss": 1.5888,
      "theoretical_loss": 3.9980751589022887,
      "tokens_seen": 412614656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044188565697091277,
      "loss": 1.4857,
      "theoretical_loss": 3.9979338377313525,
      "tokens_seen": 412745728
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4978005886077881,
      "objective/train/avg_token_score": 0.016746150329709053,
      "objective/train/docs_used": 245807,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8077481985092163,
      "objective/train/original_loss": 4.005298614501953,
      "objective/train/theoretical_loss": 3.997792573992726,
      "objective/train/tokens_used": 412876800,
      "objective/train/unlikelihood_loss": 0.1058061420917511,
      "theoretical_loss": 3.997792573992726,
      "tokens_seen": 412876800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044186559679037113,
      "loss": 1.5612,
      "theoretical_loss": 3.997792573992726,
      "tokens_seen": 412876800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418455366098295,
      "loss": 1.3567,
      "theoretical_loss": 3.997651367644848,
      "tokens_seen": 413007872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418254764292879,
      "loss": 1.674,
      "theoretical_loss": 3.9975102186462004,
      "tokens_seen": 413138944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418054162487463,
      "loss": 1.4358,
      "theoretical_loss": 3.9973691269553076,
      "tokens_seen": 413270016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044178535606820464,
      "loss": 1.4689,
      "theoretical_loss": 3.9972280925307384,
      "tokens_seen": 413401088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441765295887663,
      "loss": 1.6083,
      "theoretical_loss": 3.9970871153311025,
      "tokens_seen": 413532160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044174523570712136,
      "loss": 1.434,
      "theoretical_loss": 3.996946195315055,
      "tokens_seen": 413663232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417251755265798,
      "loss": 1.6717,
      "theoretical_loss": 3.996805332441293,
      "tokens_seen": 413794304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044170511534603814,
      "loss": 1.5077,
      "theoretical_loss": 3.996664526668555,
      "tokens_seen": 413925376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416850551654965,
      "loss": 1.5328,
      "theoretical_loss": 3.996523777955626,
      "tokens_seen": 414056448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044166499498495487,
      "loss": 1.4311,
      "theoretical_loss": 3.9963830862613303,
      "tokens_seen": 414187520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416449348044133,
      "loss": 1.5412,
      "theoretical_loss": 3.9962424515445356,
      "tokens_seen": 414318592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044162487462387164,
      "loss": 1.4443,
      "theoretical_loss": 3.9961018737641547,
      "tokens_seen": 414449664
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4602883756160736,
      "objective/train/avg_token_score": 0.015043980441987514,
      "objective/train/docs_used": 246540,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.704790472984314,
      "objective/train/original_loss": 4.011831283569336,
      "objective/train/theoretical_loss": 3.99603160621229,
      "objective/train/tokens_used": 414515200,
      "objective/train/unlikelihood_loss": 0.13677598536014557,
      "theoretical_loss": 3.99603160621229,
      "tokens_seen": 414515200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044160481444333,
      "loss": 1.4581,
      "theoretical_loss": 3.9959613528791396,
      "tokens_seen": 414580736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044158475426278837,
      "loss": 1.4903,
      "theoretical_loss": 3.9958208888484865,
      "tokens_seen": 414711808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044156469408224673,
      "loss": 1.5552,
      "theoretical_loss": 3.9956804816312355,
      "tokens_seen": 414842880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044154463390170515,
      "loss": 1.5021,
      "theoretical_loss": 3.9955401311864662,
      "tokens_seen": 414973952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415245737211635,
      "loss": 1.5002,
      "theoretical_loss": 3.995399837473302,
      "tokens_seen": 415105024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044150451354062187,
      "loss": 1.5115,
      "theoretical_loss": 3.99525960045091,
      "tokens_seen": 415236096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044148445336008023,
      "loss": 1.5434,
      "theoretical_loss": 3.9951194200784976,
      "tokens_seen": 415367168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044146439317953865,
      "loss": 1.5491,
      "theoretical_loss": 3.9949792963153143,
      "tokens_seen": 415498240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441444332998997,
      "loss": 1.5161,
      "theoretical_loss": 3.994839229120654,
      "tokens_seen": 415629312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414242728184554,
      "loss": 1.5123,
      "theoretical_loss": 3.99469921845385,
      "tokens_seen": 415760384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044140421263791374,
      "loss": 1.4513,
      "theoretical_loss": 3.9945592642742795,
      "tokens_seen": 415891456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413841524573721,
      "loss": 1.5792,
      "theoretical_loss": 3.9944193665413605,
      "tokens_seen": 416022528
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.38825756311416626,
      "objective/train/avg_token_score": 0.025925040245056152,
      "objective/train/docs_used": 247824,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4125205278396606,
      "objective/train/original_loss": 3.7366883754730225,
      "objective/train/theoretical_loss": 3.994279525214554,
      "objective/train/tokens_used": 416153600,
      "objective/train/unlikelihood_loss": 0.2123020440340042,
      "theoretical_loss": 3.994279525214554,
      "tokens_seen": 416153600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413640922768305,
      "loss": 1.4642,
      "theoretical_loss": 3.994279525214554,
      "tokens_seen": 416153600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413440320962889,
      "loss": 1.6119,
      "theoretical_loss": 3.994139740253363,
      "tokens_seen": 416284672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044132397191574724,
      "loss": 1.4493,
      "theoretical_loss": 3.99400001161733,
      "tokens_seen": 416415744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413039117352056,
      "loss": 1.4857,
      "theoretical_loss": 3.9938603392660426,
      "tokens_seen": 416546816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441283851554664,
      "loss": 1.5087,
      "theoretical_loss": 3.993720723159127,
      "tokens_seen": 416677888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412637913741224,
      "loss": 1.5877,
      "theoretical_loss": 3.9935811632562532,
      "tokens_seen": 416808960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044124373119358074,
      "loss": 1.5386,
      "theoretical_loss": 3.9934416595171323,
      "tokens_seen": 416940032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412236710130391,
      "loss": 1.5819,
      "theoretical_loss": 3.993302211901516,
      "tokens_seen": 417071104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044120361083249747,
      "loss": 1.5589,
      "theoretical_loss": 3.993162820369199,
      "tokens_seen": 417202176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411835506519559,
      "loss": 1.5473,
      "theoretical_loss": 3.993023484880016,
      "tokens_seen": 417333248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044116349047141425,
      "loss": 1.5015,
      "theoretical_loss": 3.9928842053938443,
      "tokens_seen": 417464320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411434302908726,
      "loss": 1.3862,
      "theoretical_loss": 3.9927449818706005,
      "tokens_seen": 417595392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044112337011033097,
      "loss": 1.5006,
      "theoretical_loss": 3.9926058142702456,
      "tokens_seen": 417726464
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4471529722213745,
      "objective/train/avg_token_score": 0.010523061268031597,
      "objective/train/docs_used": 248559,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6317007541656494,
      "objective/train/original_loss": 3.7148232460021973,
      "objective/train/theoretical_loss": 3.9925362514286498,
      "objective/train/tokens_used": 417792000,
      "objective/train/unlikelihood_loss": 0.19979330897331238,
      "theoretical_loss": 3.9925362514286498,
      "tokens_seen": 417792000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411033099297894,
      "loss": 1.4831,
      "theoretical_loss": 3.992466702552779,
      "tokens_seen": 417857536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044108324974924775,
      "loss": 1.5975,
      "theoretical_loss": 3.9923276466782425,
      "tokens_seen": 417988608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410631895687061,
      "loss": 1.6912,
      "theoretical_loss": 3.992188646606719,
      "tokens_seen": 418119680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410431293881645,
      "loss": 1.4185,
      "theoretical_loss": 3.992049702298332,
      "tokens_seen": 418250752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044102306920762284,
      "loss": 1.4789,
      "theoretical_loss": 3.991910813713246,
      "tokens_seen": 418381824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044100300902708125,
      "loss": 1.4325,
      "theoretical_loss": 3.991771980811667,
      "tokens_seen": 418512896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004409829488465396,
      "loss": 1.6477,
      "theoretical_loss": 3.9916332035538415,
      "tokens_seen": 418643968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440962888665998,
      "loss": 1.6004,
      "theoretical_loss": 3.991494481900056,
      "tokens_seen": 418775040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044094282848545634,
      "loss": 1.4793,
      "theoretical_loss": 3.99135581581064,
      "tokens_seen": 418906112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044092276830491476,
      "loss": 1.5306,
      "theoretical_loss": 3.9912172052459614,
      "tokens_seen": 419037184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004409027081243731,
      "loss": 1.6564,
      "theoretical_loss": 3.991078650166429,
      "tokens_seen": 419168256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408826479438315,
      "loss": 1.5377,
      "theoretical_loss": 3.990940150532494,
      "tokens_seen": 419299328
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.3536168038845062,
      "objective/train/avg_token_score": 0.01034875400364399,
      "objective/train/docs_used": 249785,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2462135553359985,
      "objective/train/original_loss": 4.00803804397583,
      "objective/train/theoretical_loss": 3.990801706304647,
      "objective/train/tokens_used": 419430400,
      "objective/train/unlikelihood_loss": 0.17689494788646698,
      "theoretical_loss": 3.990801706304647,
      "tokens_seen": 419430400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044086258776328984,
      "loss": 1.5763,
      "theoretical_loss": 3.990801706304647,
      "tokens_seen": 419430400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408425275827482,
      "loss": 1.4898,
      "theoretical_loss": 3.990663317443418,
      "tokens_seen": 419561472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408224674022066,
      "loss": 1.5875,
      "theoretical_loss": 3.9905249839093786,
      "tokens_seen": 419692544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440802407221665,
      "loss": 1.5456,
      "theoretical_loss": 3.990386705663142,
      "tokens_seen": 419823616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044078234704112335,
      "loss": 1.6324,
      "theoretical_loss": 3.990248482665359,
      "tokens_seen": 419954688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407622868605817,
      "loss": 1.6974,
      "theoretical_loss": 3.9901103148767225,
      "tokens_seen": 420085760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407422266800401,
      "loss": 1.5273,
      "theoretical_loss": 3.989972202257966,
      "tokens_seen": 420216832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407221664994985,
      "loss": 1.321,
      "theoretical_loss": 3.989834144769861,
      "tokens_seen": 420347904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044070210631895685,
      "loss": 1.5629,
      "theoretical_loss": 3.989696142373221,
      "tokens_seen": 420478976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044068204613841527,
      "loss": 1.6409,
      "theoretical_loss": 3.9895581950288994,
      "tokens_seen": 420610048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406619859578737,
      "loss": 1.4717,
      "theoretical_loss": 3.989420302697789,
      "tokens_seen": 420741120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044064192577733204,
      "loss": 1.5218,
      "theoretical_loss": 3.989282465340823,
      "tokens_seen": 420872192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406218655967904,
      "loss": 1.4617,
      "theoretical_loss": 3.9891446829189734,
      "tokens_seen": 421003264
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4630987346172333,
      "objective/train/avg_token_score": 0.02803061157464981,
      "objective/train/docs_used": 250392,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7076679468154907,
      "objective/train/original_loss": 4.153679370880127,
      "objective/train/theoretical_loss": 3.989075812296533,
      "objective/train/tokens_used": 421068800,
      "objective/train/unlikelihood_loss": 0.13347437977790833,
      "theoretical_loss": 3.989075812296533,
      "tokens_seen": 421068800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044060180541624877,
      "loss": 1.6244,
      "theoretical_loss": 3.989006955393254,
      "tokens_seen": 421134336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044058174523570713,
      "loss": 1.7096,
      "theoretical_loss": 3.9888692827247167,
      "tokens_seen": 421265408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044056168505516555,
      "loss": 1.4498,
      "theoretical_loss": 3.988731664874454,
      "tokens_seen": 421396480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405416248746239,
      "loss": 1.3687,
      "theoretical_loss": 3.988594101803598,
      "tokens_seen": 421527552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044052156469408227,
      "loss": 1.687,
      "theoretical_loss": 3.9884565934733196,
      "tokens_seen": 421658624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044050150451354063,
      "loss": 1.444,
      "theoretical_loss": 3.9883191398448306,
      "tokens_seen": 421789696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044048144433299905,
      "loss": 1.3943,
      "theoretical_loss": 3.988181740879382,
      "tokens_seen": 421920768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404613841524574,
      "loss": 1.4322,
      "theoretical_loss": 3.988044396538263,
      "tokens_seen": 422051840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404413239719158,
      "loss": 1.5675,
      "theoretical_loss": 3.9879071067828034,
      "tokens_seen": 422182912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044042126379137414,
      "loss": 1.5548,
      "theoretical_loss": 3.987769871574373,
      "tokens_seen": 422313984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404012036108325,
      "loss": 1.4586,
      "theoretical_loss": 3.987632690874379,
      "tokens_seen": 422445056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403811434302909,
      "loss": 1.6742,
      "theoretical_loss": 3.98749556464427,
      "tokens_seen": 422576128
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.3960777223110199,
      "objective/train/avg_token_score": 0.019322294741868973,
      "objective/train/docs_used": 251199,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4102435111999512,
      "objective/train/original_loss": 4.012388706207275,
      "objective/train/theoretical_loss": 3.987358492845532,
      "objective/train/tokens_used": 422707200,
      "objective/train/unlikelihood_loss": 0.17995640635490417,
      "theoretical_loss": 3.987358492845532,
      "tokens_seen": 422707200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403610832497493,
      "loss": 1.4228,
      "theoretical_loss": 3.987358492845532,
      "tokens_seen": 422707200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044034102306920764,
      "loss": 1.4906,
      "theoretical_loss": 3.987221475439691,
      "tokens_seen": 422838272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440320962888666,
      "loss": 1.3949,
      "theoretical_loss": 3.9870845123883116,
      "tokens_seen": 422969344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403009027081244,
      "loss": 1.3426,
      "theoretical_loss": 3.9869476036529994,
      "tokens_seen": 423100416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402808425275828,
      "loss": 1.562,
      "theoretical_loss": 3.986810749195396,
      "tokens_seen": 423231488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044026078234704114,
      "loss": 1.8139,
      "theoretical_loss": 3.9866739489771845,
      "tokens_seen": 423362560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402407221664995,
      "loss": 1.6235,
      "theoretical_loss": 3.986537202960085,
      "tokens_seen": 423493632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044022066198595787,
      "loss": 1.6967,
      "theoretical_loss": 3.986400511105858,
      "tokens_seen": 423624704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402006018054163,
      "loss": 1.3629,
      "theoretical_loss": 3.986263873376301,
      "tokens_seen": 423755776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044018054162487465,
      "loss": 1.6447,
      "theoretical_loss": 3.9861272897332527,
      "tokens_seen": 423886848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440160481444333,
      "loss": 1.5283,
      "theoretical_loss": 3.9859907601385887,
      "tokens_seen": 424017920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044014042126379137,
      "loss": 1.4187,
      "theoretical_loss": 3.985854284554223,
      "tokens_seen": 424148992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401203610832498,
      "loss": 1.6492,
      "theoretical_loss": 3.98571786294211,
      "tokens_seen": 424280064
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4866202473640442,
      "objective/train/avg_token_score": 0.008693613111972809,
      "objective/train/docs_used": 252247,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6004599332809448,
      "objective/train/original_loss": 3.3173439502716064,
      "objective/train/theoretical_loss": 3.9856496723637687,
      "objective/train/tokens_used": 424345600,
      "objective/train/unlikelihood_loss": 0.19723816215991974,
      "theoretical_loss": 3.9856496723637687,
      "tokens_seen": 424345600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044010030090270815,
      "loss": 1.6577,
      "theoretical_loss": 3.985581495264241,
      "tokens_seen": 424411136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400802407221665,
      "loss": 1.5225,
      "theoretical_loss": 3.9854451814826457,
      "tokens_seen": 424542208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400601805416249,
      "loss": 1.4646,
      "theoretical_loss": 3.985308921559394,
      "tokens_seen": 424673280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044004012036108324,
      "loss": 1.4294,
      "theoretical_loss": 3.9851727154565926,
      "tokens_seen": 424804352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044002006018054165,
      "loss": 1.6139,
      "theoretical_loss": 3.9850365631363873,
      "tokens_seen": 424935424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044,
      "loss": 1.5823,
      "theoretical_loss": 3.984900464560961,
      "tokens_seen": 425066496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399799398194584,
      "loss": 1.6172,
      "theoretical_loss": 3.984764419692537,
      "tokens_seen": 425197568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043995987963891674,
      "loss": 1.6824,
      "theoretical_loss": 3.984628428493374,
      "tokens_seen": 425328640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043993981945837516,
      "loss": 1.5701,
      "theoretical_loss": 3.984492490925772,
      "tokens_seen": 425459712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399197592778335,
      "loss": 1.5669,
      "theoretical_loss": 3.9843566069520664,
      "tokens_seen": 425590784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398996990972919,
      "loss": 1.3379,
      "theoretical_loss": 3.984220776534632,
      "tokens_seen": 425721856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043987963891675024,
      "loss": 1.6932,
      "theoretical_loss": 3.9840849996358814,
      "tokens_seen": 425852928
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.35037878155708313,
      "objective/train/avg_token_score": 0.010006745345890522,
      "objective/train/docs_used": 253004,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2638418674468994,
      "objective/train/original_loss": 3.8855535984039307,
      "objective/train/theoretical_loss": 3.9839492762182647,
      "objective/train/tokens_used": 425984000,
      "objective/train/unlikelihood_loss": 0.18474146723747253,
      "theoretical_loss": 3.9839492762182647,
      "tokens_seen": 425984000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398595787362086,
      "loss": 1.5895,
      "theoretical_loss": 3.9839492762182647,
      "tokens_seen": 425984000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439839518555667,
      "loss": 1.5817,
      "theoretical_loss": 3.9838136062442704,
      "tokens_seen": 426115072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398194583751254,
      "loss": 1.6915,
      "theoretical_loss": 3.9836779896764254,
      "tokens_seen": 426246144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043979939819458375,
      "loss": 1.7129,
      "theoretical_loss": 3.983542426477292,
      "tokens_seen": 426377216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397793380140421,
      "loss": 1.6312,
      "theoretical_loss": 3.983406916609473,
      "tokens_seen": 426508288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397592778335005,
      "loss": 1.5439,
      "theoretical_loss": 3.983271460035608,
      "tokens_seen": 426639360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397392176529589,
      "loss": 1.5743,
      "theoretical_loss": 3.983136056718373,
      "tokens_seen": 426770432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043971915747241725,
      "loss": 1.5065,
      "theoretical_loss": 3.9830007066204836,
      "tokens_seen": 426901504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396990972918756,
      "loss": 1.5345,
      "theoretical_loss": 3.982865409704692,
      "tokens_seen": 427032576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043967903711133403,
      "loss": 1.6371,
      "theoretical_loss": 3.9827301659337864,
      "tokens_seen": 427163648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396589769307924,
      "loss": 1.626,
      "theoretical_loss": 3.982594975270595,
      "tokens_seen": 427294720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043963891675025075,
      "loss": 1.5697,
      "theoretical_loss": 3.9824598376779825,
      "tokens_seen": 427425792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396188565697091,
      "loss": 1.4803,
      "theoretical_loss": 3.98232475311885,
      "tokens_seen": 427556864
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.33889296650886536,
      "objective/train/avg_token_score": 0.04264990985393524,
      "objective/train/docs_used": 254260,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.204171061515808,
      "objective/train/original_loss": 4.482820987701416,
      "objective/train/theoretical_loss": 3.9822572307152564,
      "objective/train/tokens_used": 427622400,
      "objective/train/unlikelihood_loss": 0.09470682591199875,
      "theoretical_loss": 3.9822572307152564,
      "tokens_seen": 427622400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395987963891675,
      "loss": 1.5879,
      "theoretical_loss": 3.9821897215561375,
      "tokens_seen": 427687936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395787362086259,
      "loss": 1.4706,
      "theoretical_loss": 3.98205474295282,
      "tokens_seen": 427819008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043955867602808426,
      "loss": 1.558,
      "theoretical_loss": 3.9819198172719124,
      "tokens_seen": 427950080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395386158475426,
      "loss": 1.6488,
      "theoretical_loss": 3.9817849444764644,
      "tokens_seen": 428081152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439518555667001,
      "loss": 1.641,
      "theoretical_loss": 3.9816501245295646,
      "tokens_seen": 428212224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394984954864594,
      "loss": 1.6366,
      "theoretical_loss": 3.981515357394338,
      "tokens_seen": 428343296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043947843530591776,
      "loss": 1.3471,
      "theoretical_loss": 3.981380643033945,
      "tokens_seen": 428474368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394583751253761,
      "loss": 1.5653,
      "theoretical_loss": 3.981245981411586,
      "tokens_seen": 428605440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394383149448345,
      "loss": 1.573,
      "theoretical_loss": 3.981111372490496,
      "tokens_seen": 428736512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043941825476429285,
      "loss": 1.6642,
      "theoretical_loss": 3.9809768162339476,
      "tokens_seen": 428867584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043939819458375126,
      "loss": 1.5006,
      "theoretical_loss": 3.9808423126052506,
      "tokens_seen": 428998656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393781344032096,
      "loss": 1.4711,
      "theoretical_loss": 3.9807078615677502,
      "tokens_seen": 429129728
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4456256031990051,
      "objective/train/avg_token_score": 0.0028587188571691513,
      "objective/train/docs_used": 255066,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7612452507019043,
      "objective/train/original_loss": 4.201564311981201,
      "objective/train/theoretical_loss": 3.9805734630848306,
      "objective/train/tokens_used": 429260800,
      "objective/train/unlikelihood_loss": 0.11549005657434464,
      "theoretical_loss": 3.9805734630848306,
      "tokens_seen": 429260800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439358074222668,
      "loss": 1.6208,
      "theoretical_loss": 3.9805734630848306,
      "tokens_seen": 429260800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043933801404212635,
      "loss": 1.3378,
      "theoretical_loss": 3.9804391171199103,
      "tokens_seen": 429391872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043931795386158477,
      "loss": 1.4226,
      "theoretical_loss": 3.9803048236364456,
      "tokens_seen": 429522944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043929789368104313,
      "loss": 1.437,
      "theoretical_loss": 3.98017058259793,
      "tokens_seen": 429654016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392778335005015,
      "loss": 1.6993,
      "theoretical_loss": 3.9800363939678918,
      "tokens_seen": 429785088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043925777331995985,
      "loss": 1.454,
      "theoretical_loss": 3.979902257709897,
      "tokens_seen": 429916160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392377131394182,
      "loss": 1.5987,
      "theoretical_loss": 3.979768173787548,
      "tokens_seen": 430047232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043921765295887663,
      "loss": 1.7553,
      "theoretical_loss": 3.9796341421644827,
      "tokens_seen": 430178304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439197592778335,
      "loss": 1.5489,
      "theoretical_loss": 3.979500162804377,
      "tokens_seen": 430309376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391775325977934,
      "loss": 1.4274,
      "theoretical_loss": 3.9793662356709407,
      "tokens_seen": 430440448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043915747241725177,
      "loss": 1.6129,
      "theoretical_loss": 3.9792323607279223,
      "tokens_seen": 430571520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391374122367102,
      "loss": 1.4735,
      "theoretical_loss": 3.9790985379391053,
      "tokens_seen": 430702592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043911735205616855,
      "loss": 1.4495,
      "theoretical_loss": 3.978964767268309,
      "tokens_seen": 430833664
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.35037878155708313,
      "objective/train/avg_token_score": 0.03396159037947655,
      "objective/train/docs_used": 256338,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.339438557624817,
      "objective/train/original_loss": 4.461732387542725,
      "objective/train/theoretical_loss": 3.9788979014658716,
      "objective/train/tokens_used": 430899200,
      "objective/train/unlikelihood_loss": 0.13601770997047424,
      "theoretical_loss": 3.9788979014658716,
      "tokens_seen": 430899200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390972918756269,
      "loss": 1.6812,
      "theoretical_loss": 3.978831048679389,
      "tokens_seen": 430964736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390772316950853,
      "loss": 1.5037,
      "theoretical_loss": 3.9786973821362377,
      "tokens_seen": 431095808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043905717151454364,
      "loss": 1.3236,
      "theoretical_loss": 3.978563767602782,
      "tokens_seen": 431226880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043903711133400205,
      "loss": 1.7428,
      "theoretical_loss": 3.9784302050429874,
      "tokens_seen": 431357952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390170511534604,
      "loss": 1.2738,
      "theoretical_loss": 3.9782966944208518,
      "tokens_seen": 431489024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389969909729188,
      "loss": 1.6967,
      "theoretical_loss": 3.9781632357004124,
      "tokens_seen": 431620096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043897693079237714,
      "loss": 1.6041,
      "theoretical_loss": 3.9780298288457394,
      "tokens_seen": 431751168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043895687061183556,
      "loss": 1.4963,
      "theoretical_loss": 3.977896473820941,
      "tokens_seen": 431882240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389368104312939,
      "loss": 1.4479,
      "theoretical_loss": 3.977763170590159,
      "tokens_seen": 432013312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389167502507523,
      "loss": 1.5352,
      "theoretical_loss": 3.9776299191175726,
      "tokens_seen": 432144384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043889669007021064,
      "loss": 1.6516,
      "theoretical_loss": 3.977496719367396,
      "tokens_seen": 432275456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438876629889669,
      "loss": 1.6093,
      "theoretical_loss": 3.9773635713038784,
      "tokens_seen": 432406528
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.46212121844291687,
      "objective/train/avg_token_score": 0.020001525059342384,
      "objective/train/docs_used": 257001,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5973520278930664,
      "objective/train/original_loss": 3.78364634513855,
      "objective/train/theoretical_loss": 3.9772304748913054,
      "objective/train/tokens_used": 432537600,
      "objective/train/unlikelihood_loss": 0.15700258314609528,
      "theoretical_loss": 3.9772304748913054,
      "tokens_seen": 432537600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388565697091274,
      "loss": 1.5176,
      "theoretical_loss": 3.9772304748913054,
      "tokens_seen": 432537600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388365095285858,
      "loss": 1.5094,
      "theoretical_loss": 3.977097430093999,
      "tokens_seen": 432668672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043881644934804415,
      "loss": 1.5875,
      "theoretical_loss": 3.9769644368763135,
      "tokens_seen": 432799744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387963891675025,
      "loss": 1.5692,
      "theoretical_loss": 3.9768314952026422,
      "tokens_seen": 432930816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387763289869609,
      "loss": 1.4838,
      "theoretical_loss": 3.976698605037411,
      "tokens_seen": 433061888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387562688064193,
      "loss": 1.4281,
      "theoretical_loss": 3.976565766345083,
      "tokens_seen": 433192960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043873620862587765,
      "loss": 1.4939,
      "theoretical_loss": 3.976432979090155,
      "tokens_seen": 433324032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438716148445336,
      "loss": 1.5756,
      "theoretical_loss": 3.9763002432371604,
      "tokens_seen": 433455104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043869608826479443,
      "loss": 1.6247,
      "theoretical_loss": 3.9761675587506673,
      "tokens_seen": 433586176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386760280842528,
      "loss": 1.4458,
      "theoretical_loss": 3.976034925595278,
      "tokens_seen": 433717248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043865596790371115,
      "loss": 1.7759,
      "theoretical_loss": 3.9759023437356316,
      "tokens_seen": 433848320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386359077231695,
      "loss": 1.5281,
      "theoretical_loss": 3.9757698131364005,
      "tokens_seen": 433979392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386158475426279,
      "loss": 1.6324,
      "theoretical_loss": 3.9756373337622932,
      "tokens_seen": 434110464
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4903470277786255,
      "objective/train/avg_token_score": 0.012690459378063679,
      "objective/train/docs_used": 257982,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6558393239974976,
      "objective/train/original_loss": 3.627277374267578,
      "objective/train/theoretical_loss": 3.975571113273642,
      "objective/train/tokens_used": 434176000,
      "objective/train/unlikelihood_loss": 0.18798591196537018,
      "theoretical_loss": 3.975571113273642,
      "tokens_seen": 434176000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385957873620863,
      "loss": 1.5611,
      "theoretical_loss": 3.975504905578053,
      "tokens_seen": 434241536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043857572718154466,
      "loss": 1.6058,
      "theoretical_loss": 3.9753725285484576,
      "tokens_seen": 434372608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438555667001003,
      "loss": 1.6332,
      "theoretical_loss": 3.9752402026383207,
      "tokens_seen": 434503680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385356068204614,
      "loss": 1.6041,
      "theoretical_loss": 3.9751079278124886,
      "tokens_seen": 434634752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385155466399198,
      "loss": 1.4406,
      "theoretical_loss": 3.9749757040358453,
      "tokens_seen": 434765824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043849548645937816,
      "loss": 1.5341,
      "theoretical_loss": 3.974843531273307,
      "tokens_seen": 434896896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384754262788365,
      "loss": 1.5085,
      "theoretical_loss": 3.974711409489826,
      "tokens_seen": 435027968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384553660982949,
      "loss": 1.3682,
      "theoretical_loss": 3.9745793386503885,
      "tokens_seen": 435159040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043843530591775325,
      "loss": 1.5537,
      "theoretical_loss": 3.9744473187200158,
      "tokens_seen": 435290112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043841524573721166,
      "loss": 1.5423,
      "theoretical_loss": 3.974315349663763,
      "tokens_seen": 435421184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043839518555667,
      "loss": 1.5619,
      "theoretical_loss": 3.974183431446721,
      "tokens_seen": 435552256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383751253761284,
      "loss": 1.5113,
      "theoretical_loss": 3.9740515640340135,
      "tokens_seen": 435683328
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.3175097703933716,
      "objective/train/avg_token_score": 0.01103399321436882,
      "objective/train/docs_used": 258654,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1034122705459595,
      "objective/train/original_loss": 3.728519916534424,
      "objective/train/theoretical_loss": 3.973919747390801,
      "objective/train/tokens_used": 435814400,
      "objective/train/unlikelihood_loss": 0.20263297855854034,
      "theoretical_loss": 3.973919747390801,
      "tokens_seen": 435814400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043835506519558675,
      "loss": 1.4324,
      "theoretical_loss": 3.973919747390801,
      "tokens_seen": 435814400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043833500501504517,
      "loss": 1.7232,
      "theoretical_loss": 3.9737879814822747,
      "tokens_seen": 435945472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043831494483450353,
      "loss": 1.5674,
      "theoretical_loss": 3.9736562662736636,
      "tokens_seen": 436076544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382948846539619,
      "loss": 1.6733,
      "theoretical_loss": 3.9735246017302295,
      "tokens_seen": 436207616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043827482447342025,
      "loss": 1.6952,
      "theoretical_loss": 3.9733929878172685,
      "tokens_seen": 436338688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382547642928786,
      "loss": 1.6721,
      "theoretical_loss": 3.9732614245001106,
      "tokens_seen": 436469760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043823470411233703,
      "loss": 1.4846,
      "theoretical_loss": 3.9731299117441203,
      "tokens_seen": 436600832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382146439317954,
      "loss": 1.4607,
      "theoretical_loss": 3.972998449514697,
      "tokens_seen": 436731904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043819458375125376,
      "loss": 1.5541,
      "theoretical_loss": 3.972867037777273,
      "tokens_seen": 436862976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381745235707121,
      "loss": 1.5188,
      "theoretical_loss": 3.9727356764973143,
      "tokens_seen": 436994048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043815446339017053,
      "loss": 1.5887,
      "theoretical_loss": 3.972604365640322,
      "tokens_seen": 437125120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381344032096289,
      "loss": 1.5462,
      "theoretical_loss": 3.97247310517183,
      "tokens_seen": 437256192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043811434302908726,
      "loss": 1.6721,
      "theoretical_loss": 3.972341895057408,
      "tokens_seen": 437387264
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.3990102708339691,
      "objective/train/avg_token_score": 0.014954428188502789,
      "objective/train/docs_used": 259916,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3668731451034546,
      "objective/train/original_loss": 3.7425804138183594,
      "objective/train/theoretical_loss": 3.972276308872222,
      "objective/train/tokens_used": 437452800,
      "objective/train/unlikelihood_loss": 0.18754182755947113,
      "theoretical_loss": 3.972276308872222,
      "tokens_seen": 437452800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380942828485456,
      "loss": 1.5532,
      "theoretical_loss": 3.972210735262657,
      "tokens_seen": 437518336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438074222668004,
      "loss": 1.7842,
      "theoretical_loss": 3.9720796257532136,
      "tokens_seen": 437649408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380541624874624,
      "loss": 1.5287,
      "theoretical_loss": 3.9719485664947483,
      "tokens_seen": 437780480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043803410230692076,
      "loss": 1.6341,
      "theoretical_loss": 3.9718175574529635,
      "tokens_seen": 437911552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380140421263791,
      "loss": 1.4245,
      "theoretical_loss": 3.971686598593596,
      "tokens_seen": 438042624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379939819458375,
      "loss": 1.5868,
      "theoretical_loss": 3.9715556898824182,
      "tokens_seen": 438173696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379739217652959,
      "loss": 1.6966,
      "theoretical_loss": 3.9714248312852343,
      "tokens_seen": 438304768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043795386158475427,
      "loss": 1.494,
      "theoretical_loss": 3.9712940227678803,
      "tokens_seen": 438435840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043793380140421263,
      "loss": 1.6852,
      "theoretical_loss": 3.971163264296229,
      "tokens_seen": 438566912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437913741223671,
      "loss": 1.4822,
      "theoretical_loss": 3.9710325558361856,
      "tokens_seen": 438697984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043789368104312935,
      "loss": 1.6556,
      "theoretical_loss": 3.970901897353688,
      "tokens_seen": 438829056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043787362086258777,
      "loss": 1.7292,
      "theoretical_loss": 3.9707712888147073,
      "tokens_seen": 438960128
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.3905791640281677,
      "objective/train/avg_token_score": 0.019270719960331917,
      "objective/train/docs_used": 260524,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4323198795318604,
      "objective/train/original_loss": 4.0027384757995605,
      "objective/train/theoretical_loss": 3.9706407301852487,
      "objective/train/tokens_used": 439091200,
      "objective/train/unlikelihood_loss": 0.153401181101799,
      "theoretical_loss": 3.9706407301852487,
      "tokens_seen": 439091200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043785356068204613,
      "loss": 1.6313,
      "theoretical_loss": 3.9706407301852487,
      "tokens_seen": 439091200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378335005015045,
      "loss": 1.6079,
      "theoretical_loss": 3.9705102214313506,
      "tokens_seen": 439222272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043781344032096286,
      "loss": 1.5786,
      "theoretical_loss": 3.970379762519084,
      "tokens_seen": 439353344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043779338014042127,
      "loss": 1.6811,
      "theoretical_loss": 3.9702493534145544,
      "tokens_seen": 439484416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043777331995987963,
      "loss": 1.5142,
      "theoretical_loss": 3.970118994083899,
      "tokens_seen": 439615488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437753259779338,
      "loss": 1.5964,
      "theoretical_loss": 3.969988684493289,
      "tokens_seen": 439746560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043773319959879636,
      "loss": 1.4616,
      "theoretical_loss": 3.969858424608928,
      "tokens_seen": 439877632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377131394182547,
      "loss": 1.6263,
      "theoretical_loss": 3.969728214397053,
      "tokens_seen": 440008704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043769307923771314,
      "loss": 1.5717,
      "theoretical_loss": 3.969598053823934,
      "tokens_seen": 440139776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043767301905717155,
      "loss": 1.5974,
      "theoretical_loss": 3.969467942855875,
      "tokens_seen": 440270848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376529588766299,
      "loss": 1.5539,
      "theoretical_loss": 3.9693378814592104,
      "tokens_seen": 440401920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376328986960883,
      "loss": 1.6919,
      "theoretical_loss": 3.969207869600309,
      "tokens_seen": 440532992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376128385155467,
      "loss": 1.5798,
      "theoretical_loss": 3.9690779072455733,
      "tokens_seen": 440664064
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4020650088787079,
      "objective/train/avg_token_score": 0.015384318307042122,
      "objective/train/docs_used": 261739,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4516440629959106,
      "objective/train/original_loss": 3.8263213634490967,
      "objective/train/theoretical_loss": 3.969012944621776,
      "objective/train/tokens_used": 440729600,
      "objective/train/unlikelihood_loss": 0.15117530524730682,
      "theoretical_loss": 3.969012944621776,
      "tokens_seen": 440729600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043759277833500506,
      "loss": 1.5678,
      "theoretical_loss": 3.968947994361436,
      "tokens_seen": 440795136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375727181544634,
      "loss": 1.5547,
      "theoretical_loss": 3.9688181309143657,
      "tokens_seen": 440926208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375526579739218,
      "loss": 1.4414,
      "theoretical_loss": 3.9686883168708613,
      "tokens_seen": 441057280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375325977933802,
      "loss": 1.6541,
      "theoretical_loss": 3.9685585521974556,
      "tokens_seen": 441188352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043751253761283856,
      "loss": 1.68,
      "theoretical_loss": 3.968428836860712,
      "tokens_seen": 441319424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374924774322969,
      "loss": 1.7028,
      "theoretical_loss": 3.96829917082723,
      "tokens_seen": 441450496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374724172517553,
      "loss": 1.6575,
      "theoretical_loss": 3.968169554063639,
      "tokens_seen": 441581568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043745235707121365,
      "loss": 1.6168,
      "theoretical_loss": 3.968039986536601,
      "tokens_seen": 441712640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043743229689067206,
      "loss": 1.7393,
      "theoretical_loss": 3.9679104682128106,
      "tokens_seen": 441843712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374122367101304,
      "loss": 1.7,
      "theoretical_loss": 3.9677809990589967,
      "tokens_seen": 441974784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373921765295888,
      "loss": 1.4573,
      "theoretical_loss": 3.9676515790419176,
      "tokens_seen": 442105856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043737211634904715,
      "loss": 1.5554,
      "theoretical_loss": 3.967522208128366,
      "tokens_seen": 442236928
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.43737781047821045,
      "objective/train/avg_token_score": 0.016550948843359947,
      "objective/train/docs_used": 262441,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5659313201904297,
      "objective/train/original_loss": 3.8397927284240723,
      "objective/train/theoretical_loss": 3.9673928862851655,
      "objective/train/tokens_used": 442368000,
      "objective/train/unlikelihood_loss": 0.18158647418022156,
      "theoretical_loss": 3.9673928862851655,
      "tokens_seen": 442368000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043735205616850557,
      "loss": 1.5389,
      "theoretical_loss": 3.9673928862851655,
      "tokens_seen": 442368000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043733199598796393,
      "loss": 1.6709,
      "theoretical_loss": 3.9672636134791737,
      "tokens_seen": 442499072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373119358074223,
      "loss": 1.6792,
      "theoretical_loss": 3.9671343896772786,
      "tokens_seen": 442630144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043729187562688065,
      "loss": 1.5301,
      "theoretical_loss": 3.967005214846401,
      "tokens_seen": 442761216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437271815446339,
      "loss": 1.6247,
      "theoretical_loss": 3.9668760889534944,
      "tokens_seen": 442892288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043725175526579743,
      "loss": 1.574,
      "theoretical_loss": 3.966747011965544,
      "tokens_seen": 443023360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372316950852558,
      "loss": 1.6382,
      "theoretical_loss": 3.9666179838495657,
      "tokens_seen": 443154432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043721163490471416,
      "loss": 1.5853,
      "theoretical_loss": 3.96648900457261,
      "tokens_seen": 443285504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371915747241725,
      "loss": 1.5827,
      "theoretical_loss": 3.9663600741017575,
      "tokens_seen": 443416576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043717151454363093,
      "loss": 1.5746,
      "theoretical_loss": 3.9662311924041203,
      "tokens_seen": 443547648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371514543630893,
      "loss": 1.4653,
      "theoretical_loss": 3.966102359446845,
      "tokens_seen": 443678720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043713139418254766,
      "loss": 1.5448,
      "theoretical_loss": 3.9659735751971064,
      "tokens_seen": 443809792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437111334002006,
      "loss": 1.6434,
      "theoretical_loss": 3.9658448396221138,
      "tokens_seen": 443940864
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.5477761626243591,
      "objective/train/avg_token_score": 0.017558608204126358,
      "objective/train/docs_used": 263839,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9970813989639282,
      "objective/train/original_loss": 3.930330753326416,
      "objective/train/theoretical_loss": 3.9657804900774085,
      "objective/train/tokens_used": 444006400,
      "objective/train/unlikelihood_loss": 0.15241247415542603,
      "theoretical_loss": 3.9657804900774085,
      "tokens_seen": 444006400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370912738214644,
      "loss": 1.8145,
      "theoretical_loss": 3.9657161526891076,
      "tokens_seen": 444071936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370712136409228,
      "loss": 1.6294,
      "theoretical_loss": 3.9655875143653594,
      "tokens_seen": 444203008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043705115346038116,
      "loss": 1.5046,
      "theoretical_loss": 3.965458924618173,
      "tokens_seen": 444334080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370310932798395,
      "loss": 1.842,
      "theoretical_loss": 3.9653303834148828,
      "tokens_seen": 444465152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370110330992979,
      "loss": 1.5623,
      "theoretical_loss": 3.965201890722857,
      "tokens_seen": 444596224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369909729187563,
      "loss": 1.5008,
      "theoretical_loss": 3.965073446509493,
      "tokens_seen": 444727296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043697091273821467,
      "loss": 1.5226,
      "theoretical_loss": 3.9649450507422204,
      "tokens_seen": 444858368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043695085255767303,
      "loss": 1.7281,
      "theoretical_loss": 3.964816703388501,
      "tokens_seen": 444989440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369307923771314,
      "loss": 1.5668,
      "theoretical_loss": 3.9646884044158277,
      "tokens_seen": 445120512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043691073219658975,
      "loss": 1.543,
      "theoretical_loss": 3.9645601537917248,
      "tokens_seen": 445251584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043689067201604817,
      "loss": 1.6472,
      "theoretical_loss": 3.9644319514837467,
      "tokens_seen": 445382656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043687061183550653,
      "loss": 1.6752,
      "theoretical_loss": 3.9643037974594817,
      "tokens_seen": 445513728
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.3345552384853363,
      "objective/train/avg_token_score": 0.010209061205387115,
      "objective/train/docs_used": 264540,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3382853269577026,
      "objective/train/original_loss": 4.270390033721924,
      "objective/train/theoretical_loss": 3.9641756916865463,
      "objective/train/tokens_used": 445644800,
      "objective/train/unlikelihood_loss": 0.15310260653495789,
      "theoretical_loss": 3.9641756916865463,
      "tokens_seen": 445644800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004368505516549649,
      "loss": 1.6619,
      "theoretical_loss": 3.9641756916865463,
      "tokens_seen": 445644800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043683049147442326,
      "loss": 1.4821,
      "theoretical_loss": 3.9640476341325916,
      "tokens_seen": 445775872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043681043129388167,
      "loss": 1.6354,
      "theoretical_loss": 3.9639196247652966,
      "tokens_seen": 445906944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043679037111334003,
      "loss": 1.4849,
      "theoretical_loss": 3.9637916635523736,
      "tokens_seen": 446038016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367703109327984,
      "loss": 1.4699,
      "theoretical_loss": 3.9636637504615653,
      "tokens_seen": 446169088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043675025075225676,
      "loss": 1.4246,
      "theoretical_loss": 3.9635358854606455,
      "tokens_seen": 446300160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367301905717151,
      "loss": 1.6212,
      "theoretical_loss": 3.963408068517419,
      "tokens_seen": 446431232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043671013039117354,
      "loss": 1.7125,
      "theoretical_loss": 3.963280299599722,
      "tokens_seen": 446562304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366900702106319,
      "loss": 1.4525,
      "theoretical_loss": 3.9631525786754214,
      "tokens_seen": 446693376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043667001003009026,
      "loss": 1.5724,
      "theoretical_loss": 3.9630249057124147,
      "tokens_seen": 446824448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366499498495486,
      "loss": 1.3806,
      "theoretical_loss": 3.9628972806786305,
      "tokens_seen": 446955520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043662988966900704,
      "loss": 1.7016,
      "theoretical_loss": 3.9627697035420284,
      "tokens_seen": 447086592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366098294884654,
      "loss": 1.6344,
      "theoretical_loss": 3.9626421742705986,
      "tokens_seen": 447217664
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.34390273690223694,
      "objective/train/avg_token_score": 0.016636604443192482,
      "objective/train/docs_used": 265849,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2081854343414307,
      "objective/train/original_loss": 4.167159557342529,
      "objective/train/theoretical_loss": 3.9625784275743294,
      "objective/train/tokens_used": 447283200,
      "objective/train/unlikelihood_loss": 0.16087184846401215,
      "theoretical_loss": 3.9625784275743294,
      "tokens_seen": 447283200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043658976930792376,
      "loss": 1.4704,
      "theoretical_loss": 3.9625146928323627,
      "tokens_seen": 447348736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043656970912738213,
      "loss": 1.3915,
      "theoretical_loss": 3.9623872591953715,
      "tokens_seen": 447479808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365496489468405,
      "loss": 1.6532,
      "theoretical_loss": 3.9622598733277083,
      "tokens_seen": 447610880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365295887662989,
      "loss": 1.5888,
      "theoretical_loss": 3.962132535197486,
      "tokens_seen": 447741952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043650952858575727,
      "loss": 1.3799,
      "theoretical_loss": 3.9620052447728478,
      "tokens_seen": 447873024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043648946840521563,
      "loss": 1.5288,
      "theoretical_loss": 3.9618780020219693,
      "tokens_seen": 448004096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436469408224674,
      "loss": 1.5376,
      "theoretical_loss": 3.961750806913054,
      "tokens_seen": 448135168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364493480441324,
      "loss": 1.5963,
      "theoretical_loss": 3.961623659414338,
      "tokens_seen": 448266240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043642928786359077,
      "loss": 1.5641,
      "theoretical_loss": 3.961496559494087,
      "tokens_seen": 448397312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043640922768304913,
      "loss": 1.3678,
      "theoretical_loss": 3.9613695071205974,
      "tokens_seen": 448528384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363891675025075,
      "loss": 1.4971,
      "theoretical_loss": 3.9612425022621953,
      "tokens_seen": 448659456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363691073219659,
      "loss": 1.5659,
      "theoretical_loss": 3.961115544887238,
      "tokens_seen": 448790528
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.3955278694629669,
      "objective/train/avg_token_score": 0.013740437105298042,
      "objective/train/docs_used": 266141,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3886208534240723,
      "objective/train/original_loss": 3.7399308681488037,
      "objective/train/theoretical_loss": 3.960988634964113,
      "objective/train/tokens_used": 448921600,
      "objective/train/unlikelihood_loss": 0.21553213894367218,
      "theoretical_loss": 3.960988634964113,
      "tokens_seen": 448921600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363490471414243,
      "loss": 1.6211,
      "theoretical_loss": 3.960988634964113,
      "tokens_seen": 448921600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043632898696088264,
      "loss": 1.5591,
      "theoretical_loss": 3.9608617724612367,
      "tokens_seen": 449052672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436308926780341,
      "loss": 1.3739,
      "theoretical_loss": 3.960734957347058,
      "tokens_seen": 449183744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043628886659979936,
      "loss": 1.6942,
      "theoretical_loss": 3.9606081895900545,
      "tokens_seen": 449314816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362688064192578,
      "loss": 1.6639,
      "theoretical_loss": 3.9604814691587347,
      "tokens_seen": 449445888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043624874623871614,
      "loss": 1.4516,
      "theoretical_loss": 3.960354796021636,
      "tokens_seen": 449576960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362286860581745,
      "loss": 1.4179,
      "theoretical_loss": 3.960228170147327,
      "tokens_seen": 449708032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043620862587763286,
      "loss": 1.3733,
      "theoretical_loss": 3.960101591504406,
      "tokens_seen": 449839104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361885656970913,
      "loss": 1.5806,
      "theoretical_loss": 3.9599750600615025,
      "tokens_seen": 449970176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361685055165497,
      "loss": 1.3499,
      "theoretical_loss": 3.959848575787273,
      "tokens_seen": 450101248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043614844533600806,
      "loss": 1.55,
      "theoretical_loss": 3.9597221386504073,
      "tokens_seen": 450232320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361283851554664,
      "loss": 1.5825,
      "theoretical_loss": 3.9595957486196225,
      "tokens_seen": 450363392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361083249749248,
      "loss": 1.4968,
      "theoretical_loss": 3.9594694056636675,
      "tokens_seen": 450494464
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.3633919954299927,
      "objective/train/avg_token_score": 0.05124437436461449,
      "objective/train/docs_used": 266764,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3248687982559204,
      "objective/train/original_loss": 4.117973327636719,
      "objective/train/theoretical_loss": 3.9594062518289923,
      "objective/train/tokens_used": 450560000,
      "objective/train/unlikelihood_loss": 0.1388237476348877,
      "theoretical_loss": 3.9594062518289923,
      "tokens_seen": 450560000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360882647943832,
      "loss": 1.3466,
      "theoretical_loss": 3.9593431097513188,
      "tokens_seen": 450625536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043606820461384156,
      "loss": 1.5877,
      "theoretical_loss": 3.9592168608513854,
      "tokens_seen": 450756608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360481444332999,
      "loss": 1.4062,
      "theoretical_loss": 3.9590906589327037,
      "tokens_seen": 450887680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360280842527583,
      "loss": 1.3673,
      "theoretical_loss": 3.958964503964141,
      "tokens_seen": 451018752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360080240722167,
      "loss": 1.4466,
      "theoretical_loss": 3.9588383959145945,
      "tokens_seen": 451149824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043598796389167507,
      "loss": 1.4174,
      "theoretical_loss": 3.95871233475299,
      "tokens_seen": 451280896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043596790371113343,
      "loss": 1.5887,
      "theoretical_loss": 3.9585863204482834,
      "tokens_seen": 451411968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004359478435305918,
      "loss": 1.5364,
      "theoretical_loss": 3.9584603529694604,
      "tokens_seen": 451543040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043592778335005015,
      "loss": 1.6187,
      "theoretical_loss": 3.958334432285537,
      "tokens_seen": 451674112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043590772316950857,
      "loss": 1.5349,
      "theoretical_loss": 3.9582085583655564,
      "tokens_seen": 451805184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043588766298896693,
      "loss": 1.3348,
      "theoretical_loss": 3.9580827311785924,
      "tokens_seen": 451936256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358676028084253,
      "loss": 1.4182,
      "theoretical_loss": 3.9579569506937498,
      "tokens_seen": 452067328
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.41373410820961,
      "objective/train/avg_token_score": 0.03839399293065071,
      "objective/train/docs_used": 268192,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.432794213294983,
      "objective/train/original_loss": 4.107253551483154,
      "objective/train/theoretical_loss": 3.9578312168801597,
      "objective/train/tokens_used": 452198400,
      "objective/train/unlikelihood_loss": 0.13305017352104187,
      "theoretical_loss": 3.9578312168801597,
      "tokens_seen": 452198400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043584754262788366,
      "loss": 1.4714,
      "theoretical_loss": 3.9578312168801597,
      "tokens_seen": 452198400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043582748244734207,
      "loss": 1.683,
      "theoretical_loss": 3.9577055297069856,
      "tokens_seen": 452329472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043580742226680043,
      "loss": 1.4859,
      "theoretical_loss": 3.9575798891434184,
      "tokens_seen": 452460544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357873620862588,
      "loss": 1.4563,
      "theoretical_loss": 3.957454295158679,
      "tokens_seen": 452591616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043576730190571716,
      "loss": 1.5263,
      "theoretical_loss": 3.9573287477220167,
      "tokens_seen": 452722688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357472417251755,
      "loss": 1.6256,
      "theoretical_loss": 3.957203246802711,
      "tokens_seen": 452853760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043572718154463394,
      "loss": 1.6589,
      "theoretical_loss": 3.9570777923700704,
      "tokens_seen": 452984832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357071213640923,
      "loss": 1.3931,
      "theoretical_loss": 3.956952384393432,
      "tokens_seen": 453115904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043568706118355066,
      "loss": 1.5144,
      "theoretical_loss": 3.9568270228421616,
      "tokens_seen": 453246976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435667001003009,
      "loss": 1.4202,
      "theoretical_loss": 3.9567017076856565,
      "tokens_seen": 453378048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043564694082246744,
      "loss": 1.5056,
      "theoretical_loss": 3.9565764388933395,
      "tokens_seen": 453509120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356268806419258,
      "loss": 1.5587,
      "theoretical_loss": 3.9564512164346652,
      "tokens_seen": 453640192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043560682046138417,
      "loss": 1.6948,
      "theoretical_loss": 3.956326040279116,
      "tokens_seen": 453771264
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4516128897666931,
      "objective/train/avg_token_score": 0.02117766998708248,
      "objective/train/docs_used": 268717,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.552639126777649,
      "objective/train/original_loss": 4.055312156677246,
      "objective/train/theoretical_loss": 3.956263469555484,
      "objective/train/tokens_used": 453836800,
      "objective/train/unlikelihood_loss": 0.12762990593910217,
      "theoretical_loss": 3.956263469555484,
      "tokens_seen": 453836800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043558676028084253,
      "loss": 1.6062,
      "theoretical_loss": 3.956200910396202,
      "tokens_seen": 453902336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355667001003009,
      "loss": 1.3842,
      "theoretical_loss": 3.956075826755465,
      "tokens_seen": 454033408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355466399197593,
      "loss": 1.5806,
      "theoretical_loss": 3.955950789326474,
      "tokens_seen": 454164480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043552657973921767,
      "loss": 1.4914,
      "theoretical_loss": 3.955825798078826,
      "tokens_seen": 454295552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043550651955867603,
      "loss": 1.4215,
      "theoretical_loss": 3.9557008529821474,
      "tokens_seen": 454426624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354864593781344,
      "loss": 1.3912,
      "theoretical_loss": 3.955575954006095,
      "tokens_seen": 454557696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354663991975928,
      "loss": 1.5038,
      "theoretical_loss": 3.955451101120352,
      "tokens_seen": 454688768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043544633901705117,
      "loss": 1.6058,
      "theoretical_loss": 3.9553262942946312,
      "tokens_seen": 454819840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043542627883650953,
      "loss": 1.4004,
      "theoretical_loss": 3.9552015334986743,
      "tokens_seen": 454950912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354062186559679,
      "loss": 1.6557,
      "theoretical_loss": 3.95507681870225,
      "tokens_seen": 455081984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353861584754263,
      "loss": 1.3329,
      "theoretical_loss": 3.954952149875159,
      "tokens_seen": 455213056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353660982948847,
      "loss": 1.506,
      "theoretical_loss": 3.954827526987226,
      "tokens_seen": 455344128
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.35752686858177185,
      "objective/train/avg_token_score": 0.016128363087773323,
      "objective/train/docs_used": 269967,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1892235279083252,
      "objective/train/original_loss": 3.846442937850952,
      "objective/train/theoretical_loss": 3.954702950008308,
      "objective/train/tokens_used": 455475200,
      "objective/train/unlikelihood_loss": 0.21875594556331635,
      "theoretical_loss": 3.954702950008308,
      "tokens_seen": 455475200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043534603811434304,
      "loss": 1.2484,
      "theoretical_loss": 3.954702950008308,
      "tokens_seen": 455475200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353259779338014,
      "loss": 1.3839,
      "theoretical_loss": 3.9545784189082887,
      "tokens_seen": 455606272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043530591775325976,
      "loss": 1.5678,
      "theoretical_loss": 3.9544539336570796,
      "tokens_seen": 455737344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352858575727182,
      "loss": 1.4671,
      "theoretical_loss": 3.954329494224623,
      "tokens_seen": 455868416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043526579739217654,
      "loss": 1.5832,
      "theoretical_loss": 3.9542051005808867,
      "tokens_seen": 455999488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352457372116349,
      "loss": 1.5173,
      "theoretical_loss": 3.954080752695868,
      "tokens_seen": 456130560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043522567703109326,
      "loss": 1.4201,
      "theoretical_loss": 3.953956450539594,
      "tokens_seen": 456261632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352056168505517,
      "loss": 1.634,
      "theoretical_loss": 3.953832194082117,
      "tokens_seen": 456392704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043518555667001004,
      "loss": 1.5989,
      "theoretical_loss": 3.95370798329352,
      "tokens_seen": 456523776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351654964894684,
      "loss": 1.5042,
      "theoretical_loss": 3.9535838181439127,
      "tokens_seen": 456654848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043514543630892677,
      "loss": 1.5895,
      "theoretical_loss": 3.953459698603434,
      "tokens_seen": 456785920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043512537612838513,
      "loss": 1.611,
      "theoretical_loss": 3.95333562464225,
      "tokens_seen": 456916992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043510531594784355,
      "loss": 1.5907,
      "theoretical_loss": 3.953211596230556,
      "tokens_seen": 457048064
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.5039711594581604,
      "objective/train/avg_token_score": 0.018044468015432358,
      "objective/train/docs_used": 270622,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7069590091705322,
      "objective/train/original_loss": 3.6348330974578857,
      "objective/train/theoretical_loss": 3.953149599096461,
      "objective/train/tokens_used": 457113600,
      "objective/train/unlikelihood_loss": 0.2166004627943039,
      "theoretical_loss": 3.953149599096461,
      "tokens_seen": 457113600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350852557673019,
      "loss": 1.5883,
      "theoretical_loss": 3.9530876133385737,
      "tokens_seen": 457179136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043506519558676027,
      "loss": 1.4272,
      "theoretical_loss": 3.9529636759365543,
      "tokens_seen": 457310208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043504513540621863,
      "loss": 1.448,
      "theoretical_loss": 3.952839783994776,
      "tokens_seen": 457441280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043502507522567705,
      "loss": 1.4977,
      "theoretical_loss": 3.9527159374835463,
      "tokens_seen": 457572352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350050150451354,
      "loss": 1.4147,
      "theoretical_loss": 3.9525921363731977,
      "tokens_seen": 457703424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349849548645938,
      "loss": 1.6583,
      "theoretical_loss": 3.9524683806340937,
      "tokens_seen": 457834496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043496489468405214,
      "loss": 1.6028,
      "theoretical_loss": 3.9523446702366245,
      "tokens_seen": 457965568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349448345035105,
      "loss": 1.5376,
      "theoretical_loss": 3.9522210051512077,
      "tokens_seen": 458096640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349247743229689,
      "loss": 1.4196,
      "theoretical_loss": 3.9520973853482886,
      "tokens_seen": 458227712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349047141424273,
      "loss": 1.5257,
      "theoretical_loss": 3.951973810798341,
      "tokens_seen": 458358784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043488465396188564,
      "loss": 1.5334,
      "theoretical_loss": 3.9518502814718652,
      "tokens_seen": 458489856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434864593781344,
      "loss": 1.3228,
      "theoretical_loss": 3.951726797339391,
      "tokens_seen": 458620928
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.39320626854896545,
      "objective/train/avg_token_score": 0.018131617456674576,
      "objective/train/docs_used": 271952,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.354129433631897,
      "objective/train/original_loss": 3.985093355178833,
      "objective/train/theoretical_loss": 3.9516033583714734,
      "objective/train/tokens_used": 458752000,
      "objective/train/unlikelihood_loss": 0.19177991151809692,
      "theoretical_loss": 3.9516033583714734,
      "tokens_seen": 458752000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348445336008024,
      "loss": 1.514,
      "theoretical_loss": 3.9516033583714734,
      "tokens_seen": 458752000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348244734202608,
      "loss": 1.4862,
      "theoretical_loss": 3.9514799645386973,
      "tokens_seen": 458883072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043480441323971914,
      "loss": 1.5574,
      "theoretical_loss": 3.951356615811674,
      "tokens_seen": 459014144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347843530591775,
      "loss": 1.3874,
      "theoretical_loss": 3.951233312161042,
      "tokens_seen": 459145216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043476429287863587,
      "loss": 1.5508,
      "theoretical_loss": 3.9511100535574686,
      "tokens_seen": 459276288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347442326980943,
      "loss": 1.5419,
      "theoretical_loss": 3.9509868399716472,
      "tokens_seen": 459407360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043472417251755265,
      "loss": 1.2042,
      "theoretical_loss": 3.950863671374299,
      "tokens_seen": 459538432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434704112337011,
      "loss": 1.5611,
      "theoretical_loss": 3.9507405477361726,
      "tokens_seen": 459669504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346840521564694,
      "loss": 1.4663,
      "theoretical_loss": 3.9506174690280442,
      "tokens_seen": 459800576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043466399197592784,
      "loss": 1.5522,
      "theoretical_loss": 3.9504944352207176,
      "tokens_seen": 459931648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346439317953862,
      "loss": 1.5421,
      "theoretical_loss": 3.9503714462850237,
      "tokens_seen": 460062720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043462387161484457,
      "loss": 1.557,
      "theoretical_loss": 3.950248502191819,
      "tokens_seen": 460193792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043460381143430293,
      "loss": 1.5097,
      "theoretical_loss": 3.9501256029119896,
      "tokens_seen": 460324864
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4094574749469757,
      "objective/train/avg_token_score": 0.011406373232603073,
      "objective/train/docs_used": 272724,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6004735231399536,
      "objective/train/original_loss": 4.182957649230957,
      "objective/train/theoretical_loss": 3.950064170068,
      "objective/train/tokens_used": 460390400,
      "objective/train/unlikelihood_loss": 0.14253197610378265,
      "theoretical_loss": 3.950064170068,
      "tokens_seen": 460390400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345837512537613,
      "loss": 1.5761,
      "theoretical_loss": 3.9500027484164484,
      "tokens_seen": 460455936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345636910732197,
      "loss": 1.3552,
      "theoretical_loss": 3.9498799386761334,
      "tokens_seen": 460587008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043454363089267807,
      "loss": 1.4481,
      "theoretical_loss": 3.9497571736620127,
      "tokens_seen": 460718080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043452357071213643,
      "loss": 1.4637,
      "theoretical_loss": 3.9496344533450793,
      "tokens_seen": 460849152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345035105315948,
      "loss": 1.6593,
      "theoretical_loss": 3.9495117776963538,
      "tokens_seen": 460980224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344834503510532,
      "loss": 1.495,
      "theoretical_loss": 3.9493891466868845,
      "tokens_seen": 461111296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043446339017051157,
      "loss": 1.4484,
      "theoretical_loss": 3.9492665602877457,
      "tokens_seen": 461242368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043444332998996993,
      "loss": 1.4955,
      "theoretical_loss": 3.9491440184700393,
      "tokens_seen": 461373440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344232698094283,
      "loss": 1.4501,
      "theoretical_loss": 3.9490215212048936,
      "tokens_seen": 461504512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344032096288867,
      "loss": 1.4812,
      "theoretical_loss": 3.948899068463465,
      "tokens_seen": 461635584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343831494483451,
      "loss": 1.6095,
      "theoretical_loss": 3.948776660216935,
      "tokens_seen": 461766656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043436308926780344,
      "loss": 1.4606,
      "theoretical_loss": 3.948654296436513,
      "tokens_seen": 461897728
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4204545319080353,
      "objective/train/avg_token_score": 0.009838702157139778,
      "objective/train/docs_used": 273842,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5137208700180054,
      "objective/train/original_loss": 3.8193166255950928,
      "objective/train/theoretical_loss": 3.9485319770934355,
      "objective/train/tokens_used": 462028800,
      "objective/train/unlikelihood_loss": 0.21145619451999664,
      "theoretical_loss": 3.9485319770934355,
      "tokens_seen": 462028800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343430290872618,
      "loss": 1.5253,
      "theoretical_loss": 3.9485319770934355,
      "tokens_seen": 462028800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043432296890672016,
      "loss": 1.4442,
      "theoretical_loss": 3.9484097021589646,
      "tokens_seen": 462159872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343029087261786,
      "loss": 1.3611,
      "theoretical_loss": 3.94828747160439,
      "tokens_seen": 462290944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043428284854563694,
      "loss": 1.5169,
      "theoretical_loss": 3.9481652854010285,
      "tokens_seen": 462422016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342627883650953,
      "loss": 1.602,
      "theoretical_loss": 3.9480431435202226,
      "tokens_seen": 462553088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043424272818455367,
      "loss": 1.4654,
      "theoretical_loss": 3.9479210459333407,
      "tokens_seen": 462684160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342226680040121,
      "loss": 1.4747,
      "theoretical_loss": 3.94779899261178,
      "tokens_seen": 462815232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043420260782347044,
      "loss": 1.6772,
      "theoretical_loss": 3.9476769835269634,
      "tokens_seen": 462946304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341825476429288,
      "loss": 1.6903,
      "theoretical_loss": 3.94755501865034,
      "tokens_seen": 463077376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043416248746238717,
      "loss": 1.5698,
      "theoretical_loss": 3.9474330979533843,
      "tokens_seen": 463208448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043414242728184553,
      "loss": 1.5586,
      "theoretical_loss": 3.947311221407599,
      "tokens_seen": 463339520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043412236710130395,
      "loss": 1.5921,
      "theoretical_loss": 3.9471893889845138,
      "tokens_seen": 463470592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341023069207623,
      "loss": 1.3821,
      "theoretical_loss": 3.947067600655682,
      "tokens_seen": 463601664
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.3299120366573334,
      "objective/train/avg_token_score": 0.027434295043349266,
      "objective/train/docs_used": 274430,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1526551246643066,
      "objective/train/original_loss": 3.9500741958618164,
      "objective/train/theoretical_loss": 3.94700672301773,
      "objective/train/tokens_used": 463667200,
      "objective/train/unlikelihood_loss": 0.1668168306350708,
      "theoretical_loss": 3.94700672301773,
      "tokens_seen": 463667200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043408224674022067,
      "loss": 1.4242,
      "theoretical_loss": 3.946945856392687,
      "tokens_seen": 463732736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043406218655967903,
      "loss": 1.3769,
      "theoretical_loss": 3.9468241561671342,
      "tokens_seen": 463863808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043404212637913745,
      "loss": 1.3464,
      "theoretical_loss": 3.946702499950659,
      "tokens_seen": 463994880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340220661985958,
      "loss": 1.5064,
      "theoretical_loss": 3.946580887714921,
      "tokens_seen": 464125952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340020060180542,
      "loss": 1.322,
      "theoretical_loss": 3.946459319431608,
      "tokens_seen": 464257024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043398194583751254,
      "loss": 1.5324,
      "theoretical_loss": 3.9463377950724317,
      "tokens_seen": 464388096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339618856569709,
      "loss": 1.3533,
      "theoretical_loss": 3.9462163146091305,
      "tokens_seen": 464519168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339418254764293,
      "loss": 1.5288,
      "theoretical_loss": 3.9460948780134713,
      "tokens_seen": 464650240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339217652958877,
      "loss": 1.5073,
      "theoretical_loss": 3.9459734852572437,
      "tokens_seen": 464781312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043390170511534604,
      "loss": 1.5032,
      "theoretical_loss": 3.945852136312265,
      "tokens_seen": 464912384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338816449348044,
      "loss": 1.6266,
      "theoretical_loss": 3.94573083115038,
      "tokens_seen": 465043456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338615847542628,
      "loss": 1.5327,
      "theoretical_loss": 3.9456095697434566,
      "tokens_seen": 465174528
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.32600194215774536,
      "objective/train/avg_token_score": 0.02726374939084053,
      "objective/train/docs_used": 275809,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0033814907073975,
      "objective/train/original_loss": 3.863251209259033,
      "objective/train/theoretical_loss": 3.945488352063391,
      "objective/train/tokens_used": 465305600,
      "objective/train/unlikelihood_loss": 0.22456102073192596,
      "theoretical_loss": 3.945488352063391,
      "tokens_seen": 465305600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338415245737212,
      "loss": 1.4332,
      "theoretical_loss": 3.945488352063391,
      "tokens_seen": 465305600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043382146439317954,
      "loss": 1.4683,
      "theoretical_loss": 3.945367178082104,
      "tokens_seen": 465436672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338014042126379,
      "loss": 1.4562,
      "theoretical_loss": 3.945246047771543,
      "tokens_seen": 465567744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043378134403209627,
      "loss": 1.6305,
      "theoretical_loss": 3.945124961103682,
      "tokens_seen": 465698816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337612838515547,
      "loss": 1.5113,
      "theoretical_loss": 3.945003918050518,
      "tokens_seen": 465829888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043374122367101305,
      "loss": 1.5304,
      "theoretical_loss": 3.9448829185840784,
      "tokens_seen": 465960960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337211634904714,
      "loss": 1.4951,
      "theoretical_loss": 3.944761962676412,
      "tokens_seen": 466092032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043370110330992977,
      "loss": 1.4054,
      "theoretical_loss": 3.9446410502995963,
      "tokens_seen": 466223104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336810431293882,
      "loss": 1.4125,
      "theoretical_loss": 3.9445201814257334,
      "tokens_seen": 466354176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043366098294884655,
      "loss": 1.4092,
      "theoretical_loss": 3.9443993560269504,
      "tokens_seen": 466485248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336409227683049,
      "loss": 1.3047,
      "theoretical_loss": 3.944278574075402,
      "tokens_seen": 466616320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336208625877633,
      "loss": 1.5973,
      "theoretical_loss": 3.9441578355432663,
      "tokens_seen": 466747392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043360080240722164,
      "loss": 1.6775,
      "theoretical_loss": 3.9440371404027488,
      "tokens_seen": 466878464
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4305962920188904,
      "objective/train/avg_token_score": 0.008782853372395039,
      "objective/train/docs_used": 276320,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.639618158340454,
      "objective/train/original_loss": 4.319045543670654,
      "objective/train/theoretical_loss": 3.943976809095668,
      "objective/train/tokens_used": 466944000,
      "objective/train/unlikelihood_loss": 0.18612205982208252,
      "theoretical_loss": 3.943976809095668,
      "tokens_seen": 466944000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043358074222668005,
      "loss": 1.5497,
      "theoretical_loss": 3.94391648862608,
      "tokens_seen": 467009536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335606820461384,
      "loss": 1.5624,
      "theoretical_loss": 3.9437958801855157,
      "tokens_seen": 467140608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335406218655968,
      "loss": 1.5378,
      "theoretical_loss": 3.9436753150533375,
      "tokens_seen": 467271680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043352056168505514,
      "loss": 1.7575,
      "theoretical_loss": 3.9435547932018524,
      "tokens_seen": 467402752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043350050150451356,
      "loss": 1.7,
      "theoretical_loss": 3.943434314603394,
      "tokens_seen": 467533824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334804413239719,
      "loss": 1.4659,
      "theoretical_loss": 3.9433138792303177,
      "tokens_seen": 467664896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334603811434303,
      "loss": 1.3896,
      "theoretical_loss": 3.943193487055009,
      "tokens_seen": 467795968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043344032096288864,
      "loss": 1.3853,
      "theoretical_loss": 3.9430731380498765,
      "tokens_seen": 467927040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433420260782347,
      "loss": 1.3526,
      "theoretical_loss": 3.9429528321873537,
      "tokens_seen": 468058112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334002006018054,
      "loss": 1.5852,
      "theoretical_loss": 3.9428325694398993,
      "tokens_seen": 468189184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333801404212638,
      "loss": 1.6151,
      "theoretical_loss": 3.94271234978,
      "tokens_seen": 468320256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043336008024072215,
      "loss": 1.5538,
      "theoretical_loss": 3.942592173180163,
      "tokens_seen": 468451328
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.5079423189163208,
      "objective/train/avg_token_score": 0.011020613834261894,
      "objective/train/docs_used": 277290,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8621379137039185,
      "objective/train/original_loss": 3.9069674015045166,
      "objective/train/theoretical_loss": 3.942472039612926,
      "objective/train/tokens_used": 468582400,
      "objective/train/unlikelihood_loss": 0.13427872955799103,
      "theoretical_loss": 3.942472039612926,
      "tokens_seen": 468582400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333400200601805,
      "loss": 1.5221,
      "theoretical_loss": 3.942472039612926,
      "tokens_seen": 468582400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333199598796389,
      "loss": 1.543,
      "theoretical_loss": 3.9423519490508476,
      "tokens_seen": 468713472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332998996990973,
      "loss": 1.5703,
      "theoretical_loss": 3.9422319014665144,
      "tokens_seen": 468844544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043327983951855565,
      "loss": 1.419,
      "theoretical_loss": 3.942111896832537,
      "tokens_seen": 468975616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433259779338014,
      "loss": 1.4121,
      "theoretical_loss": 3.9419919351215498,
      "tokens_seen": 469106688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043323971915747243,
      "loss": 1.481,
      "theoretical_loss": 3.941872016306215,
      "tokens_seen": 469237760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332196589769308,
      "loss": 1.5854,
      "theoretical_loss": 3.9417521403592186,
      "tokens_seen": 469368832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043319959879638915,
      "loss": 1.7019,
      "theoretical_loss": 3.94163230725327,
      "tokens_seen": 469499904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043317953861584757,
      "loss": 1.5685,
      "theoretical_loss": 3.941512516961107,
      "tokens_seen": 469630976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043315947843530593,
      "loss": 1.4466,
      "theoretical_loss": 3.9413927694554887,
      "tokens_seen": 469762048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043313941825476435,
      "loss": 1.5199,
      "theoretical_loss": 3.9412730647092022,
      "tokens_seen": 469893120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331193580742227,
      "loss": 1.4827,
      "theoretical_loss": 3.9411534026950568,
      "tokens_seen": 470024192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043309929789368107,
      "loss": 1.6589,
      "theoretical_loss": 3.9410337833858886,
      "tokens_seen": 470155264
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.3777492642402649,
      "objective/train/avg_token_score": 0.027359778061509132,
      "objective/train/docs_used": 277851,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.291692852973938,
      "objective/train/original_loss": 4.366660118103027,
      "objective/train/theoretical_loss": 3.9409739897371896,
      "objective/train/tokens_used": 470220800,
      "objective/train/unlikelihood_loss": 0.15265023708343506,
      "theoretical_loss": 3.9409739897371896,
      "tokens_seen": 470220800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043307923771313943,
      "loss": 1.3202,
      "theoretical_loss": 3.9409142067545586,
      "tokens_seen": 470286336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043305917753259785,
      "loss": 1.2957,
      "theoretical_loss": 3.940794672773951,
      "tokens_seen": 470417408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330391173520562,
      "loss": 1.5722,
      "theoretical_loss": 3.9406751814169763,
      "tokens_seen": 470548480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330190571715146,
      "loss": 1.4456,
      "theoretical_loss": 3.940555732656569,
      "tokens_seen": 470679552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043299899699097294,
      "loss": 1.4136,
      "theoretical_loss": 3.9404363264656883,
      "tokens_seen": 470810624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329789368104313,
      "loss": 1.3883,
      "theoretical_loss": 3.9403169628173185,
      "tokens_seen": 470941696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329588766298897,
      "loss": 1.5931,
      "theoretical_loss": 3.9401976416844677,
      "tokens_seen": 471072768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329388164493481,
      "loss": 1.5737,
      "theoretical_loss": 3.94007836304017,
      "tokens_seen": 471203840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043291875626880644,
      "loss": 1.6797,
      "theoretical_loss": 3.9399591268574836,
      "tokens_seen": 471334912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328986960882648,
      "loss": 1.5247,
      "theoretical_loss": 3.93983993310949,
      "tokens_seen": 471465984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328786359077232,
      "loss": 1.6422,
      "theoretical_loss": 3.9397207817692967,
      "tokens_seen": 471597056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328585757271816,
      "loss": 1.6177,
      "theoretical_loss": 3.9396016728100363,
      "tokens_seen": 471728128
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.33987048268318176,
      "objective/train/avg_token_score": 0.018589701503515244,
      "objective/train/docs_used": 279270,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2118350267410278,
      "objective/train/original_loss": 3.926861524581909,
      "objective/train/theoretical_loss": 3.939482606204863,
      "objective/train/tokens_used": 471859200,
      "objective/train/unlikelihood_loss": 0.1467958688735962,
      "theoretical_loss": 3.939482606204863,
      "tokens_seen": 471859200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043283851554663994,
      "loss": 1.5283,
      "theoretical_loss": 3.939482606204863,
      "tokens_seen": 471859200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328184553660983,
      "loss": 1.4258,
      "theoretical_loss": 3.939363581926959,
      "tokens_seen": 471990272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043279839518555667,
      "loss": 1.4664,
      "theoretical_loss": 3.939244599949528,
      "tokens_seen": 472121344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327783350050151,
      "loss": 1.3412,
      "theoretical_loss": 3.9391256602457996,
      "tokens_seen": 472252416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043275827482447345,
      "loss": 1.4782,
      "theoretical_loss": 3.9390067627890275,
      "tokens_seen": 472383488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327382146439318,
      "loss": 1.4871,
      "theoretical_loss": 3.9388879075524907,
      "tokens_seen": 472514560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043271815446339017,
      "loss": 1.6205,
      "theoretical_loss": 3.9387690945094906,
      "tokens_seen": 472645632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326980942828486,
      "loss": 1.3513,
      "theoretical_loss": 3.938650323633354,
      "tokens_seen": 472776704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043267803410230695,
      "loss": 1.6774,
      "theoretical_loss": 3.9385315948974315,
      "tokens_seen": 472907776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326579739217653,
      "loss": 1.542,
      "theoretical_loss": 3.938412908275099,
      "tokens_seen": 473038848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326379137412237,
      "loss": 1.4591,
      "theoretical_loss": 3.938294263739755,
      "tokens_seen": 473169920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043261785356068204,
      "loss": 1.5059,
      "theoretical_loss": 3.938175661264823,
      "tokens_seen": 473300992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043259779338014045,
      "loss": 1.4335,
      "theoretical_loss": 3.938057100823751,
      "tokens_seen": 473432064
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4881476163864136,
      "objective/train/avg_token_score": 0.017582368105649948,
      "objective/train/docs_used": 279967,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6766897439956665,
      "objective/train/original_loss": 3.8044707775115967,
      "objective/train/theoretical_loss": 3.9379978363576216,
      "objective/train/tokens_used": 473497600,
      "objective/train/unlikelihood_loss": 0.20410603284835815,
      "theoretical_loss": 3.9379978363576216,
      "tokens_seen": 473497600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325777331995988,
      "loss": 1.7232,
      "theoretical_loss": 3.937938582390011,
      "tokens_seen": 473563136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325576730190572,
      "loss": 1.5215,
      "theoretical_loss": 3.9378201059370976,
      "tokens_seen": 473694208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043253761283851554,
      "loss": 1.4315,
      "theoretical_loss": 3.9377016714385316,
      "tokens_seen": 473825280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043251755265797396,
      "loss": 1.5092,
      "theoretical_loss": 3.937583278867857,
      "tokens_seen": 473956352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324974924774323,
      "loss": 1.5942,
      "theoretical_loss": 3.9374649281986405,
      "tokens_seen": 474087424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324774322968907,
      "loss": 1.5666,
      "theoretical_loss": 3.937346619404475,
      "tokens_seen": 474218496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043245737211634904,
      "loss": 1.666,
      "theoretical_loss": 3.937228352458976,
      "tokens_seen": 474349568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324373119358074,
      "loss": 1.6075,
      "theoretical_loss": 3.9371101273357825,
      "tokens_seen": 474480640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324172517552658,
      "loss": 1.4956,
      "theoretical_loss": 3.9369919440085592,
      "tokens_seen": 474611712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323971915747242,
      "loss": 1.4977,
      "theoretical_loss": 3.9368738024509926,
      "tokens_seen": 474742784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043237713139418255,
      "loss": 1.6371,
      "theoretical_loss": 3.936755702636794,
      "tokens_seen": 474873856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323570712136409,
      "loss": 1.6294,
      "theoretical_loss": 3.936637644539699,
      "tokens_seen": 475004928
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.469696968793869,
      "objective/train/avg_token_score": 0.015840815380215645,
      "objective/train/docs_used": 280356,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.593431830406189,
      "objective/train/original_loss": 3.748237133026123,
      "objective/train/theoretical_loss": 3.936519628133466,
      "objective/train/tokens_used": 475136000,
      "objective/train/unlikelihood_loss": 0.19854024052619934,
      "theoretical_loss": 3.936519628133466,
      "tokens_seen": 475136000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323370110330993,
      "loss": 1.6438,
      "theoretical_loss": 3.936519628133466,
      "tokens_seen": 475136000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323169508525577,
      "loss": 1.4211,
      "theoretical_loss": 3.9364016533918766,
      "tokens_seen": 475267072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043229689067201605,
      "loss": 1.4308,
      "theoretical_loss": 3.9362837202887384,
      "tokens_seen": 475398144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322768304914744,
      "loss": 1.5334,
      "theoretical_loss": 3.9361658287978813,
      "tokens_seen": 475529216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043225677031093283,
      "loss": 1.4563,
      "theoretical_loss": 3.936047978893158,
      "tokens_seen": 475660288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322367101303912,
      "loss": 1.3508,
      "theoretical_loss": 3.9359301705484455,
      "tokens_seen": 475791360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043221664994984955,
      "loss": 1.6533,
      "theoretical_loss": 3.935812403737646,
      "tokens_seen": 475922432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321965897693079,
      "loss": 1.7426,
      "theoretical_loss": 3.9356946784346825,
      "tokens_seen": 476053504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321765295887663,
      "loss": 1.5579,
      "theoretical_loss": 3.9355769946135037,
      "tokens_seen": 476184576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321564694082247,
      "loss": 1.5945,
      "theoretical_loss": 3.9354593522480803,
      "tokens_seen": 476315648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043213640922768306,
      "loss": 1.5677,
      "theoretical_loss": 3.935341751312408,
      "tokens_seen": 476446720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321163490471414,
      "loss": 1.4719,
      "theoretical_loss": 3.935224191780505,
      "tokens_seen": 476577792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320962888665998,
      "loss": 1.7156,
      "theoretical_loss": 3.9351066736264126,
      "tokens_seen": 476708864
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4970674514770508,
      "objective/train/avg_token_score": 0.05302438512444496,
      "objective/train/docs_used": 281398,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6894476413726807,
      "objective/train/original_loss": 4.0028076171875,
      "objective/train/theoretical_loss": 3.93504793005794,
      "objective/train/tokens_used": 476774400,
      "objective/train/unlikelihood_loss": 0.18129520118236542,
      "theoretical_loss": 3.93504793005794,
      "tokens_seen": 476774400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320762286860582,
      "loss": 1.544,
      "theoretical_loss": 3.9349891968241963,
      "tokens_seen": 476839936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043205616850551656,
      "loss": 1.6091,
      "theoretical_loss": 3.934871761347945,
      "tokens_seen": 476971008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320361083249749,
      "loss": 1.3174,
      "theoretical_loss": 3.93475436717177,
      "tokens_seen": 477102080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320160481444333,
      "loss": 1.4658,
      "theoretical_loss": 3.934637014269808,
      "tokens_seen": 477233152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043199598796389165,
      "loss": 1.2387,
      "theoretical_loss": 3.9345197026162158,
      "tokens_seen": 477364224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043197592778335006,
      "loss": 1.3269,
      "theoretical_loss": 3.934402432185176,
      "tokens_seen": 477495296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319558676028084,
      "loss": 1.6676,
      "theoretical_loss": 3.9342852029508935,
      "tokens_seen": 477626368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319358074222668,
      "loss": 1.4119,
      "theoretical_loss": 3.934168014887597,
      "tokens_seen": 477757440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043191574724172515,
      "loss": 1.464,
      "theoretical_loss": 3.934050867969537,
      "tokens_seen": 477888512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043189568706118357,
      "loss": 1.5136,
      "theoretical_loss": 3.93393376217099,
      "tokens_seen": 478019584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043187562688064193,
      "loss": 1.5275,
      "theoretical_loss": 3.933816697466252,
      "tokens_seen": 478150656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318555667001003,
      "loss": 1.6569,
      "theoretical_loss": 3.9336996738296444,
      "tokens_seen": 478281728
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4156891405582428,
      "objective/train/avg_token_score": 0.021102596074342728,
      "objective/train/docs_used": 281924,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5788319110870361,
      "objective/train/original_loss": 4.477541446685791,
      "objective/train/theoretical_loss": 3.9335826912355114,
      "objective/train/tokens_used": 478412800,
      "objective/train/unlikelihood_loss": 0.12658530473709106,
      "theoretical_loss": 3.9335826912355114,
      "tokens_seen": 478412800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043183550651955865,
      "loss": 1.3921,
      "theoretical_loss": 3.9335826912355114,
      "tokens_seen": 478412800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000431815446339017,
      "loss": 1.3283,
      "theoretical_loss": 3.9334657496582195,
      "tokens_seen": 478543872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043179538615847543,
      "loss": 1.3853,
      "theoretical_loss": 3.9333488490721598,
      "tokens_seen": 478674944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317753259779338,
      "loss": 1.4112,
      "theoretical_loss": 3.933231989451744,
      "tokens_seen": 478806016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043175526579739215,
      "loss": 1.4039,
      "theoretical_loss": 3.933115170771409,
      "tokens_seen": 478937088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317352056168505,
      "loss": 1.37,
      "theoretical_loss": 3.932998393005613,
      "tokens_seen": 479068160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043171514543630893,
      "loss": 1.3066,
      "theoretical_loss": 3.9328816561288384,
      "tokens_seen": 479199232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316950852557673,
      "loss": 1.4299,
      "theoretical_loss": 3.93276496011559,
      "tokens_seen": 479330304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316750250752257,
      "loss": 1.3555,
      "theoretical_loss": 3.932648304940395,
      "tokens_seen": 479461376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316549648946841,
      "loss": 1.2894,
      "theoretical_loss": 3.9325316905778047,
      "tokens_seen": 479592448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043163490471414244,
      "loss": 1.3105,
      "theoretical_loss": 3.932415117002391,
      "tokens_seen": 479723520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043161484453360085,
      "loss": 1.4595,
      "theoretical_loss": 3.932298584188751,
      "tokens_seen": 479854592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315947843530592,
      "loss": 1.4289,
      "theoretical_loss": 3.9321820921115034,
      "tokens_seen": 479985664
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.31689882278442383,
      "objective/train/avg_token_score": 0.03435559570789337,
      "objective/train/docs_used": 283309,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2388397455215454,
      "objective/train/original_loss": 4.061135768890381,
      "objective/train/theoretical_loss": 3.932123861341101,
      "objective/train/tokens_used": 480051200,
      "objective/train/unlikelihood_loss": 0.1754794418811798,
      "theoretical_loss": 3.932123861341101,
      "tokens_seen": 480051200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315747241725176,
      "loss": 1.4026,
      "theoretical_loss": 3.932065640745289,
      "tokens_seen": 480116736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043155466399197594,
      "loss": 1.4035,
      "theoretical_loss": 3.9319492300647725,
      "tokens_seen": 480247808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043153460381143436,
      "loss": 1.4297,
      "theoretical_loss": 3.9318328600446413,
      "tokens_seen": 480378880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315145436308927,
      "loss": 1.2719,
      "theoretical_loss": 3.931716530659604,
      "tokens_seen": 480509952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314944834503511,
      "loss": 1.4338,
      "theoretical_loss": 3.9316002418843934,
      "tokens_seen": 480641024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043147442326980944,
      "loss": 1.4461,
      "theoretical_loss": 3.931483993693764,
      "tokens_seen": 480772096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314543630892678,
      "loss": 1.3509,
      "theoretical_loss": 3.9313677860624936,
      "tokens_seen": 480903168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314343029087262,
      "loss": 1.3157,
      "theoretical_loss": 3.9312516189653817,
      "tokens_seen": 481034240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314142427281846,
      "loss": 1.5189,
      "theoretical_loss": 3.9311354923772504,
      "tokens_seen": 481165312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043139418254764295,
      "loss": 1.4483,
      "theoretical_loss": 3.931019406272945,
      "tokens_seen": 481296384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313741223671013,
      "loss": 1.3292,
      "theoretical_loss": 3.9309033606273327,
      "tokens_seen": 481427456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313540621865597,
      "loss": 1.3071,
      "theoretical_loss": 3.930787355415304,
      "tokens_seen": 481558528
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.4172165095806122,
      "objective/train/avg_token_score": 0.026619194075465202,
      "objective/train/docs_used": 283520,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5104230642318726,
      "objective/train/original_loss": 3.943290948867798,
      "objective/train/theoretical_loss": 3.93067139061177,
      "objective/train/tokens_used": 481689600,
      "objective/train/unlikelihood_loss": 0.14345265924930573,
      "theoretical_loss": 3.93067139061177,
      "tokens_seen": 481689600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313340020060181,
      "loss": 1.6205,
      "theoretical_loss": 3.93067139061177,
      "tokens_seen": 481689600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043131394182547645,
      "loss": 1.3794,
      "theoretical_loss": 3.9305554661916657,
      "tokens_seen": 481820672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312938816449348,
      "loss": 1.563,
      "theoretical_loss": 3.9304395821299485,
      "tokens_seen": 481951744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043127382146439323,
      "loss": 1.543,
      "theoretical_loss": 3.930323738401597,
      "tokens_seen": 482082816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312537612838516,
      "loss": 1.3568,
      "theoretical_loss": 3.930207934981614,
      "tokens_seen": 482213888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043123370110330995,
      "loss": 1.5519,
      "theoretical_loss": 3.930092171845022,
      "tokens_seen": 482344960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312136409227683,
      "loss": 1.4007,
      "theoretical_loss": 3.929976448966868,
      "tokens_seen": 482476032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311935807422267,
      "loss": 1.3958,
      "theoretical_loss": 3.92986076632222,
      "tokens_seen": 482607104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311735205616851,
      "loss": 1.3824,
      "theoretical_loss": 3.9297451238861685,
      "tokens_seen": 482738176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043115346038114346,
      "loss": 1.4257,
      "theoretical_loss": 3.929629521633827,
      "tokens_seen": 482869248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311334002006018,
      "loss": 1.5184,
      "theoretical_loss": 3.92951395954033,
      "tokens_seen": 483000320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311133400200602,
      "loss": 1.402,
      "theoretical_loss": 3.9293984375808337,
      "tokens_seen": 483131392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310932798395186,
      "loss": 1.5764,
      "theoretical_loss": 3.929282955730519,
      "tokens_seen": 483262464
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.4565615952014923,
      "objective/train/avg_token_score": 0.03555590286850929,
      "objective/train/docs_used": 284734,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5730347633361816,
      "objective/train/original_loss": 4.092532157897949,
      "objective/train/theoretical_loss": 3.9292252298385533,
      "objective/train/tokens_used": 483328000,
      "objective/train/unlikelihood_loss": 0.2168586552143097,
      "theoretical_loss": 3.9292252298385533,
      "tokens_seen": 483328000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043107321965897696,
      "loss": 1.3438,
      "theoretical_loss": 3.929167513964585,
      "tokens_seen": 483393536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310531594784353,
      "loss": 1.3901,
      "theoretical_loss": 3.9290521122582573,
      "tokens_seen": 483524608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310330992978937,
      "loss": 1.3786,
      "theoretical_loss": 3.928936750586779,
      "tokens_seen": 483655680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043101303911735205,
      "loss": 1.4429,
      "theoretical_loss": 3.9288214289254193,
      "tokens_seen": 483786752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043099297893681046,
      "loss": 1.2574,
      "theoretical_loss": 3.928706147249466,
      "tokens_seen": 483917824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309729187562688,
      "loss": 1.2796,
      "theoretical_loss": 3.9285909055342314,
      "tokens_seen": 484048896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309528585757272,
      "loss": 1.4361,
      "theoretical_loss": 3.928475703755048,
      "tokens_seen": 484179968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043093279839518555,
      "loss": 1.5195,
      "theoretical_loss": 3.9283605418872716,
      "tokens_seen": 484311040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043091273821464397,
      "loss": 1.5687,
      "theoretical_loss": 3.9282454199062777,
      "tokens_seen": 484442112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043089267803410233,
      "loss": 1.5128,
      "theoretical_loss": 3.928130337787467,
      "tokens_seen": 484573184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308726178535607,
      "loss": 1.4754,
      "theoretical_loss": 3.9280152955062584,
      "tokens_seen": 484704256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043085255767301905,
      "loss": 1.3141,
      "theoretical_loss": 3.9279002930380953,
      "tokens_seen": 484835328
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.4302908182144165,
      "objective/train/avg_token_score": 0.014568953774869442,
      "objective/train/docs_used": 285372,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4252182245254517,
      "objective/train/original_loss": 3.9690096378326416,
      "objective/train/theoretical_loss": 3.927785330358441,
      "objective/train/tokens_used": 484966400,
      "objective/train/unlikelihood_loss": 0.10961988568305969,
      "theoretical_loss": 3.927785330358441,
      "tokens_seen": 484966400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308324974924774,
      "loss": 1.2884,
      "theoretical_loss": 3.927785330358441,
      "tokens_seen": 484966400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043081243731193583,
      "loss": 1.5775,
      "theoretical_loss": 3.9276704074427826,
      "tokens_seen": 485097472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307923771313942,
      "loss": 1.4638,
      "theoretical_loss": 3.9275555242666265,
      "tokens_seen": 485228544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043077231695085256,
      "loss": 1.4386,
      "theoretical_loss": 3.927440680805503,
      "tokens_seen": 485359616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307522567703109,
      "loss": 1.4193,
      "theoretical_loss": 3.9273258770349626,
      "tokens_seen": 485490688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043073219658976933,
      "loss": 1.6056,
      "theoretical_loss": 3.9272111129305785,
      "tokens_seen": 485621760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307121364092277,
      "loss": 1.4051,
      "theoretical_loss": 3.9270963884679437,
      "tokens_seen": 485752832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043069207622868606,
      "loss": 1.2638,
      "theoretical_loss": 3.9269817036226757,
      "tokens_seen": 485883904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306720160481444,
      "loss": 1.3818,
      "theoretical_loss": 3.9268670583704113,
      "tokens_seen": 486014976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306519558676028,
      "loss": 1.2313,
      "theoretical_loss": 3.9267524526868085,
      "tokens_seen": 486146048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306318956870612,
      "loss": 1.4043,
      "theoretical_loss": 3.9266378865475495,
      "tokens_seen": 486277120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043061183550651956,
      "loss": 1.3174,
      "theoretical_loss": 3.926523359928335,
      "tokens_seen": 486408192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305917753259779,
      "loss": 1.4299,
      "theoretical_loss": 3.926408872804889,
      "tokens_seen": 486539264
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.4375610947608948,
      "objective/train/avg_token_score": 0.00509691471233964,
      "objective/train/docs_used": 285999,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.344993233680725,
      "objective/train/original_loss": 3.588568925857544,
      "objective/train/theoretical_loss": 3.9263516440464987,
      "objective/train/tokens_used": 486604800,
      "objective/train/unlikelihood_loss": 0.1715584695339203,
      "theoretical_loss": 3.9263516440464987,
      "tokens_seen": 486604800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305717151454363,
      "loss": 1.468,
      "theoretical_loss": 3.926294425152957,
      "tokens_seen": 486670336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305516549648947,
      "loss": 1.5126,
      "theoretical_loss": 3.9261800169483037,
      "tokens_seen": 486801408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043053159478435306,
      "loss": 1.3765,
      "theoretical_loss": 3.926065648166719,
      "tokens_seen": 486932480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305115346038114,
      "loss": 1.6243,
      "theoretical_loss": 3.9259513187840103,
      "tokens_seen": 487063552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304914744232698,
      "loss": 1.4171,
      "theoretical_loss": 3.9258370287760087,
      "tokens_seen": 487194624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043047141424272815,
      "loss": 1.348,
      "theoretical_loss": 3.9257227781185655,
      "tokens_seen": 487325696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043045135406218657,
      "loss": 1.4053,
      "theoretical_loss": 3.925608566787555,
      "tokens_seen": 487456768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043043129388164493,
      "loss": 1.3852,
      "theoretical_loss": 3.9254943947588705,
      "tokens_seen": 487587840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304112337011033,
      "loss": 1.3728,
      "theoretical_loss": 3.925380262008428,
      "tokens_seen": 487718912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043039117352056165,
      "loss": 1.1108,
      "theoretical_loss": 3.925266168512165,
      "tokens_seen": 487849984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043037111334002007,
      "loss": 1.584,
      "theoretical_loss": 3.9251521142460377,
      "tokens_seen": 487981056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043035105315947843,
      "loss": 1.4063,
      "theoretical_loss": 3.925038099186027,
      "tokens_seen": 488112128
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.26423507928848267,
      "objective/train/avg_token_score": 0.009158661589026451,
      "objective/train/docs_used": 287428,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9026793837547302,
      "objective/train/original_loss": 3.2273669242858887,
      "objective/train/theoretical_loss": 3.9249241233081333,
      "objective/train/tokens_used": 488243200,
      "objective/train/unlikelihood_loss": 0.31226837635040283,
      "theoretical_loss": 3.9249241233081333,
      "tokens_seen": 488243200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303309929789368,
      "loss": 1.4517,
      "theoretical_loss": 3.9249241233081333,
      "tokens_seen": 488243200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043031093279839516,
      "loss": 1.5534,
      "theoretical_loss": 3.9248101865883767,
      "tokens_seen": 488374272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302908726178535,
      "loss": 1.4797,
      "theoretical_loss": 3.924696289002801,
      "tokens_seen": 488505344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043027081243731194,
      "loss": 1.4941,
      "theoretical_loss": 3.9245824305274697,
      "tokens_seen": 488636416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302507522567703,
      "loss": 1.4144,
      "theoretical_loss": 3.924468611138467,
      "tokens_seen": 488767488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043023069207622866,
      "loss": 1.4611,
      "theoretical_loss": 3.9243548308118994,
      "tokens_seen": 488898560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430210631895687,
      "loss": 1.4122,
      "theoretical_loss": 3.9242410895238935,
      "tokens_seen": 489029632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043019057171514544,
      "loss": 1.3031,
      "theoretical_loss": 3.9241273872505973,
      "tokens_seen": 489160704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043017051153460386,
      "loss": 1.4607,
      "theoretical_loss": 3.9240137239681787,
      "tokens_seen": 489291776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301504513540622,
      "loss": 1.5392,
      "theoretical_loss": 3.923900099652828,
      "tokens_seen": 489422848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301303911735206,
      "loss": 1.5051,
      "theoretical_loss": 3.9237865142807555,
      "tokens_seen": 489553920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430110330992979,
      "loss": 1.6593,
      "theoretical_loss": 3.9236729678281925,
      "tokens_seen": 489684992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043009027081243736,
      "loss": 1.5331,
      "theoretical_loss": 3.9235594602713926,
      "tokens_seen": 489816064
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.5066593289375305,
      "objective/train/avg_token_score": 0.013708590529859066,
      "objective/train/docs_used": 288116,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7811899185180664,
      "objective/train/original_loss": 3.8481087684631348,
      "objective/train/theoretical_loss": 3.923502721071488,
      "objective/train/tokens_used": 489881600,
      "objective/train/unlikelihood_loss": 0.17132464051246643,
      "theoretical_loss": 3.923502721071488,
      "tokens_seen": 489881600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300702106318957,
      "loss": 1.5335,
      "theoretical_loss": 3.9234459915866275,
      "tokens_seen": 489947136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300501504513541,
      "loss": 1.4411,
      "theoretical_loss": 3.923332561750192,
      "tokens_seen": 490078208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043003009027081245,
      "loss": 1.466,
      "theoretical_loss": 3.9232191707384,
      "tokens_seen": 490209280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043001003009027086,
      "loss": 1.3247,
      "theoretical_loss": 3.9231058185275884,
      "tokens_seen": 490340352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299899699097292,
      "loss": 1.5211,
      "theoretical_loss": 3.922992505094112,
      "tokens_seen": 490471424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299699097291876,
      "loss": 1.2627,
      "theoretical_loss": 3.922879230414349,
      "tokens_seen": 490602496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042994984954864595,
      "loss": 1.5509,
      "theoretical_loss": 3.9227659944646964,
      "tokens_seen": 490733568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042992978936810437,
      "loss": 1.504,
      "theoretical_loss": 3.922652797221573,
      "tokens_seen": 490864640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042990972918756273,
      "loss": 1.395,
      "theoretical_loss": 3.922539638661418,
      "tokens_seen": 490995712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298896690070211,
      "loss": 1.5169,
      "theoretical_loss": 3.92242651876069,
      "tokens_seen": 491126784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042986960882647945,
      "loss": 1.8001,
      "theoretical_loss": 3.922313437495871,
      "tokens_seen": 491257856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298495486459378,
      "loss": 1.3611,
      "theoretical_loss": 3.9222003948434603,
      "tokens_seen": 491388928
    },
    {
      "debugging/length": 516.8571428571429,
      "debugging/num_segments": 7,
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.45747801661491394,
      "objective/train/avg_token_score": 0.004869875498116016,
      "objective/train/docs_used": 289197,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5986454486846924,
      "objective/train/original_loss": 4.020472526550293,
      "objective/train/theoretical_loss": 3.92208739077998,
      "objective/train/tokens_used": 491520000,
      "objective/train/unlikelihood_loss": 0.21023108065128326,
      "theoretical_loss": 3.92208739077998,
      "tokens_seen": 491520000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042982948846539623,
      "loss": 1.4155,
      "theoretical_loss": 3.92208739077998,
      "tokens_seen": 491520000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298094282848546,
      "loss": 1.5174,
      "theoretical_loss": 3.921974425281972,
      "tokens_seen": 491651072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042978936810431296,
      "loss": 1.3887,
      "theoretical_loss": 3.921861498325998,
      "tokens_seen": 491782144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297693079237713,
      "loss": 1.4294,
      "theoretical_loss": 3.9217486098886427,
      "tokens_seen": 491913216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042974924774322973,
      "loss": 1.6334,
      "theoretical_loss": 3.9216357599465073,
      "tokens_seen": 492044288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297291875626881,
      "loss": 1.2926,
      "theoretical_loss": 3.921522948476217,
      "tokens_seen": 492175360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042970912738214646,
      "loss": 1.4125,
      "theoretical_loss": 3.921410175454416,
      "tokens_seen": 492306432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296890672016048,
      "loss": 1.4243,
      "theoretical_loss": 3.9212974408577685,
      "tokens_seen": 492437504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296690070210632,
      "loss": 1.3942,
      "theoretical_loss": 3.92118474466296,
      "tokens_seen": 492568576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296489468405216,
      "loss": 1.4767,
      "theoretical_loss": 3.9210720868466953,
      "tokens_seen": 492699648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042962888665997996,
      "loss": 1.5349,
      "theoretical_loss": 3.9209594673857,
      "tokens_seen": 492830720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296088264794383,
      "loss": 1.4584,
      "theoretical_loss": 3.9208468862567205,
      "tokens_seen": 492961792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295887662988967,
      "loss": 1.2407,
      "theoretical_loss": 3.920734343436523,
      "tokens_seen": 493092864
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.41605570912361145,
      "objective/train/avg_token_score": 0.049754299223423004,
      "objective/train/docs_used": 289918,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4484484195709229,
      "objective/train/original_loss": 4.115923881530762,
      "objective/train/theoretical_loss": 3.920678086384963,
      "objective/train/tokens_used": 493158400,
      "objective/train/unlikelihood_loss": 0.13206537067890167,
      "theoretical_loss": 3.920678086384963,
      "tokens_seen": 493158400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295687061183551,
      "loss": 1.3909,
      "theoretical_loss": 3.9206218389018943,
      "tokens_seen": 493223936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042954864593781347,
      "loss": 1.6181,
      "theoretical_loss": 3.9205093726296405,
      "tokens_seen": 493355008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042952858575727183,
      "loss": 1.4534,
      "theoretical_loss": 3.9203969445965887,
      "tokens_seen": 493486080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295085255767302,
      "loss": 1.4879,
      "theoretical_loss": 3.9202845547795864,
      "tokens_seen": 493617152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042948846539618855,
      "loss": 1.5724,
      "theoretical_loss": 3.9201722031555004,
      "tokens_seen": 493748224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042946840521564697,
      "loss": 1.5877,
      "theoretical_loss": 3.9200598897012187,
      "tokens_seen": 493879296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042944834503510533,
      "loss": 1.6108,
      "theoretical_loss": 3.9199476143936485,
      "tokens_seen": 494010368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294282848545637,
      "loss": 1.4842,
      "theoretical_loss": 3.9198353772097168,
      "tokens_seen": 494141440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042940822467402206,
      "loss": 1.3961,
      "theoretical_loss": 3.9197231781263726,
      "tokens_seen": 494272512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042938816449348047,
      "loss": 1.4718,
      "theoretical_loss": 3.919611017120583,
      "tokens_seen": 494403584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042936810431293883,
      "loss": 1.172,
      "theoretical_loss": 3.9194988941693354,
      "tokens_seen": 494534656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293480441323972,
      "loss": 1.5461,
      "theoretical_loss": 3.9193868092496382,
      "tokens_seen": 494665728
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.38336998224258423,
      "objective/train/avg_token_score": 0.007834085263311863,
      "objective/train/docs_used": 291263,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3380773067474365,
      "objective/train/original_loss": 3.855975866317749,
      "objective/train/theoretical_loss": 3.919274762338519,
      "objective/train/tokens_used": 494796800,
      "objective/train/unlikelihood_loss": 0.20109692215919495,
      "theoretical_loss": 3.919274762338519,
      "tokens_seen": 494796800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042932798395185556,
      "loss": 1.6039,
      "theoretical_loss": 3.919274762338519,
      "tokens_seen": 494796800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293079237713139,
      "loss": 1.2938,
      "theoretical_loss": 3.9191627534130253,
      "tokens_seen": 494927872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042928786359077234,
      "loss": 1.5174,
      "theoretical_loss": 3.919050782450225,
      "tokens_seen": 495058944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292678034102307,
      "loss": 1.3587,
      "theoretical_loss": 3.9189388494272057,
      "tokens_seen": 495190016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042924774322968906,
      "loss": 1.4829,
      "theoretical_loss": 3.918826954321074,
      "tokens_seen": 495321088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292276830491474,
      "loss": 1.5397,
      "theoretical_loss": 3.9187150971089584,
      "tokens_seen": 495452160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042920762286860584,
      "loss": 1.445,
      "theoretical_loss": 3.9186032777680055,
      "tokens_seen": 495583232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291875626880642,
      "loss": 1.6188,
      "theoretical_loss": 3.918491496275383,
      "tokens_seen": 495714304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042916750250752256,
      "loss": 1.4414,
      "theoretical_loss": 3.918379752608276,
      "tokens_seen": 495845376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291474423269809,
      "loss": 1.2727,
      "theoretical_loss": 3.9182680467438926,
      "tokens_seen": 495976448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291273821464393,
      "loss": 1.3956,
      "theoretical_loss": 3.918156378659459,
      "tokens_seen": 496107520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291073219658977,
      "loss": 1.4127,
      "theoretical_loss": 3.9180447483322203,
      "tokens_seen": 496238592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042908726178535607,
      "loss": 1.3812,
      "theoretical_loss": 3.9179331557394432,
      "tokens_seen": 496369664
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.36589688062667847,
      "objective/train/avg_token_score": 0.019114283844828606,
      "objective/train/docs_used": 291574,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3350896835327148,
      "objective/train/original_loss": 4.14976167678833,
      "objective/train/theoretical_loss": 3.9178773735863786,
      "objective/train/tokens_used": 496435200,
      "objective/train/unlikelihood_loss": 0.12984949350357056,
      "theoretical_loss": 3.9178773735863786,
      "tokens_seen": 496435200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042906720160481443,
      "loss": 1.3808,
      "theoretical_loss": 3.917821600858413,
      "tokens_seen": 496500736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290471414242728,
      "loss": 1.5891,
      "theoretical_loss": 3.917710083666434,
      "tokens_seen": 496631808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290270812437312,
      "loss": 1.279,
      "theoretical_loss": 3.9175986041408315,
      "tokens_seen": 496762880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042900702106318957,
      "loss": 1.3422,
      "theoretical_loss": 3.9174871622589498,
      "tokens_seen": 496893952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042898696088264793,
      "loss": 1.4238,
      "theoretical_loss": 3.9173757579981534,
      "tokens_seen": 497025024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289669007021063,
      "loss": 1.4951,
      "theoretical_loss": 3.9172643913358254,
      "tokens_seen": 497156096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289468405215647,
      "loss": 1.4327,
      "theoretical_loss": 3.917153062249368,
      "tokens_seen": 497287168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289267803410231,
      "loss": 1.3294,
      "theoretical_loss": 3.9170417707162057,
      "tokens_seen": 497418240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042890672016048144,
      "loss": 1.5096,
      "theoretical_loss": 3.9169305167137796,
      "tokens_seen": 497549312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288866599799398,
      "loss": 1.587,
      "theoretical_loss": 3.9168193002195513,
      "tokens_seen": 497680384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042886659979939816,
      "loss": 1.3454,
      "theoretical_loss": 3.916708121211002,
      "tokens_seen": 497811456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288465396188566,
      "loss": 1.4182,
      "theoretical_loss": 3.916596979665632,
      "tokens_seen": 497942528
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.35257819294929504,
      "objective/train/avg_token_score": 0.013382690958678722,
      "objective/train/docs_used": 293024,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3515236377716064,
      "objective/train/original_loss": 3.946211576461792,
      "objective/train/theoretical_loss": 3.9164858755609613,
      "objective/train/tokens_used": 498073600,
      "objective/train/unlikelihood_loss": 0.19796288013458252,
      "theoretical_loss": 3.9164858755609613,
      "tokens_seen": 498073600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042882647943831494,
      "loss": 1.4908,
      "theoretical_loss": 3.9164858755609613,
      "tokens_seen": 498073600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288064192577733,
      "loss": 1.438,
      "theoretical_loss": 3.9163748088745303,
      "tokens_seen": 498204672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042878635907723166,
      "loss": 1.3417,
      "theoretical_loss": 3.916263779583896,
      "tokens_seen": 498335744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287662988966901,
      "loss": 1.5937,
      "theoretical_loss": 3.916152787666638,
      "tokens_seen": 498466816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042874623871614844,
      "loss": 1.4663,
      "theoretical_loss": 3.9160418331003526,
      "tokens_seen": 498597888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287261785356068,
      "loss": 1.4601,
      "theoretical_loss": 3.9159309158626576,
      "tokens_seen": 498728960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042870611835506517,
      "loss": 1.2813,
      "theoretical_loss": 3.915820035931188,
      "tokens_seen": 498860032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042868605817452353,
      "loss": 1.2875,
      "theoretical_loss": 3.9157091932836,
      "tokens_seen": 498991104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428665997993982,
      "loss": 1.5178,
      "theoretical_loss": 3.9155983878975675,
      "tokens_seen": 499122176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042864593781344036,
      "loss": 1.5188,
      "theoretical_loss": 3.9154876197507846,
      "tokens_seen": 499253248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286258776328987,
      "loss": 1.5098,
      "theoretical_loss": 3.9153768888209646,
      "tokens_seen": 499384320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286058174523571,
      "loss": 1.3121,
      "theoretical_loss": 3.915266195085839,
      "tokens_seen": 499515392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285857572718155,
      "loss": 1.5902,
      "theoretical_loss": 3.91515553852316,
      "tokens_seen": 499646464
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.4228983521461487,
      "objective/train/avg_token_score": 0.019886652007699013,
      "objective/train/docs_used": 293514,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.436700701713562,
      "objective/train/original_loss": 3.5851337909698486,
      "objective/train/theoretical_loss": 3.91510022417454,
      "objective/train/tokens_used": 499712000,
      "objective/train/unlikelihood_loss": 0.24415582418441772,
      "theoretical_loss": 3.91510022417454,
      "tokens_seen": 499712000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042856569709127387,
      "loss": 1.4769,
      "theoretical_loss": 3.915044919110697,
      "tokens_seen": 499777536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042854563691073223,
      "loss": 1.4117,
      "theoretical_loss": 3.9149343368262404,
      "tokens_seen": 499908608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285255767301906,
      "loss": 1.4151,
      "theoretical_loss": 3.9148237916475983,
      "tokens_seen": 500039680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042850551654964895,
      "loss": 1.4644,
      "theoretical_loss": 3.914713283552599,
      "tokens_seen": 500170752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042848545636910737,
      "loss": 1.4707,
      "theoretical_loss": 3.914602812519089,
      "tokens_seen": 500301824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042846539618856573,
      "loss": 1.3893,
      "theoretical_loss": 3.914492378524934,
      "tokens_seen": 500432896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284453360080241,
      "loss": 1.3221,
      "theoretical_loss": 3.9143819815480194,
      "tokens_seen": 500563968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042842527582748246,
      "loss": 1.4009,
      "theoretical_loss": 3.9142716215662494,
      "tokens_seen": 500695040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042840521564694087,
      "loss": 1.4476,
      "theoretical_loss": 3.9141612985575454,
      "tokens_seen": 500826112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042838515546639923,
      "loss": 1.3708,
      "theoretical_loss": 3.91405101249985,
      "tokens_seen": 500957184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283650952858576,
      "loss": 1.5202,
      "theoretical_loss": 3.9139407633711243,
      "tokens_seen": 501088256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042834503510531596,
      "loss": 1.393,
      "theoretical_loss": 3.9138305511493474,
      "tokens_seen": 501219328
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.34488025307655334,
      "objective/train/avg_token_score": 0.03448681905865669,
      "objective/train/docs_used": 294722,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1151436567306519,
      "objective/train/original_loss": 3.941049337387085,
      "objective/train/theoretical_loss": 3.9137203758125176,
      "objective/train/tokens_used": 501350400,
      "objective/train/unlikelihood_loss": 0.2037958949804306,
      "theoretical_loss": 3.9137203758125176,
      "tokens_seen": 501350400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283249749247743,
      "loss": 1.4004,
      "theoretical_loss": 3.9137203758125176,
      "tokens_seen": 501350400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042830491474423274,
      "loss": 1.2935,
      "theoretical_loss": 3.913610237338653,
      "tokens_seen": 501481472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282848545636911,
      "loss": 1.2546,
      "theoretical_loss": 3.913500135705789,
      "tokens_seen": 501612544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042826479438314946,
      "loss": 1.5618,
      "theoretical_loss": 3.913390070891981,
      "tokens_seen": 501743616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282447342026078,
      "loss": 1.3867,
      "theoretical_loss": 3.9132800428753027,
      "tokens_seen": 501874688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042822467402206624,
      "loss": 1.2633,
      "theoretical_loss": 3.9131700516338475,
      "tokens_seen": 502005760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282046138415246,
      "loss": 1.4745,
      "theoretical_loss": 3.913060097145725,
      "tokens_seen": 502136832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042818455366098296,
      "loss": 1.46,
      "theoretical_loss": 3.9129501793890666,
      "tokens_seen": 502267904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281644934804413,
      "loss": 1.2897,
      "theoretical_loss": 3.9128402983420214,
      "tokens_seen": 502398976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281444332998997,
      "loss": 1.2421,
      "theoretical_loss": 3.912730453982756,
      "tokens_seen": 502530048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281243731193581,
      "loss": 1.3616,
      "theoretical_loss": 3.912620646289457,
      "tokens_seen": 502661120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042810431293881647,
      "loss": 1.5323,
      "theoretical_loss": 3.9125108752403293,
      "tokens_seen": 502792192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042808425275827483,
      "loss": 1.2956,
      "theoretical_loss": 3.9124011408135964,
      "tokens_seen": 502923264
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.4918743968009949,
      "objective/train/avg_token_score": 0.003718223888427019,
      "objective/train/docs_used": 295112,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6487371921539307,
      "objective/train/original_loss": 4.021548748016357,
      "objective/train/theoretical_loss": 3.9123462873268275,
      "objective/train/tokens_used": 502988800,
      "objective/train/unlikelihood_loss": 0.127858504652977,
      "theoretical_loss": 3.9123462873268275,
      "tokens_seen": 502988800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280641925777332,
      "loss": 1.3411,
      "theoretical_loss": 3.9122914429874998,
      "tokens_seen": 503054336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280441323971916,
      "loss": 1.5446,
      "theoretical_loss": 3.9121817817403004,
      "tokens_seen": 503185408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042802407221664997,
      "loss": 1.2649,
      "theoretical_loss": 3.912072157050278,
      "tokens_seen": 503316480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042800401203610833,
      "loss": 1.2767,
      "theoretical_loss": 3.91196256889573,
      "tokens_seen": 503447552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279839518555667,
      "loss": 1.2984,
      "theoretical_loss": 3.911853017254973,
      "tokens_seen": 503578624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279638916750251,
      "loss": 1.3402,
      "theoretical_loss": 3.911743502106341,
      "tokens_seen": 503709696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279438314944835,
      "loss": 1.4761,
      "theoretical_loss": 3.9116340234281886,
      "tokens_seen": 503840768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042792377131394184,
      "loss": 1.4801,
      "theoretical_loss": 3.9115245811988864,
      "tokens_seen": 503971840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279037111334002,
      "loss": 1.4165,
      "theoretical_loss": 3.911415175396825,
      "tokens_seen": 504102912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042788365095285856,
      "loss": 1.4903,
      "theoretical_loss": 3.9113058060004136,
      "tokens_seen": 504233984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427863590772317,
      "loss": 1.3463,
      "theoretical_loss": 3.911196472988079,
      "tokens_seen": 504365056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042784353059177534,
      "loss": 1.4082,
      "theoretical_loss": 3.911087176338266,
      "tokens_seen": 504496128
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.2879398763179779,
      "objective/train/avg_token_score": 0.030777588486671448,
      "objective/train/docs_used": 296352,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.946451723575592,
      "objective/train/original_loss": 3.832697629928589,
      "objective/train/theoretical_loss": 3.910977916029439,
      "objective/train/tokens_used": 504627200,
      "objective/train/unlikelihood_loss": 0.18606476485729218,
      "theoretical_loss": 3.910977916029439,
      "tokens_seen": 504627200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278234704112337,
      "loss": 1.31,
      "theoretical_loss": 3.910977916029439,
      "tokens_seen": 504627200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042780341023069206,
      "loss": 1.3279,
      "theoretical_loss": 3.91086869204008,
      "tokens_seen": 504758272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277833500501505,
      "loss": 1.4205,
      "theoretical_loss": 3.9107595043486905,
      "tokens_seen": 504889344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042776328986960884,
      "loss": 1.5258,
      "theoretical_loss": 3.9106503529337875,
      "tokens_seen": 505020416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277432296890672,
      "loss": 1.5069,
      "theoretical_loss": 3.910541237773909,
      "tokens_seen": 505151488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042772316950852557,
      "loss": 1.5058,
      "theoretical_loss": 3.9104321588476108,
      "tokens_seen": 505282560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042770310932798393,
      "loss": 1.4835,
      "theoretical_loss": 3.9103231161334655,
      "tokens_seen": 505413632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042768304914744235,
      "loss": 1.3955,
      "theoretical_loss": 3.910214109610065,
      "tokens_seen": 505544704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276629889669007,
      "loss": 1.4016,
      "theoretical_loss": 3.9101051392560198,
      "tokens_seen": 505675776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042764292878635907,
      "loss": 1.4402,
      "theoretical_loss": 3.9099962050499584,
      "tokens_seen": 505806848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042762286860581743,
      "loss": 1.4021,
      "theoretical_loss": 3.9098873069705267,
      "tokens_seen": 505937920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042760280842527585,
      "loss": 1.4597,
      "theoretical_loss": 3.909778444996389,
      "tokens_seen": 506068992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275827482447342,
      "loss": 1.469,
      "theoretical_loss": 3.9096696191062277,
      "tokens_seen": 506200064
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.37237292528152466,
      "objective/train/avg_token_score": 0.01380022894591093,
      "objective/train/docs_used": 297118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3744415044784546,
      "objective/train/original_loss": 4.307300090789795,
      "objective/train/theoretical_loss": 3.909615219685982,
      "objective/train/tokens_used": 506265600,
      "objective/train/unlikelihood_loss": 0.17518089711666107,
      "theoretical_loss": 3.909615219685982,
      "tokens_seen": 506265600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275626880641926,
      "loss": 1.4586,
      "theoretical_loss": 3.9095608292787443,
      "tokens_seen": 506331136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042754262788365094,
      "loss": 1.4717,
      "theoretical_loss": 3.9094520754926574,
      "tokens_seen": 506462208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275225677031093,
      "loss": 1.4785,
      "theoretical_loss": 3.909343357726703,
      "tokens_seen": 506593280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275025075225677,
      "loss": 1.4522,
      "theoretical_loss": 3.9092346759596373,
      "tokens_seen": 506724352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274824473420261,
      "loss": 1.4267,
      "theoretical_loss": 3.9091260301702326,
      "tokens_seen": 506855424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042746238716148444,
      "loss": 1.6183,
      "theoretical_loss": 3.90901742033728,
      "tokens_seen": 506986496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274423269809428,
      "loss": 1.4882,
      "theoretical_loss": 3.9089088464395876,
      "tokens_seen": 507117568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274222668004012,
      "loss": 1.3987,
      "theoretical_loss": 3.9088003084559833,
      "tokens_seen": 507248640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274022066198596,
      "loss": 1.6277,
      "theoretical_loss": 3.9086918063653115,
      "tokens_seen": 507379712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042738214643931794,
      "loss": 1.5008,
      "theoretical_loss": 3.9085833401464347,
      "tokens_seen": 507510784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273620862587763,
      "loss": 1.4856,
      "theoretical_loss": 3.908474909778234,
      "tokens_seen": 507641856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042734202607823467,
      "loss": 1.3557,
      "theoretical_loss": 3.9083665152396074,
      "tokens_seen": 507772928
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.35135629773139954,
      "objective/train/avg_token_score": 0.02974272146821022,
      "objective/train/docs_used": 298304,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.351550579071045,
      "objective/train/original_loss": 4.083728790283203,
      "objective/train/theoretical_loss": 3.908258156509472,
      "objective/train/tokens_used": 507904000,
      "objective/train/unlikelihood_loss": 0.15935710072517395,
      "theoretical_loss": 3.908258156509472,
      "tokens_seen": 507904000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273219658976931,
      "loss": 1.406,
      "theoretical_loss": 3.908258156509472,
      "tokens_seen": 507904000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042730190571715145,
      "loss": 1.5871,
      "theoretical_loss": 3.9081498335667613,
      "tokens_seen": 508035072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272818455366098,
      "loss": 1.2993,
      "theoretical_loss": 3.908041546390428,
      "tokens_seen": 508166144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042726178535606817,
      "loss": 1.4989,
      "theoretical_loss": 3.9079332949594408,
      "tokens_seen": 508297216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272417251755266,
      "loss": 1.417,
      "theoretical_loss": 3.907825079252789,
      "tokens_seen": 508428288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042722166499498495,
      "loss": 1.4882,
      "theoretical_loss": 3.907716899249477,
      "tokens_seen": 508559360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272016048144433,
      "loss": 1.4883,
      "theoretical_loss": 3.9076087549285274,
      "tokens_seen": 508690432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271815446339017,
      "loss": 1.5095,
      "theoretical_loss": 3.9075006462689825,
      "tokens_seen": 508821504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271614844533601,
      "loss": 1.4736,
      "theoretical_loss": 3.9073925732498997,
      "tokens_seen": 508952576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271414242728185,
      "loss": 1.4869,
      "theoretical_loss": 3.907284535850356,
      "tokens_seen": 509083648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042712136409227687,
      "loss": 1.401,
      "theoretical_loss": 3.9071765340494453,
      "tokens_seen": 509214720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042710130391173523,
      "loss": 1.3473,
      "theoretical_loss": 3.907068567826278,
      "tokens_seen": 509345792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270812437311936,
      "loss": 1.3851,
      "theoretical_loss": 3.906960637159985,
      "tokens_seen": 509476864
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.29997557401657104,
      "objective/train/avg_token_score": 0.05142614245414734,
      "objective/train/docs_used": 298986,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.076449990272522,
      "objective/train/original_loss": 4.7175726890563965,
      "objective/train/theoretical_loss": 3.9069066851541487,
      "objective/train/tokens_used": 509542400,
      "objective/train/unlikelihood_loss": 0.11390677094459534,
      "theoretical_loss": 3.9069066851541487,
      "tokens_seen": 509542400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427061183550652,
      "loss": 1.4249,
      "theoretical_loss": 3.906852742029712,
      "tokens_seen": 509607936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042704112337011037,
      "loss": 1.4852,
      "theoretical_loss": 3.9067448824146243,
      "tokens_seen": 509739008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042702106318956873,
      "loss": 1.411,
      "theoretical_loss": 3.906637058293903,
      "tokens_seen": 509870080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270010030090271,
      "loss": 1.31,
      "theoretical_loss": 3.9065292696467475,
      "tokens_seen": 510001152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269809428284855,
      "loss": 1.5244,
      "theoretical_loss": 3.906421516452376,
      "tokens_seen": 510132224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269608826479439,
      "loss": 1.5396,
      "theoretical_loss": 3.9063137986900216,
      "tokens_seen": 510263296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042694082246740224,
      "loss": 1.4904,
      "theoretical_loss": 3.9062061163389377,
      "tokens_seen": 510394368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269207622868606,
      "loss": 1.4771,
      "theoretical_loss": 3.906098469378393,
      "tokens_seen": 510525440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042690070210631896,
      "loss": 1.4655,
      "theoretical_loss": 3.905990857787675,
      "tokens_seen": 510656512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268806419257774,
      "loss": 1.5493,
      "theoretical_loss": 3.905883281546088,
      "tokens_seen": 510787584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042686058174523574,
      "loss": 1.604,
      "theoretical_loss": 3.9057757406329534,
      "tokens_seen": 510918656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268405215646941,
      "loss": 1.4769,
      "theoretical_loss": 3.9056682350276106,
      "tokens_seen": 511049728
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.5116080045700073,
      "objective/train/avg_token_score": 0.008771687746047974,
      "objective/train/docs_used": 299996,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8161802291870117,
      "objective/train/original_loss": 3.781794548034668,
      "objective/train/theoretical_loss": 3.905560764709417,
      "objective/train/tokens_used": 511180800,
      "objective/train/unlikelihood_loss": 0.15718472003936768,
      "theoretical_loss": 3.905560764709417,
      "tokens_seen": 511180800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042682046138415246,
      "loss": 1.3467,
      "theoretical_loss": 3.905560764709417,
      "tokens_seen": 511180800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268004012036109,
      "loss": 1.2939,
      "theoretical_loss": 3.905453329657745,
      "tokens_seen": 511311872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042678034102306924,
      "loss": 1.4485,
      "theoretical_loss": 3.9053459298519875,
      "tokens_seen": 511442944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267602808425276,
      "loss": 1.3541,
      "theoretical_loss": 3.905238565271553,
      "tokens_seen": 511574016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042674022066198597,
      "loss": 1.4509,
      "theoretical_loss": 3.9051312358958663,
      "tokens_seen": 511705088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042672016048144433,
      "loss": 1.3144,
      "theoretical_loss": 3.905023941704371,
      "tokens_seen": 511836160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042670010030090275,
      "loss": 1.5859,
      "theoretical_loss": 3.9049166826765283,
      "tokens_seen": 511967232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266800401203611,
      "loss": 1.4643,
      "theoretical_loss": 3.9048094587918154,
      "tokens_seen": 512098304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042665997993981947,
      "loss": 1.62,
      "theoretical_loss": 3.904702270029727,
      "tokens_seen": 512229376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042663991975927783,
      "loss": 1.474,
      "theoretical_loss": 3.9045951163697756,
      "tokens_seen": 512360448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042661985957873625,
      "loss": 1.4682,
      "theoretical_loss": 3.904487997791491,
      "tokens_seen": 512491520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265997993981946,
      "loss": 1.313,
      "theoretical_loss": 3.904380914274419,
      "tokens_seen": 512622592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426579739217653,
      "loss": 1.5021,
      "theoretical_loss": 3.9042738657981237,
      "tokens_seen": 512753664
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.28787878155708313,
      "objective/train/avg_token_score": 0.02153957076370716,
      "objective/train/docs_used": 300643,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0317378044128418,
      "objective/train/original_loss": 4.142201900482178,
      "objective/train/theoretical_loss": 3.9042203546938854,
      "objective/train/tokens_used": 512819200,
      "objective/train/unlikelihood_loss": 0.20485782623291016,
      "theoretical_loss": 3.9042203546938854,
      "tokens_seen": 512819200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042655967903711134,
      "loss": 1.438,
      "theoretical_loss": 3.9041668523421853,
      "tokens_seen": 512884736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265396188565697,
      "loss": 1.5775,
      "theoretical_loss": 3.9040598738862027,
      "tokens_seen": 513015808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265195586760281,
      "loss": 1.4891,
      "theoretical_loss": 3.90395293040979,
      "tokens_seen": 513146880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264994984954865,
      "loss": 1.6842,
      "theoretical_loss": 3.90384602189258,
      "tokens_seen": 513277952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042647943831494484,
      "loss": 1.5024,
      "theoretical_loss": 3.9037391483142208,
      "tokens_seen": 513409024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264593781344032,
      "loss": 1.6385,
      "theoretical_loss": 3.9036323096543795,
      "tokens_seen": 513540096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264393179538616,
      "loss": 1.5277,
      "theoretical_loss": 3.9035255058927394,
      "tokens_seen": 513671168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042641925777332,
      "loss": 1.5275,
      "theoretical_loss": 3.903418737009,
      "tokens_seen": 513802240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042639919759277834,
      "loss": 1.3673,
      "theoretical_loss": 3.903312002982879,
      "tokens_seen": 513933312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263791374122367,
      "loss": 1.2568,
      "theoretical_loss": 3.9032053037941106,
      "tokens_seen": 514064384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042635907723169507,
      "loss": 1.3819,
      "theoretical_loss": 3.903098639422445,
      "tokens_seen": 514195456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263390170511535,
      "loss": 1.5151,
      "theoretical_loss": 3.902992009847652,
      "tokens_seen": 514326528
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.45747801661491394,
      "objective/train/avg_token_score": 0.014842421747744083,
      "objective/train/docs_used": 301818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6070836782455444,
      "objective/train/original_loss": 3.8919403553009033,
      "objective/train/theoretical_loss": 3.9028854150495143,
      "objective/train/tokens_used": 514457600,
      "objective/train/unlikelihood_loss": 0.16238826513290405,
      "theoretical_loss": 3.9028854150495143,
      "tokens_seen": 514457600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042631895687061185,
      "loss": 1.4055,
      "theoretical_loss": 3.9028854150495143,
      "tokens_seen": 514457600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262988966900702,
      "loss": 1.4375,
      "theoretical_loss": 3.9027788550078357,
      "tokens_seen": 514588672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042627883650952857,
      "loss": 1.461,
      "theoretical_loss": 3.9026723297024337,
      "tokens_seen": 514719744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426258776328987,
      "loss": 1.4347,
      "theoretical_loss": 3.9025658391131444,
      "tokens_seen": 514850816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042623871614844535,
      "loss": 1.5544,
      "theoretical_loss": 3.9024593832198198,
      "tokens_seen": 514981888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262186559679037,
      "loss": 1.4037,
      "theoretical_loss": 3.9023529620023294,
      "tokens_seen": 515112960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261985957873621,
      "loss": 1.4876,
      "theoretical_loss": 3.902246575440559,
      "tokens_seen": 515244032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042617853560682044,
      "loss": 1.3702,
      "theoretical_loss": 3.9021402235144116,
      "tokens_seen": 515375104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042615847542627885,
      "loss": 1.3778,
      "theoretical_loss": 3.9020339062038065,
      "tokens_seen": 515506176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261384152457372,
      "loss": 1.4036,
      "theoretical_loss": 3.90192762348868,
      "tokens_seen": 515637248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261183550651956,
      "loss": 1.4649,
      "theoretical_loss": 3.901821375348985,
      "tokens_seen": 515768320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042609829488465394,
      "loss": 1.5229,
      "theoretical_loss": 3.9017151617646917,
      "tokens_seen": 515899392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042607823470411236,
      "loss": 1.5503,
      "theoretical_loss": 3.9016089827157865,
      "tokens_seen": 516030464
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.44012707471847534,
      "objective/train/avg_token_score": 0.02129729837179184,
      "objective/train/docs_used": 302480,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6454428434371948,
      "objective/train/original_loss": 4.097105026245117,
      "objective/train/theoretical_loss": 3.9015559061358545,
      "objective/train/tokens_used": 516096000,
      "objective/train/unlikelihood_loss": 0.17164163291454315,
      "theoretical_loss": 3.9015559061358545,
      "tokens_seen": 516096000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260581745235707,
      "loss": 1.5246,
      "theoretical_loss": 3.901502838182272,
      "tokens_seen": 516161536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260381143430291,
      "loss": 1.397,
      "theoretical_loss": 3.9013967281441677,
      "tokens_seen": 516292608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042601805416248744,
      "loss": 1.4353,
      "theoretical_loss": 3.901290652581511,
      "tokens_seen": 516423680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259979939819458,
      "loss": 1.5998,
      "theoretical_loss": 3.9011846114743536,
      "tokens_seen": 516554752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259779338014042,
      "loss": 1.5021,
      "theoretical_loss": 3.9010786048027666,
      "tokens_seen": 516685824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259578736208626,
      "loss": 1.5956,
      "theoretical_loss": 3.900972632546835,
      "tokens_seen": 516816896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042593781344032095,
      "loss": 1.5683,
      "theoretical_loss": 3.9008666946866613,
      "tokens_seen": 516947968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259177532597793,
      "loss": 1.486,
      "theoretical_loss": 3.900760791202366,
      "tokens_seen": 517079040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258976930792377,
      "loss": 1.378,
      "theoretical_loss": 3.9006549220740836,
      "tokens_seen": 517210112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258776328986961,
      "loss": 1.4012,
      "theoretical_loss": 3.9005490872819673,
      "tokens_seen": 517341184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042585757271815445,
      "loss": 1.3931,
      "theoretical_loss": 3.9004432868061856,
      "tokens_seen": 517472256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258375125376128,
      "loss": 1.5117,
      "theoretical_loss": 3.9003375206269233,
      "tokens_seen": 517603328
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.3442082107067108,
      "objective/train/avg_token_score": 0.01602189801633358,
      "objective/train/docs_used": 303675,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2318471670150757,
      "objective/train/original_loss": 4.042038440704346,
      "objective/train/theoretical_loss": 3.9002317887243834,
      "objective/train/tokens_used": 517734400,
      "objective/train/unlikelihood_loss": 0.17470526695251465,
      "theoretical_loss": 3.9002317887243834,
      "tokens_seen": 517734400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258174523570712,
      "loss": 1.4767,
      "theoretical_loss": 3.9002317887243834,
      "tokens_seen": 517734400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257973921765296,
      "loss": 1.3376,
      "theoretical_loss": 3.900126091078783,
      "tokens_seen": 517865472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042577733199598795,
      "loss": 1.3587,
      "theoretical_loss": 3.9000204276703565,
      "tokens_seen": 517996544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257572718154463,
      "loss": 1.3704,
      "theoretical_loss": 3.8999147984793554,
      "tokens_seen": 518127616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257372116349047,
      "loss": 1.47,
      "theoretical_loss": 3.8998092034860474,
      "tokens_seen": 518258688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257171514543631,
      "loss": 1.3936,
      "theoretical_loss": 3.8997036426707155,
      "tokens_seen": 518389760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042569709127382145,
      "loss": 1.4903,
      "theoretical_loss": 3.89959811601366,
      "tokens_seen": 518520832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256770310932798,
      "loss": 1.4924,
      "theoretical_loss": 3.899492623495198,
      "tokens_seen": 518651904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042565697091273823,
      "loss": 1.5132,
      "theoretical_loss": 3.8993871650956606,
      "tokens_seen": 518782976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042563691073219665,
      "loss": 1.4211,
      "theoretical_loss": 3.8992817407953986,
      "tokens_seen": 518914048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425616850551655,
      "loss": 1.6041,
      "theoretical_loss": 3.8991763505747765,
      "tokens_seen": 519045120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255967903711134,
      "loss": 1.3711,
      "theoretical_loss": 3.899070994414176,
      "tokens_seen": 519176192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042557673019057174,
      "loss": 1.5216,
      "theoretical_loss": 3.8989656722939943,
      "tokens_seen": 519307264
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.5273093581199646,
      "objective/train/avg_token_score": 0.004553263075649738,
      "objective/train/docs_used": 304298,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.785210132598877,
      "objective/train/original_loss": 3.7178406715393066,
      "objective/train/theoretical_loss": 3.8989130239929395,
      "objective/train/tokens_used": 519372800,
      "objective/train/unlikelihood_loss": 0.14067475497722626,
      "theoretical_loss": 3.8989130239929395,
      "tokens_seen": 519372800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255566700100301,
      "loss": 1.581,
      "theoretical_loss": 3.8988603841946463,
      "tokens_seen": 519438336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255366098294885,
      "loss": 1.3521,
      "theoretical_loss": 3.898755130096562,
      "tokens_seen": 519569408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255165496489469,
      "loss": 1.5186,
      "theoretical_loss": 3.8986499099801875,
      "tokens_seen": 519700480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042549648946840524,
      "loss": 1.5265,
      "theoretical_loss": 3.8985447238259856,
      "tokens_seen": 519831552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254764292878636,
      "loss": 1.5886,
      "theoretical_loss": 3.8984395716144347,
      "tokens_seen": 519962624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425456369107322,
      "loss": 1.2698,
      "theoretical_loss": 3.8983344533260307,
      "tokens_seen": 520093696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254363089267804,
      "loss": 1.4448,
      "theoretical_loss": 3.8982293689412835,
      "tokens_seen": 520224768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042541624874623874,
      "loss": 1.5163,
      "theoretical_loss": 3.898124318440721,
      "tokens_seen": 520355840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253961885656971,
      "loss": 1.527,
      "theoretical_loss": 3.8980193018048857,
      "tokens_seen": 520486912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042537612838515547,
      "loss": 1.6545,
      "theoretical_loss": 3.8979143190143377,
      "tokens_seen": 520617984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253560682046139,
      "loss": 1.4184,
      "theoretical_loss": 3.8978093700496514,
      "tokens_seen": 520749056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042533600802407225,
      "loss": 1.4471,
      "theoretical_loss": 3.897704454891419,
      "tokens_seen": 520880128
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.42644184827804565,
      "objective/train/avg_token_score": 0.012661627493798733,
      "objective/train/docs_used": 305033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3695693016052246,
      "objective/train/original_loss": 3.7167294025421143,
      "objective/train/theoretical_loss": 3.897599573520247,
      "objective/train/tokens_used": 521011200,
      "objective/train/unlikelihood_loss": 0.21720042824745178,
      "theoretical_loss": 3.897599573520247,
      "tokens_seen": 521011200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253159478435306,
      "loss": 1.4253,
      "theoretical_loss": 3.897599573520247,
      "tokens_seen": 521011200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042529588766298897,
      "loss": 1.4934,
      "theoretical_loss": 3.8974947259167596,
      "tokens_seen": 521142272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252758274824474,
      "loss": 1.4735,
      "theoretical_loss": 3.8973899120615956,
      "tokens_seen": 521273344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042525576730190575,
      "loss": 1.4753,
      "theoretical_loss": 3.897285131935411,
      "tokens_seen": 521404416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252357071213641,
      "loss": 1.6052,
      "theoretical_loss": 3.897180385518876,
      "tokens_seen": 521535488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252156469408225,
      "loss": 1.4427,
      "theoretical_loss": 3.8970756727926785,
      "tokens_seen": 521666560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042519558676028084,
      "loss": 1.5753,
      "theoretical_loss": 3.896970993737522,
      "tokens_seen": 521797632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042517552657973925,
      "loss": 1.5369,
      "theoretical_loss": 3.8968663483341253,
      "tokens_seen": 521928704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251554663991976,
      "loss": 1.3264,
      "theoretical_loss": 3.8967617365632217,
      "tokens_seen": 522059776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425135406218656,
      "loss": 1.3828,
      "theoretical_loss": 3.8966571584055645,
      "tokens_seen": 522190848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042511534603811434,
      "loss": 1.5941,
      "theoretical_loss": 3.8965526138419184,
      "tokens_seen": 522321920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042509528585757276,
      "loss": 1.284,
      "theoretical_loss": 3.896448102853067,
      "tokens_seen": 522452992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250752256770311,
      "loss": 1.4964,
      "theoretical_loss": 3.8963436254198074,
      "tokens_seen": 522584064
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.4902859330177307,
      "objective/train/avg_token_score": 0.007528013549745083,
      "objective/train/docs_used": 306243,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7427791357040405,
      "objective/train/original_loss": 3.7014918327331543,
      "objective/train/theoretical_loss": 3.8962913992805284,
      "objective/train/tokens_used": 522649600,
      "objective/train/unlikelihood_loss": 0.16139429807662964,
      "theoretical_loss": 3.8962913992805284,
      "tokens_seen": 522649600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250551654964895,
      "loss": 1.5339,
      "theoretical_loss": 3.896239181522954,
      "tokens_seen": 522715136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042503510531594784,
      "loss": 1.5216,
      "theoretical_loss": 3.8961347711433376,
      "tokens_seen": 522846208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250150451354062,
      "loss": 1.4623,
      "theoretical_loss": 3.8960303942618024,
      "tokens_seen": 522977280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249949849548646,
      "loss": 1.346,
      "theoretical_loss": 3.8959260508592104,
      "tokens_seen": 523108352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424974924774323,
      "loss": 1.6433,
      "theoretical_loss": 3.895821740916439,
      "tokens_seen": 523239424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042495486459378135,
      "loss": 1.3234,
      "theoretical_loss": 3.89571746441438,
      "tokens_seen": 523370496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249348044132397,
      "loss": 1.546,
      "theoretical_loss": 3.8956132213339423,
      "tokens_seen": 523501568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249147442326981,
      "loss": 1.4065,
      "theoretical_loss": 3.8955090116560505,
      "tokens_seen": 523632640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248946840521565,
      "loss": 1.494,
      "theoretical_loss": 3.8954048353616435,
      "tokens_seen": 523763712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042487462387161485,
      "loss": 1.6275,
      "theoretical_loss": 3.895300692431677,
      "tokens_seen": 523894784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248545636910732,
      "loss": 1.3638,
      "theoretical_loss": 3.895196582847123,
      "tokens_seen": 524025856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248345035105316,
      "loss": 1.4729,
      "theoretical_loss": 3.8950925065889663,
      "tokens_seen": 524156928
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.4181329309940338,
      "objective/train/avg_token_score": 0.01068844273686409,
      "objective/train/docs_used": 306972,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3113154172897339,
      "objective/train/original_loss": 3.7316195964813232,
      "objective/train/theoretical_loss": 3.8949884636382106,
      "objective/train/tokens_used": 524288000,
      "objective/train/unlikelihood_loss": 0.1870473474264145,
      "theoretical_loss": 3.8949884636382106,
      "tokens_seen": 524288000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042481444332999,
      "loss": 1.5167,
      "theoretical_loss": 3.8949884636382106,
      "tokens_seen": 524288000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042479438314944835,
      "loss": 1.4619,
      "theoretical_loss": 3.894884453975873,
      "tokens_seen": 524419072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247743229689067,
      "loss": 1.4227,
      "theoretical_loss": 3.894780477582988,
      "tokens_seen": 524550144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247542627883651,
      "loss": 1.5495,
      "theoretical_loss": 3.8946765344406034,
      "tokens_seen": 524681216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247342026078235,
      "loss": 1.362,
      "theoretical_loss": 3.8945726245297836,
      "tokens_seen": 524812288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042471414242728186,
      "loss": 1.4983,
      "theoretical_loss": 3.894468747831609,
      "tokens_seen": 524943360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246940822467402,
      "loss": 1.7403,
      "theoretical_loss": 3.8943649043271744,
      "tokens_seen": 525074432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246740220661986,
      "loss": 1.5129,
      "theoretical_loss": 3.894261093997592,
      "tokens_seen": 525205504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424653961885657,
      "loss": 1.535,
      "theoretical_loss": 3.8941573168239874,
      "tokens_seen": 525336576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042463390170511536,
      "loss": 1.4884,
      "theoretical_loss": 3.894053572787502,
      "tokens_seen": 525467648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246138415245737,
      "loss": 1.6132,
      "theoretical_loss": 3.8939498618692934,
      "tokens_seen": 525598720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245937813440321,
      "loss": 1.5335,
      "theoretical_loss": 3.8938461840505347,
      "tokens_seen": 525729792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042457372116349044,
      "loss": 1.4542,
      "theoretical_loss": 3.893742539312414,
      "tokens_seen": 525860864
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.43389540910720825,
      "objective/train/avg_token_score": 0.013585958629846573,
      "objective/train/docs_used": 308263,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.377792477607727,
      "objective/train/original_loss": 3.8568532466888428,
      "objective/train/theoretical_loss": 3.8936907293427176,
      "objective/train/tokens_used": 525926400,
      "objective/train/unlikelihood_loss": 0.1510537713766098,
      "theoretical_loss": 3.8936907293427176,
      "tokens_seen": 525926400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042455366098294886,
      "loss": 1.3444,
      "theoretical_loss": 3.893638927636134,
      "tokens_seen": 525991936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245336008024072,
      "loss": 1.5525,
      "theoretical_loss": 3.8935353490029136,
      "tokens_seen": 526123008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245135406218656,
      "loss": 1.395,
      "theoretical_loss": 3.8934318033939874,
      "tokens_seen": 526254080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042449348044132395,
      "loss": 1.4722,
      "theoretical_loss": 3.8933282907906044,
      "tokens_seen": 526385152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042447342026078236,
      "loss": 1.4891,
      "theoretical_loss": 3.89322481117403,
      "tokens_seen": 526516224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244533600802407,
      "loss": 1.3796,
      "theoretical_loss": 3.893121364525544,
      "tokens_seen": 526647296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244332998996991,
      "loss": 1.4325,
      "theoretical_loss": 3.893017950826441,
      "tokens_seen": 526778368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042441323971915745,
      "loss": 1.519,
      "theoretical_loss": 3.892914570058033,
      "tokens_seen": 526909440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243931795386158,
      "loss": 1.5274,
      "theoretical_loss": 3.8928112222016455,
      "tokens_seen": 527040512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042437311935807423,
      "loss": 1.4975,
      "theoretical_loss": 3.8927079072386186,
      "tokens_seen": 527171584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243530591775326,
      "loss": 1.3574,
      "theoretical_loss": 3.8926046251503092,
      "tokens_seen": 527302656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042433299899699095,
      "loss": 1.4539,
      "theoretical_loss": 3.892501375918089,
      "tokens_seen": 527433728
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.3826979398727417,
      "objective/train/avg_token_score": 0.035878412425518036,
      "objective/train/docs_used": 309026,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3176822662353516,
      "objective/train/original_loss": 4.551613807678223,
      "objective/train/theoretical_loss": 3.892398159523345,
      "objective/train/tokens_used": 527564800,
      "objective/train/unlikelihood_loss": 0.1040167510509491,
      "theoretical_loss": 3.892398159523345,
      "tokens_seen": 527564800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243129388164493,
      "loss": 1.4981,
      "theoretical_loss": 3.892398159523345,
      "tokens_seen": 527564800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042429287863590773,
      "loss": 1.5112,
      "theoretical_loss": 3.892294975947477,
      "tokens_seen": 527695872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242728184553661,
      "loss": 1.4593,
      "theoretical_loss": 3.8921918251719054,
      "tokens_seen": 527826944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042425275827482446,
      "loss": 1.6896,
      "theoretical_loss": 3.8920887071780594,
      "tokens_seen": 527958016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242326980942828,
      "loss": 1.4538,
      "theoretical_loss": 3.891985621947388,
      "tokens_seen": 528089088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242126379137412,
      "loss": 1.5069,
      "theoretical_loss": 3.891882569461353,
      "tokens_seen": 528220160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241925777331996,
      "loss": 1.6161,
      "theoretical_loss": 3.8917795497014316,
      "tokens_seen": 528351232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424172517552658,
      "loss": 1.315,
      "theoretical_loss": 3.891676562649116,
      "tokens_seen": 528482304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241524573721164,
      "loss": 1.4037,
      "theoretical_loss": 3.891573608285915,
      "tokens_seen": 528613376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042413239719157474,
      "loss": 1.5572,
      "theoretical_loss": 3.89147068659335,
      "tokens_seen": 528744448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042411233701103316,
      "loss": 1.5609,
      "theoretical_loss": 3.891367797552959,
      "tokens_seen": 528875520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240922768304915,
      "loss": 1.5477,
      "theoretical_loss": 3.891264941146295,
      "tokens_seen": 529006592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240722166499499,
      "loss": 1.4536,
      "theoretical_loss": 3.8911621173549253,
      "tokens_seen": 529137664
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.5902981162071228,
      "objective/train/avg_token_score": 0.02185690961778164,
      "objective/train/docs_used": 310459,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0812337398529053,
      "objective/train/original_loss": 3.739985227584839,
      "objective/train/theoretical_loss": 3.89111071768422,
      "objective/train/tokens_used": 529203200,
      "objective/train/unlikelihood_loss": 0.1313895732164383,
      "theoretical_loss": 3.89111071768422,
      "tokens_seen": 529203200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042405215646940824,
      "loss": 1.5481,
      "theoretical_loss": 3.8910593261604323,
      "tokens_seen": 529268736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240320962888666,
      "loss": 1.3968,
      "theoretical_loss": 3.8909565675444138,
      "tokens_seen": 529399808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424012036108325,
      "loss": 1.4023,
      "theoretical_loss": 3.8908538414884823,
      "tokens_seen": 529530880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239919759277834,
      "loss": 1.3563,
      "theoretical_loss": 3.890751147974265,
      "tokens_seen": 529661952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042397191574724175,
      "loss": 1.6032,
      "theoretical_loss": 3.890648486983405,
      "tokens_seen": 529793024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239518555667001,
      "loss": 1.5371,
      "theoretical_loss": 3.8905458584975583,
      "tokens_seen": 529924096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239317953861585,
      "loss": 1.5931,
      "theoretical_loss": 3.8904432624983976,
      "tokens_seen": 530055168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239117352056169,
      "loss": 1.5401,
      "theoretical_loss": 3.89034069896761,
      "tokens_seen": 530186240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042389167502507525,
      "loss": 1.4977,
      "theoretical_loss": 3.8902381678868974,
      "tokens_seen": 530317312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238716148445336,
      "loss": 1.4138,
      "theoretical_loss": 3.890135669237976,
      "tokens_seen": 530448384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423851554663992,
      "loss": 1.4784,
      "theoretical_loss": 3.8900332030025773,
      "tokens_seen": 530579456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238314944834504,
      "loss": 1.3966,
      "theoretical_loss": 3.889930769162448,
      "tokens_seen": 530710528
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.38929617404937744,
      "objective/train/avg_token_score": 0.033991601318120956,
      "objective/train/docs_used": 311126,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3811792135238647,
      "objective/train/original_loss": 4.070007801055908,
      "objective/train/theoretical_loss": 3.889828367699349,
      "objective/train/tokens_used": 530841600,
      "objective/train/unlikelihood_loss": 0.13424085080623627,
      "theoretical_loss": 3.889828367699349,
      "tokens_seen": 530841600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042381143430290875,
      "loss": 1.4535,
      "theoretical_loss": 3.889828367699349,
      "tokens_seen": 530841600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237913741223671,
      "loss": 1.4853,
      "theoretical_loss": 3.8897259985950554,
      "tokens_seen": 530972672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237713139418255,
      "loss": 1.5492,
      "theoretical_loss": 3.889623661831359,
      "tokens_seen": 531103744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237512537612839,
      "loss": 1.432,
      "theoretical_loss": 3.8895213573900635,
      "tokens_seen": 531234816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042373119358074226,
      "loss": 1.404,
      "theoretical_loss": 3.8894190852529906,
      "tokens_seen": 531365888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237111334002006,
      "loss": 1.3642,
      "theoretical_loss": 3.889316845401974,
      "tokens_seen": 531496960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423691073219659,
      "loss": 1.4066,
      "theoretical_loss": 3.8892146378188643,
      "tokens_seen": 531628032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236710130391174,
      "loss": 1.576,
      "theoretical_loss": 3.8891124624855236,
      "tokens_seen": 531759104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042365095285857576,
      "loss": 1.4427,
      "theoretical_loss": 3.889010319383832,
      "tokens_seen": 531890176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236308926780341,
      "loss": 1.3313,
      "theoretical_loss": 3.888908208495683,
      "tokens_seen": 532021248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236108324974925,
      "loss": 1.5414,
      "theoretical_loss": 3.888806129802984,
      "tokens_seen": 532152320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042359077231695085,
      "loss": 1.5081,
      "theoretical_loss": 3.8887040832876587,
      "tokens_seen": 532283392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042357071213640926,
      "loss": 1.5332,
      "theoretical_loss": 3.888602068931643,
      "tokens_seen": 532414464
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.3950391113758087,
      "objective/train/avg_token_score": 0.011321122758090496,
      "objective/train/docs_used": 312276,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4738465547561646,
      "objective/train/original_loss": 3.9406888484954834,
      "objective/train/theoretical_loss": 3.888551073807736,
      "objective/train/tokens_used": 532480000,
      "objective/train/unlikelihood_loss": 0.1509210616350174,
      "theoretical_loss": 3.888551073807736,
      "tokens_seen": 532480000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235506519558676,
      "loss": 1.4874,
      "theoretical_loss": 3.8885000867168893,
      "tokens_seen": 532545536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423530591775326,
      "loss": 1.4809,
      "theoretical_loss": 3.888398136625364,
      "tokens_seen": 532676608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042351053159478435,
      "loss": 1.4155,
      "theoretical_loss": 3.8882962186390486,
      "tokens_seen": 532807680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042349047141424276,
      "loss": 1.4746,
      "theoretical_loss": 3.8881943327399378,
      "tokens_seen": 532938752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042347041123370113,
      "loss": 1.5315,
      "theoretical_loss": 3.8880924789100417,
      "tokens_seen": 533069824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234503510531595,
      "loss": 1.4988,
      "theoretical_loss": 3.8879906571313847,
      "tokens_seen": 533200896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042343029087261785,
      "loss": 1.4565,
      "theoretical_loss": 3.887888867386007,
      "tokens_seen": 533331968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234102306920762,
      "loss": 1.4381,
      "theoretical_loss": 3.88778710965596,
      "tokens_seen": 533463040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042339017051153463,
      "loss": 1.4021,
      "theoretical_loss": 3.887685383923314,
      "tokens_seen": 533594112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423370110330993,
      "loss": 1.5644,
      "theoretical_loss": 3.887583690170149,
      "tokens_seen": 533725184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042335005015045135,
      "loss": 1.5509,
      "theoretical_loss": 3.887482028378564,
      "tokens_seen": 533856256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233299899699097,
      "loss": 1.4937,
      "theoretical_loss": 3.8873803985306683,
      "tokens_seen": 533987328
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.49890029430389404,
      "objective/train/avg_token_score": 0.006122749298810959,
      "objective/train/docs_used": 312792,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6340548992156982,
      "objective/train/original_loss": 3.637239456176758,
      "objective/train/theoretical_loss": 3.8872788006085894,
      "objective/train/tokens_used": 534118400,
      "objective/train/unlikelihood_loss": 0.22649259865283966,
      "theoretical_loss": 3.8872788006085894,
      "tokens_seen": 534118400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042330992978936813,
      "loss": 1.5323,
      "theoretical_loss": 3.8872788006085894,
      "tokens_seen": 534118400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232898696088265,
      "loss": 1.3563,
      "theoretical_loss": 3.887177234594466,
      "tokens_seen": 534249472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042326980942828486,
      "loss": 1.2005,
      "theoretical_loss": 3.8870757004704526,
      "tokens_seen": 534380544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232497492477432,
      "loss": 1.5123,
      "theoretical_loss": 3.8869741982187183,
      "tokens_seen": 534511616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232296890672016,
      "loss": 1.5522,
      "theoretical_loss": 3.886872727821446,
      "tokens_seen": 534642688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042320962888666,
      "loss": 1.5628,
      "theoretical_loss": 3.886771289260834,
      "tokens_seen": 534773760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042318956870611836,
      "loss": 1.6368,
      "theoretical_loss": 3.886669882519093,
      "tokens_seen": 534904832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231695085255767,
      "loss": 1.6142,
      "theoretical_loss": 3.886568507578448,
      "tokens_seen": 535035904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231494483450351,
      "loss": 1.4894,
      "theoretical_loss": 3.886467164421142,
      "tokens_seen": 535166976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231293881644935,
      "loss": 1.5009,
      "theoretical_loss": 3.886365853029427,
      "tokens_seen": 535298048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042310932798395186,
      "loss": 1.3925,
      "theoretical_loss": 3.886264573385574,
      "tokens_seen": 535429120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230892678034102,
      "loss": 1.4485,
      "theoretical_loss": 3.886163325471864,
      "tokens_seen": 535560192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230692076228686,
      "loss": 1.4131,
      "theoretical_loss": 3.8860621092705956,
      "tokens_seen": 535691264
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.4310850501060486,
      "objective/train/avg_token_score": 0.006272444501519203,
      "objective/train/docs_used": 314102,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.511539101600647,
      "objective/train/original_loss": 3.6116085052490234,
      "objective/train/theoretical_loss": 3.8860115130565984,
      "objective/train/tokens_used": 535756800,
      "objective/train/unlikelihood_loss": 0.3091060221195221,
      "theoretical_loss": 3.8860115130565984,
      "tokens_seen": 535756800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042304914744232695,
      "loss": 1.4495,
      "theoretical_loss": 3.8859609247640794,
      "tokens_seen": 535822336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042302908726178537,
      "loss": 1.4349,
      "theoretical_loss": 3.885859771934642,
      "tokens_seen": 535953408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042300902708124373,
      "loss": 1.1708,
      "theoretical_loss": 3.8857586507646227,
      "tokens_seen": 536084480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229889669007021,
      "loss": 1.4465,
      "theoretical_loss": 3.8856575612363753,
      "tokens_seen": 536215552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042296890672016045,
      "loss": 1.5385,
      "theoretical_loss": 3.885556503332268,
      "tokens_seen": 536346624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042294884653961887,
      "loss": 1.5361,
      "theoretical_loss": 3.885455477034683,
      "tokens_seen": 536477696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042292878635907723,
      "loss": 1.4312,
      "theoretical_loss": 3.8853544823260178,
      "tokens_seen": 536608768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229087261785356,
      "loss": 1.4874,
      "theoretical_loss": 3.885253519188682,
      "tokens_seen": 536739840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042288866599799396,
      "loss": 1.431,
      "theoretical_loss": 3.8851525876050994,
      "tokens_seen": 536870912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228686058174523,
      "loss": 1.3144,
      "theoretical_loss": 3.8850516875577092,
      "tokens_seen": 537001984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042284854563691074,
      "loss": 1.4203,
      "theoretical_loss": 3.884950819028965,
      "tokens_seen": 537133056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228284854563691,
      "loss": 1.5352,
      "theoretical_loss": 3.8848499820013327,
      "tokens_seen": 537264128
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.45100194215774536,
      "objective/train/avg_token_score": 0.006222711876034737,
      "objective/train/docs_used": 314596,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7215205430984497,
      "objective/train/original_loss": 3.9992692470550537,
      "objective/train/theoretical_loss": 3.8847491764572926,
      "objective/train/tokens_used": 537395200,
      "objective/train/unlikelihood_loss": 0.1294986605644226,
      "theoretical_loss": 3.8847491764572926,
      "tokens_seen": 537395200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042280842527582746,
      "loss": 1.4693,
      "theoretical_loss": 3.8847491764572926,
      "tokens_seen": 537395200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227883650952858,
      "loss": 1.3508,
      "theoretical_loss": 3.884648402379341,
      "tokens_seen": 537526272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042276830491474424,
      "loss": 1.4451,
      "theoretical_loss": 3.8845476597499857,
      "tokens_seen": 537657344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227482447342026,
      "loss": 1.41,
      "theoretical_loss": 3.8844469485517488,
      "tokens_seen": 537788416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042272818455366096,
      "loss": 1.4082,
      "theoretical_loss": 3.8843462687671684,
      "tokens_seen": 537919488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227081243731193,
      "loss": 1.3873,
      "theoretical_loss": 3.884245620378794,
      "tokens_seen": 538050560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226880641925777,
      "loss": 1.5178,
      "theoretical_loss": 3.8841450033691913,
      "tokens_seen": 538181632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042266800401203616,
      "loss": 1.4785,
      "theoretical_loss": 3.884044417720938,
      "tokens_seen": 538312704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226479438314945,
      "loss": 1.527,
      "theoretical_loss": 3.8839438634166275,
      "tokens_seen": 538443776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226278836509529,
      "loss": 1.4577,
      "theoretical_loss": 3.883843340438865,
      "tokens_seen": 538574848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042260782347041125,
      "loss": 1.3839,
      "theoretical_loss": 3.8837428487702716,
      "tokens_seen": 538705920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042258776328986966,
      "loss": 1.5392,
      "theoretical_loss": 3.883642388393481,
      "tokens_seen": 538836992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422567703109328,
      "loss": 1.5504,
      "theoretical_loss": 3.8835419592911413,
      "tokens_seen": 538968064
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.32624632120132446,
      "objective/train/avg_token_score": 0.007150240708142519,
      "objective/train/docs_used": 315427,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1273283958435059,
      "objective/train/original_loss": 3.809842348098755,
      "objective/train/theoretical_loss": 3.8834917564624716,
      "objective/train/tokens_used": 539033600,
      "objective/train/unlikelihood_loss": 0.1816972941160202,
      "theoretical_loss": 3.8834917564624716,
      "tokens_seen": 539033600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225476429287864,
      "loss": 1.3572,
      "theoretical_loss": 3.883441561445914,
      "tokens_seen": 539099136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042252758274824475,
      "loss": 1.5028,
      "theoretical_loss": 3.8833411948404755,
      "tokens_seen": 539230208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042250752256770317,
      "loss": 1.3948,
      "theoretical_loss": 3.8832408594575147,
      "tokens_seen": 539361280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042248746238716153,
      "loss": 1.2123,
      "theoretical_loss": 3.883140555279735,
      "tokens_seen": 539492352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224674022066199,
      "loss": 1.3314,
      "theoretical_loss": 3.8830402822898535,
      "tokens_seen": 539623424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042244734202607825,
      "loss": 1.4584,
      "theoretical_loss": 3.8829400404706003,
      "tokens_seen": 539754496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224272818455366,
      "loss": 1.4907,
      "theoretical_loss": 3.8828398298047206,
      "tokens_seen": 539885568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042240722166499503,
      "loss": 1.5874,
      "theoretical_loss": 3.8827396502749734,
      "tokens_seen": 540016640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223871614844534,
      "loss": 1.4599,
      "theoretical_loss": 3.882639501864129,
      "tokens_seen": 540147712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042236710130391176,
      "loss": 1.3443,
      "theoretical_loss": 3.882539384554974,
      "tokens_seen": 540278784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223470411233701,
      "loss": 1.4215,
      "theoretical_loss": 3.8824392983303078,
      "tokens_seen": 540409856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042232698094282853,
      "loss": 1.5224,
      "theoretical_loss": 3.8823392431729435,
      "tokens_seen": 540540928
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.3801930546760559,
      "objective/train/avg_token_score": 0.03133194521069527,
      "objective/train/docs_used": 316530,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4709320068359375,
      "objective/train/original_loss": 4.609292030334473,
      "objective/train/theoretical_loss": 3.882239219065708,
      "objective/train/tokens_used": 540672000,
      "objective/train/unlikelihood_loss": 0.10041511058807373,
      "theoretical_loss": 3.882239219065708,
      "tokens_seen": 540672000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223069207622869,
      "loss": 1.531,
      "theoretical_loss": 3.882239219065708,
      "tokens_seen": 540672000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042228686058174526,
      "loss": 1.3437,
      "theoretical_loss": 3.882139225991441,
      "tokens_seen": 540803072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222668004012036,
      "loss": 1.6213,
      "theoretical_loss": 3.882039263932997,
      "tokens_seen": 540934144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422246740220662,
      "loss": 1.3605,
      "theoretical_loss": 3.8819393328732446,
      "tokens_seen": 541065216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222266800401204,
      "loss": 1.459,
      "theoretical_loss": 3.881839432795064,
      "tokens_seen": 541196288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042220661985957876,
      "loss": 1.525,
      "theoretical_loss": 3.88173956368135,
      "tokens_seen": 541327360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221865596790371,
      "loss": 1.541,
      "theoretical_loss": 3.8816397255150115,
      "tokens_seen": 541458432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221664994984955,
      "loss": 1.2984,
      "theoretical_loss": 3.8815399182789703,
      "tokens_seen": 541589504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221464393179539,
      "loss": 1.4195,
      "theoretical_loss": 3.881440141956162,
      "tokens_seen": 541720576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042212637913741226,
      "loss": 1.3955,
      "theoretical_loss": 3.8813403965295357,
      "tokens_seen": 541851648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221063189568706,
      "loss": 1.4605,
      "theoretical_loss": 3.881240681982054,
      "tokens_seen": 541982720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422086258776329,
      "loss": 1.4174,
      "theoretical_loss": 3.881140998296694,
      "tokens_seen": 542113792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042206619859578735,
      "loss": 1.4816,
      "theoretical_loss": 3.881041345456444,
      "tokens_seen": 542244864
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.44330400228500366,
      "objective/train/avg_token_score": 0.020557841286063194,
      "objective/train/docs_used": 317270,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5593016147613525,
      "objective/train/original_loss": 3.649369955062866,
      "objective/train/theoretical_loss": 3.8809915305979237,
      "objective/train/tokens_used": 542310400,
      "objective/train/unlikelihood_loss": 0.21089822053909302,
      "theoretical_loss": 3.8809915305979237,
      "tokens_seen": 542310400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042204613841524577,
      "loss": 1.4163,
      "theoretical_loss": 3.880941723444308,
      "tokens_seen": 542375936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042202607823470413,
      "loss": 1.5283,
      "theoretical_loss": 3.8808421322433024,
      "tokens_seen": 542507008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220060180541625,
      "loss": 1.469,
      "theoretical_loss": 3.880742571836457,
      "tokens_seen": 542638080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042198595787362085,
      "loss": 1.3293,
      "theoretical_loss": 3.880643042206816,
      "tokens_seen": 542769152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042196589769307927,
      "loss": 1.4443,
      "theoretical_loss": 3.8805435433374353,
      "tokens_seen": 542900224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042194583751253763,
      "loss": 1.4672,
      "theoretical_loss": 3.8804440752113862,
      "tokens_seen": 543031296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000421925777331996,
      "loss": 1.4297,
      "theoretical_loss": 3.880344637811752,
      "tokens_seen": 543162368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042190571715145436,
      "loss": 1.4319,
      "theoretical_loss": 3.8802452311216302,
      "tokens_seen": 543293440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218856569709127,
      "loss": 1.5163,
      "theoretical_loss": 3.8801458551241312,
      "tokens_seen": 543424512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042186559679037114,
      "loss": 1.471,
      "theoretical_loss": 3.880046509802378,
      "tokens_seen": 543555584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218455366098295,
      "loss": 1.4938,
      "theoretical_loss": 3.8799471951395095,
      "tokens_seen": 543686656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042182547642928786,
      "loss": 1.5712,
      "theoretical_loss": 3.8798479111186754,
      "tokens_seen": 543817728
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.3662634491920471,
      "objective/train/avg_token_score": 0.021842803806066513,
      "objective/train/docs_used": 318285,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.198555588722229,
      "objective/train/original_loss": 3.5977704524993896,
      "objective/train/theoretical_loss": 3.879748657723039,
      "objective/train/tokens_used": 543948800,
      "objective/train/unlikelihood_loss": 0.24941086769104004,
      "theoretical_loss": 3.879748657723039,
      "tokens_seen": 543948800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218054162487462,
      "loss": 1.4546,
      "theoretical_loss": 3.879748657723039,
      "tokens_seen": 543948800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042178535606820464,
      "loss": 1.4712,
      "theoretical_loss": 3.879649434935778,
      "tokens_seen": 544079872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000421765295887663,
      "loss": 1.6343,
      "theoretical_loss": 3.879550242740083,
      "tokens_seen": 544210944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042174523570712136,
      "loss": 1.5444,
      "theoretical_loss": 3.879451081119158,
      "tokens_seen": 544342016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217251755265797,
      "loss": 1.6409,
      "theoretical_loss": 3.8793519500562192,
      "tokens_seen": 544473088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004217051153460381,
      "loss": 1.521,
      "theoretical_loss": 3.879252849534498,
      "tokens_seen": 544604160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216850551654965,
      "loss": 1.4057,
      "theoretical_loss": 3.8791537795372366,
      "tokens_seen": 544735232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042166499498495487,
      "loss": 1.6129,
      "theoretical_loss": 3.8790547400476925,
      "tokens_seen": 544866304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042164493480441323,
      "loss": 1.5271,
      "theoretical_loss": 3.878955731049136,
      "tokens_seen": 544997376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216248746238716,
      "loss": 1.496,
      "theoretical_loss": 3.8788567525248494,
      "tokens_seen": 545128448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042160481444333,
      "loss": 1.3628,
      "theoretical_loss": 3.8787578044581292,
      "tokens_seen": 545259520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042158475426278837,
      "loss": 1.4522,
      "theoretical_loss": 3.8786588868322847,
      "tokens_seen": 545390592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042156469408224673,
      "loss": 1.3902,
      "theoretical_loss": 3.8785599996306397,
      "tokens_seen": 545521664
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.4580278694629669,
      "objective/train/avg_token_score": 0.007631353102624416,
      "objective/train/docs_used": 318961,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5138821601867676,
      "objective/train/original_loss": 3.7409567832946777,
      "objective/train/theoretical_loss": 3.8785105674336835,
      "objective/train/tokens_used": 545587200,
      "objective/train/unlikelihood_loss": 0.14674793183803558,
      "theoretical_loss": 3.8785105674336835,
      "tokens_seen": 545587200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215446339017051,
      "loss": 1.4147,
      "theoretical_loss": 3.8784611428365285,
      "tokens_seen": 545652736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215245737211635,
      "loss": 1.4939,
      "theoretical_loss": 3.878362316433301,
      "tokens_seen": 545783808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215045135406219,
      "loss": 1.4692,
      "theoretical_loss": 3.8782635204043188,
      "tokens_seen": 545914880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042148445336008024,
      "loss": 1.4711,
      "theoretical_loss": 3.8781647547329565,
      "tokens_seen": 546045952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214643931795386,
      "loss": 1.5982,
      "theoretical_loss": 3.8780660194026035,
      "tokens_seen": 546177024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042144433299899696,
      "loss": 1.5872,
      "theoretical_loss": 3.8779673143966606,
      "tokens_seen": 546308096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214242728184554,
      "loss": 1.6839,
      "theoretical_loss": 3.8778686396985416,
      "tokens_seen": 546439168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042140421263791374,
      "loss": 1.4686,
      "theoretical_loss": 3.8777699952916738,
      "tokens_seen": 546570240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213841524573721,
      "loss": 1.6047,
      "theoretical_loss": 3.877671381159499,
      "tokens_seen": 546701312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042136409227683046,
      "loss": 1.331,
      "theoretical_loss": 3.877572797285469,
      "tokens_seen": 546832384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213440320962889,
      "loss": 1.3915,
      "theoretical_loss": 3.877474243653051,
      "tokens_seen": 546963456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042132397191574724,
      "loss": 1.5564,
      "theoretical_loss": 3.877375720245725,
      "tokens_seen": 547094528
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.41086265444755554,
      "objective/train/avg_token_score": 0.00896799098700285,
      "objective/train/docs_used": 320353,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4873528480529785,
      "objective/train/original_loss": 3.9619879722595215,
      "objective/train/theoretical_loss": 3.8772772270469824,
      "objective/train/tokens_used": 547225600,
      "objective/train/unlikelihood_loss": 0.1763414591550827,
      "theoretical_loss": 3.8772772270469824,
      "tokens_seen": 547225600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213039117352056,
      "loss": 1.5364,
      "theoretical_loss": 3.8772772270469824,
      "tokens_seen": 547225600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042128385155466397,
      "loss": 1.4582,
      "theoretical_loss": 3.877178764040329,
      "tokens_seen": 547356672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042126379137412233,
      "loss": 1.5835,
      "theoretical_loss": 3.8770803312092834,
      "tokens_seen": 547487744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042124373119358075,
      "loss": 1.3389,
      "theoretical_loss": 3.8769819285373766,
      "tokens_seen": 547618816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212236710130391,
      "loss": 1.518,
      "theoretical_loss": 3.8768835560081527,
      "tokens_seen": 547749888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042120361083249747,
      "loss": 1.417,
      "theoretical_loss": 3.8767852136051686,
      "tokens_seen": 547880960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042118355065195583,
      "loss": 1.46,
      "theoretical_loss": 3.876686901311995,
      "tokens_seen": 548012032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211634904714143,
      "loss": 1.6196,
      "theoretical_loss": 3.8765886191122148,
      "tokens_seen": 548143104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042114343029087266,
      "loss": 1.4089,
      "theoretical_loss": 3.876490366989423,
      "tokens_seen": 548274176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042112337011033103,
      "loss": 1.4457,
      "theoretical_loss": 3.876392144927229,
      "tokens_seen": 548405248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211033099297894,
      "loss": 1.4709,
      "theoretical_loss": 3.8762939529092537,
      "tokens_seen": 548536320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042108324974924775,
      "loss": 1.6226,
      "theoretical_loss": 3.876195790919132,
      "tokens_seen": 548667392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042106318956870617,
      "loss": 1.5583,
      "theoretical_loss": 3.8760976589405107,
      "tokens_seen": 548798464
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.4123900234699249,
      "objective/train/avg_token_score": 0.014366233721375465,
      "objective/train/docs_used": 320989,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.500076174736023,
      "objective/train/original_loss": 3.8920931816101074,
      "objective/train/theoretical_loss": 3.8760486042004056,
      "objective/train/tokens_used": 548864000,
      "objective/train/unlikelihood_loss": 0.15627138316631317,
      "theoretical_loss": 3.8760486042004056,
      "tokens_seen": 548864000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042104312938816453,
      "loss": 1.426,
      "theoretical_loss": 3.8759995569570496,
      "tokens_seen": 548929536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210230692076229,
      "loss": 1.4949,
      "theoretical_loss": 3.875901484952422,
      "tokens_seen": 549060608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042100300902708125,
      "loss": 1.5319,
      "theoretical_loss": 3.8758034429103128,
      "tokens_seen": 549191680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042098294884653967,
      "loss": 1.6009,
      "theoretical_loss": 3.875705430814421,
      "tokens_seen": 549322752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042096288866599803,
      "loss": 1.2854,
      "theoretical_loss": 3.8756074486484566,
      "tokens_seen": 549453824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209428284854564,
      "loss": 1.3986,
      "theoretical_loss": 3.8755094963961447,
      "tokens_seen": 549584896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042092276830491476,
      "loss": 1.4487,
      "theoretical_loss": 3.8754115740412214,
      "tokens_seen": 549715968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209027081243731,
      "loss": 1.4448,
      "theoretical_loss": 3.8753136815674356,
      "tokens_seen": 549847040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042088264794383154,
      "loss": 1.5002,
      "theoretical_loss": 3.875215818958549,
      "tokens_seen": 549978112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208625877632899,
      "loss": 1.6046,
      "theoretical_loss": 3.8751179861983367,
      "tokens_seen": 550109184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042084252758274826,
      "loss": 1.3805,
      "theoretical_loss": 3.8750201832705864,
      "tokens_seen": 550240256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208224674022066,
      "loss": 1.6252,
      "theoretical_loss": 3.8749224101590976,
      "tokens_seen": 550371328
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.3899071216583252,
      "objective/train/avg_token_score": 0.00873059593141079,
      "objective/train/docs_used": 322374,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3493757247924805,
      "objective/train/original_loss": 3.7656335830688477,
      "objective/train/theoretical_loss": 3.8748246668476827,
      "objective/train/tokens_used": 550502400,
      "objective/train/unlikelihood_loss": 0.2136422097682953,
      "theoretical_loss": 3.8748246668476827,
      "tokens_seen": 550502400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042080240722166504,
      "loss": 1.5383,
      "theoretical_loss": 3.8748246668476827,
      "tokens_seen": 550502400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207823470411234,
      "loss": 1.4345,
      "theoretical_loss": 3.8747269533201676,
      "tokens_seen": 550633472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042076228686058176,
      "loss": 1.4031,
      "theoretical_loss": 3.8746292695603897,
      "tokens_seen": 550764544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207422266800401,
      "loss": 1.6027,
      "theoretical_loss": 3.8745316155521996,
      "tokens_seen": 550895616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207221664994985,
      "loss": 1.6353,
      "theoretical_loss": 3.874433991279461,
      "tokens_seen": 551026688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207021063189569,
      "loss": 1.5923,
      "theoretical_loss": 3.874336396726049,
      "tokens_seen": 551157760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042068204613841527,
      "loss": 1.5882,
      "theoretical_loss": 3.874238831875852,
      "tokens_seen": 551288832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042066198595787363,
      "loss": 1.5571,
      "theoretical_loss": 3.8741412967127706,
      "tokens_seen": 551419904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420641925777332,
      "loss": 1.3437,
      "theoretical_loss": 3.8740437912207186,
      "tokens_seen": 551550976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206218655967904,
      "loss": 1.463,
      "theoretical_loss": 3.873946315383622,
      "tokens_seen": 551682048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042060180541624877,
      "loss": 1.4278,
      "theoretical_loss": 3.8738488691854194,
      "tokens_seen": 551813120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042058174523570713,
      "loss": 1.2151,
      "theoretical_loss": 3.873751452610061,
      "tokens_seen": 551944192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205616850551655,
      "loss": 1.4381,
      "theoretical_loss": 3.873654065641511,
      "tokens_seen": 552075264
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.48796433210372925,
      "objective/train/avg_token_score": 0.016625795513391495,
      "objective/train/docs_used": 322945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7378815412521362,
      "objective/train/original_loss": 4.003293037414551,
      "objective/train/theoretical_loss": 3.8736053832547808,
      "objective/train/tokens_used": 552140800,
      "objective/train/unlikelihood_loss": 0.14501631259918213,
      "theoretical_loss": 3.8736053832547808,
      "tokens_seen": 552140800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205416248746239,
      "loss": 1.4769,
      "theoretical_loss": 3.873556708263745,
      "tokens_seen": 552206336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205215646940823,
      "loss": 1.4547,
      "theoretical_loss": 3.8734593804607513,
      "tokens_seen": 552337408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042050150451354064,
      "loss": 1.5517,
      "theoretical_loss": 3.8733620822165316,
      "tokens_seen": 552468480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420481444332999,
      "loss": 1.4435,
      "theoretical_loss": 3.8732648135150978,
      "tokens_seen": 552599552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042046138415245736,
      "loss": 1.565,
      "theoretical_loss": 3.873167574340477,
      "tokens_seen": 552730624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204413239719158,
      "loss": 1.3768,
      "theoretical_loss": 3.873070364676707,
      "tokens_seen": 552861696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042042126379137414,
      "loss": 1.5248,
      "theoretical_loss": 3.8729731845078383,
      "tokens_seen": 552992768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204012036108325,
      "loss": 1.1907,
      "theoretical_loss": 3.872876033817934,
      "tokens_seen": 553123840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042038114343029086,
      "loss": 1.432,
      "theoretical_loss": 3.872778912591069,
      "tokens_seen": 553254912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203610832497493,
      "loss": 1.4321,
      "theoretical_loss": 3.8726818208113323,
      "tokens_seen": 553385984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042034102306920764,
      "loss": 1.3276,
      "theoretical_loss": 3.872584758462823,
      "tokens_seen": 553517056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420320962888666,
      "loss": 1.6415,
      "theoretical_loss": 3.8724877255296533,
      "tokens_seen": 553648128
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.4739125072956085,
      "objective/train/avg_token_score": 0.016837503761053085,
      "objective/train/docs_used": 324208,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4981387853622437,
      "objective/train/original_loss": 3.6097116470336914,
      "objective/train/theoretical_loss": 3.8723907219959486,
      "objective/train/tokens_used": 553779200,
      "objective/train/unlikelihood_loss": 0.127224862575531,
      "theoretical_loss": 3.8723907219959486,
      "tokens_seen": 553779200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042030090270812437,
      "loss": 1.446,
      "theoretical_loss": 3.8723907219959486,
      "tokens_seen": 553779200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042028084252758273,
      "loss": 1.4663,
      "theoretical_loss": 3.872293747845846,
      "tokens_seen": 553910272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042026078234704115,
      "loss": 1.4597,
      "theoretical_loss": 3.872196803063495,
      "tokens_seen": 554041344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202407221664995,
      "loss": 1.5731,
      "theoretical_loss": 3.8720998876330572,
      "tokens_seen": 554172416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042022066198595787,
      "loss": 1.4404,
      "theoretical_loss": 3.8720030015387064,
      "tokens_seen": 554303488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042020060180541623,
      "loss": 1.4862,
      "theoretical_loss": 3.8719061447646297,
      "tokens_seen": 554434560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042018054162487465,
      "loss": 1.2866,
      "theoretical_loss": 3.8718093172950248,
      "tokens_seen": 554565632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420160481444333,
      "loss": 1.4176,
      "theoretical_loss": 3.8717125191141024,
      "tokens_seen": 554696704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201404212637914,
      "loss": 1.2709,
      "theoretical_loss": 3.8716157502060864,
      "tokens_seen": 554827776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042012036108324974,
      "loss": 1.4054,
      "theoretical_loss": 3.871519010555211,
      "tokens_seen": 554958848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201003009027081,
      "loss": 1.329,
      "theoretical_loss": 3.871422300145725,
      "tokens_seen": 555089920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200802407221665,
      "loss": 1.2811,
      "theoretical_loss": 3.871325618961887,
      "tokens_seen": 555220992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200601805416249,
      "loss": 1.4898,
      "theoretical_loss": 3.871228966987969,
      "tokens_seen": 555352064
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.40713587403297424,
      "objective/train/avg_token_score": 0.009826954454183578,
      "objective/train/docs_used": 324930,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.531923532485962,
      "objective/train/original_loss": 3.965177297592163,
      "objective/train/theoretical_loss": 3.871180651949819,
      "objective/train/tokens_used": 555417600,
      "objective/train/unlikelihood_loss": 0.14890779554843903,
      "theoretical_loss": 3.871180651949819,
      "tokens_seen": 555417600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042004012036108324,
      "loss": 1.4487,
      "theoretical_loss": 3.871132344208256,
      "tokens_seen": 555483136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200200601805416,
      "loss": 1.3508,
      "theoretical_loss": 3.8710357506070427,
      "tokens_seen": 555614208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042,
      "loss": 1.5513,
      "theoretical_loss": 3.870939186168638,
      "tokens_seen": 555745280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199799398194584,
      "loss": 1.314,
      "theoretical_loss": 3.8708426508773632,
      "tokens_seen": 555876352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041995987963891674,
      "loss": 1.319,
      "theoretical_loss": 3.8707461447175495,
      "tokens_seen": 556007424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199398194583751,
      "loss": 1.4497,
      "theoretical_loss": 3.870649667673543,
      "tokens_seen": 556138496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041991975927783347,
      "loss": 1.3447,
      "theoretical_loss": 3.8705532197296995,
      "tokens_seen": 556269568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198996990972919,
      "loss": 1.375,
      "theoretical_loss": 3.870456800870388,
      "tokens_seen": 556400640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041987963891675025,
      "loss": 1.4515,
      "theoretical_loss": 3.8703604110799903,
      "tokens_seen": 556531712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198595787362086,
      "loss": 1.3271,
      "theoretical_loss": 3.8702640503428984,
      "tokens_seen": 556662784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041983951855566697,
      "loss": 1.3552,
      "theoretical_loss": 3.8701677186435184,
      "tokens_seen": 556793856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198194583751254,
      "loss": 1.4606,
      "theoretical_loss": 3.8700714159662666,
      "tokens_seen": 556924928
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.32086998224258423,
      "objective/train/avg_token_score": 0.02666344866156578,
      "objective/train/docs_used": 326252,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0482410192489624,
      "objective/train/original_loss": 4.232210636138916,
      "objective/train/theoretical_loss": 3.869975142295573,
      "objective/train/tokens_used": 557056000,
      "objective/train/unlikelihood_loss": 0.14249522984027863,
      "theoretical_loss": 3.869975142295573,
      "tokens_seen": 557056000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041979939819458375,
      "loss": 1.5389,
      "theoretical_loss": 3.869975142295573,
      "tokens_seen": 557056000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197793380140421,
      "loss": 1.3538,
      "theoretical_loss": 3.8698788976158776,
      "tokens_seen": 557187072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041975927783350047,
      "loss": 1.315,
      "theoretical_loss": 3.8697826819116345,
      "tokens_seen": 557318144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041973921765295883,
      "loss": 1.6038,
      "theoretical_loss": 3.8696864951673087,
      "tokens_seen": 557449216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041971915747241725,
      "loss": 1.4843,
      "theoretical_loss": 3.8695903373673772,
      "tokens_seen": 557580288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196990972918756,
      "loss": 1.5741,
      "theoretical_loss": 3.869494208496329,
      "tokens_seen": 557711360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419679037111334,
      "loss": 1.5376,
      "theoretical_loss": 3.8693981085386655,
      "tokens_seen": 557842432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196589769307924,
      "loss": 1.4148,
      "theoretical_loss": 3.8693020374788993,
      "tokens_seen": 557973504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196389167502508,
      "loss": 1.5087,
      "theoretical_loss": 3.8692059953015563,
      "tokens_seen": 558104576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041961885656970917,
      "loss": 1.2978,
      "theoretical_loss": 3.8691099819911723,
      "tokens_seen": 558235648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041959879638916753,
      "loss": 1.468,
      "theoretical_loss": 3.8690139975322966,
      "tokens_seen": 558366720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195787362086259,
      "loss": 1.275,
      "theoretical_loss": 3.8689180419094895,
      "tokens_seen": 558497792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195586760280843,
      "loss": 1.5295,
      "theoretical_loss": 3.868822115107324,
      "tokens_seen": 558628864
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.48692572116851807,
      "objective/train/avg_token_score": 0.008406232111155987,
      "objective/train/docs_used": 326886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.637534499168396,
      "objective/train/original_loss": 3.788033962249756,
      "objective/train/theoretical_loss": 3.8687741625091636,
      "objective/train/tokens_used": 558694400,
      "objective/train/unlikelihood_loss": 0.1396278291940689,
      "theoretical_loss": 3.8687741625091636,
      "tokens_seen": 558694400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195386158475427,
      "loss": 1.4419,
      "theoretical_loss": 3.8687262171103836,
      "tokens_seen": 558759936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041951855566700104,
      "loss": 1.6244,
      "theoretical_loss": 3.868630347903266,
      "tokens_seen": 558891008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194984954864594,
      "loss": 1.3994,
      "theoretical_loss": 3.868534507470579,
      "tokens_seen": 559022080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041947843530591776,
      "loss": 1.4045,
      "theoretical_loss": 3.868438695796941,
      "tokens_seen": 559153152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194583751253762,
      "loss": 1.3849,
      "theoretical_loss": 3.8683429128669857,
      "tokens_seen": 559284224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041943831494483454,
      "loss": 1.3363,
      "theoretical_loss": 3.868247158665356,
      "tokens_seen": 559415296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194182547642929,
      "loss": 1.3883,
      "theoretical_loss": 3.868151433176707,
      "tokens_seen": 559546368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041939819458375126,
      "loss": 1.5001,
      "theoretical_loss": 3.868055736385706,
      "tokens_seen": 559677440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193781344032097,
      "loss": 1.4969,
      "theoretical_loss": 3.867960068277033,
      "tokens_seen": 559808512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041935807422266804,
      "loss": 1.3821,
      "theoretical_loss": 3.867864428835377,
      "tokens_seen": 559939584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193380140421264,
      "loss": 1.4965,
      "theoretical_loss": 3.8677688180454415,
      "tokens_seen": 560070656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041931795386158477,
      "loss": 1.5585,
      "theoretical_loss": 3.8676732358919406,
      "tokens_seen": 560201728
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.2959432899951935,
      "objective/train/avg_token_score": 0.025653989985585213,
      "objective/train/docs_used": 327930,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9278034567832947,
      "objective/train/original_loss": 3.611701011657715,
      "objective/train/theoretical_loss": 3.8675776823595998,
      "objective/train/tokens_used": 560332800,
      "objective/train/unlikelihood_loss": 0.2902003228664398,
      "theoretical_loss": 3.8675776823595998,
      "tokens_seen": 560332800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041929789368104313,
      "loss": 1.3127,
      "theoretical_loss": 3.8675776823595998,
      "tokens_seen": 560332800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041927783350050155,
      "loss": 1.5624,
      "theoretical_loss": 3.867482157433157,
      "tokens_seen": 560463872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192577733199599,
      "loss": 1.3398,
      "theoretical_loss": 3.8673866610973624,
      "tokens_seen": 560594944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041923771313941827,
      "loss": 1.4846,
      "theoretical_loss": 3.867291193336976,
      "tokens_seen": 560726016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041921765295887663,
      "loss": 1.5351,
      "theoretical_loss": 3.867195754136771,
      "tokens_seen": 560857088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041919759277833505,
      "loss": 1.4848,
      "theoretical_loss": 3.867100343481532,
      "tokens_seen": 560988160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191775325977934,
      "loss": 1.3812,
      "theoretical_loss": 3.867004961356054,
      "tokens_seen": 561119232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191574724172518,
      "loss": 1.3775,
      "theoretical_loss": 3.8669096077451464,
      "tokens_seen": 561250304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041913741223671014,
      "loss": 1.6188,
      "theoretical_loss": 3.8668142826336274,
      "tokens_seen": 561381376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191173520561685,
      "loss": 1.4986,
      "theoretical_loss": 3.8667189860063282,
      "tokens_seen": 561512448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190972918756269,
      "loss": 1.3638,
      "theoretical_loss": 3.866623717848092,
      "tokens_seen": 561643520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190772316950853,
      "loss": 1.4552,
      "theoretical_loss": 3.8665284781437714,
      "tokens_seen": 561774592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041905717151454364,
      "loss": 1.4597,
      "theoretical_loss": 3.866433266878234,
      "tokens_seen": 561905664
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.5047653913497925,
      "objective/train/avg_token_score": 0.00482219597324729,
      "objective/train/docs_used": 328746,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.737789511680603,
      "objective/train/original_loss": 3.9125659465789795,
      "objective/train/theoretical_loss": 3.8663856719052827,
      "objective/train/tokens_used": 561971200,
      "objective/train/unlikelihood_loss": 0.14458067715168,
      "theoretical_loss": 3.8663856719052827,
      "tokens_seen": 561971200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419037111334002,
      "loss": 1.5193,
      "theoretical_loss": 3.8663380840363564,
      "tokens_seen": 562036736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190170511534604,
      "loss": 1.362,
      "theoretical_loss": 3.866242929603027,
      "tokens_seen": 562167808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189969909729188,
      "loss": 1.4048,
      "theoretical_loss": 3.866147803563148,
      "tokens_seen": 562298880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041897693079237714,
      "loss": 1.6211,
      "theoretical_loss": 3.8660527059016294,
      "tokens_seen": 562429952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189568706118355,
      "loss": 1.4695,
      "theoretical_loss": 3.8659576366033956,
      "tokens_seen": 562561024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041893681043129387,
      "loss": 1.4011,
      "theoretical_loss": 3.8658625956533816,
      "tokens_seen": 562692096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189167502507523,
      "loss": 1.3756,
      "theoretical_loss": 3.8657675830365346,
      "tokens_seen": 562823168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041889669007021065,
      "loss": 1.4237,
      "theoretical_loss": 3.8656725987378118,
      "tokens_seen": 562954240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418876629889669,
      "loss": 1.4832,
      "theoretical_loss": 3.865577642742183,
      "tokens_seen": 563085312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041885656970912737,
      "loss": 1.4726,
      "theoretical_loss": 3.86548271503463,
      "tokens_seen": 563216384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188365095285858,
      "loss": 1.3446,
      "theoretical_loss": 3.865387815600144,
      "tokens_seen": 563347456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041881644934804415,
      "loss": 1.5273,
      "theoretical_loss": 3.86529294442373,
      "tokens_seen": 563478528
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.6345918774604797,
      "objective/train/avg_token_score": 0.00891077984124422,
      "objective/train/docs_used": 329425,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0925443172454834,
      "objective/train/original_loss": 3.6826608180999756,
      "objective/train/theoretical_loss": 3.8651981014904027,
      "objective/train/tokens_used": 563609600,
      "objective/train/unlikelihood_loss": 0.14516757428646088,
      "theoretical_loss": 3.8651981014904027,
      "tokens_seen": 563609600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187963891675025,
      "loss": 1.452,
      "theoretical_loss": 3.8651981014904027,
      "tokens_seen": 563609600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187763289869609,
      "loss": 1.2785,
      "theoretical_loss": 3.8651032867851893,
      "tokens_seen": 563740672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041875626880641924,
      "loss": 1.3895,
      "theoretical_loss": 3.8650085002931283,
      "tokens_seen": 563871744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041873620862587765,
      "loss": 1.4778,
      "theoretical_loss": 3.8649137419992687,
      "tokens_seen": 564002816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418716148445336,
      "loss": 1.4185,
      "theoretical_loss": 3.8648190118886725,
      "tokens_seen": 564133888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186960882647944,
      "loss": 1.4829,
      "theoretical_loss": 3.8647243099464106,
      "tokens_seen": 564264960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041867602808425274,
      "loss": 1.6106,
      "theoretical_loss": 3.8646296361575687,
      "tokens_seen": 564396032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041865596790371115,
      "loss": 1.3867,
      "theoretical_loss": 3.8645349905072406,
      "tokens_seen": 564527104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186359077231695,
      "loss": 1.4051,
      "theoretical_loss": 3.864440372980533,
      "tokens_seen": 564658176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186158475426279,
      "loss": 1.5175,
      "theoretical_loss": 3.8643457835625634,
      "tokens_seen": 564789248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041859578736208624,
      "loss": 1.4996,
      "theoretical_loss": 3.864251222238462,
      "tokens_seen": 564920320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185757271815446,
      "loss": 1.4721,
      "theoretical_loss": 3.864156688993369,
      "tokens_seen": 565051392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418555667001003,
      "loss": 1.4107,
      "theoretical_loss": 3.864062183812436,
      "tokens_seen": 565182464
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.37713831663131714,
      "objective/train/avg_token_score": 0.01374147180467844,
      "objective/train/docs_used": 329612,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2935733795166016,
      "objective/train/original_loss": 3.755645275115967,
      "objective/train/theoretical_loss": 3.864014941741392,
      "objective/train/tokens_used": 565248000,
      "objective/train/unlikelihood_loss": 0.1588522344827652,
      "theoretical_loss": 3.864014941741392,
      "tokens_seen": 565248000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185356068204614,
      "loss": 1.5042,
      "theoretical_loss": 3.8639677066808256,
      "tokens_seen": 565313536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041851554663991974,
      "loss": 1.5188,
      "theoretical_loss": 3.8638732575837125,
      "tokens_seen": 565444608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184954864593781,
      "loss": 1.5116,
      "theoretical_loss": 3.863778836506283,
      "tokens_seen": 565575680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184754262788365,
      "loss": 1.5429,
      "theoretical_loss": 3.8636844434337325,
      "tokens_seen": 565706752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184553660982949,
      "loss": 1.6697,
      "theoretical_loss": 3.863590078351271,
      "tokens_seen": 565837824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041843530591775325,
      "loss": 1.4738,
      "theoretical_loss": 3.8634957412441167,
      "tokens_seen": 565968896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184152457372116,
      "loss": 1.5514,
      "theoretical_loss": 3.8634014320975005,
      "tokens_seen": 566099968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041839518555666997,
      "loss": 1.5185,
      "theoretical_loss": 3.863307150896664,
      "tokens_seen": 566231040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183751253761284,
      "loss": 1.4115,
      "theoretical_loss": 3.8632128976268603,
      "tokens_seen": 566362112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041835506519558675,
      "loss": 1.4309,
      "theoretical_loss": 3.8631186722733535,
      "tokens_seen": 566493184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183350050150451,
      "loss": 1.5064,
      "theoretical_loss": 3.8630244748214198,
      "tokens_seen": 566624256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183149448345035,
      "loss": 1.5129,
      "theoretical_loss": 3.862930305256345,
      "tokens_seen": 566755328
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.4685361683368683,
      "objective/train/avg_token_score": 0.020750796422362328,
      "objective/train/docs_used": 331071,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7844135761260986,
      "objective/train/original_loss": 4.0298967361450195,
      "objective/train/theoretical_loss": 3.8628361635634265,
      "objective/train/tokens_used": 566886400,
      "objective/train/unlikelihood_loss": 0.1526690572500229,
      "theoretical_loss": 3.8628361635634265,
      "tokens_seen": 566886400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182948846539619,
      "loss": 1.4882,
      "theoretical_loss": 3.8628361635634265,
      "tokens_seen": 566886400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041827482447342025,
      "loss": 1.5128,
      "theoretical_loss": 3.8627420497279745,
      "tokens_seen": 567017472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182547642928786,
      "loss": 1.6453,
      "theoretical_loss": 3.8626479637353075,
      "tokens_seen": 567148544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418234704112337,
      "loss": 1.5273,
      "theoretical_loss": 3.8625539055707576,
      "tokens_seen": 567279616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182146439317954,
      "loss": 1.5139,
      "theoretical_loss": 3.862459875219667,
      "tokens_seen": 567410688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041819458375125376,
      "loss": 1.4496,
      "theoretical_loss": 3.8623658726673886,
      "tokens_seen": 567541760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181745235707121,
      "loss": 1.5045,
      "theoretical_loss": 3.862271897899287,
      "tokens_seen": 567672832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041815446339017054,
      "loss": 1.5649,
      "theoretical_loss": 3.8621779509007377,
      "tokens_seen": 567803904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181344032096289,
      "loss": 1.4613,
      "theoretical_loss": 3.862084031657128,
      "tokens_seen": 567934976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181143430290873,
      "loss": 1.5159,
      "theoretical_loss": 3.8619901401538543,
      "tokens_seen": 568066048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180942828485457,
      "loss": 1.5407,
      "theoretical_loss": 3.861896276376326,
      "tokens_seen": 568197120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041807422266800404,
      "loss": 1.4418,
      "theoretical_loss": 3.8618024403099636,
      "tokens_seen": 568328192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180541624874624,
      "loss": 1.4974,
      "theoretical_loss": 3.8617086319401968,
      "tokens_seen": 568459264
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.3284457325935364,
      "objective/train/avg_token_score": 0.011517138220369816,
      "objective/train/docs_used": 331416,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2020608186721802,
      "objective/train/original_loss": 4.407777309417725,
      "objective/train/theoretical_loss": 3.861661738136987,
      "objective/train/tokens_used": 568524800,
      "objective/train/unlikelihood_loss": 0.14563091099262238,
      "theoretical_loss": 3.861661738136987,
      "tokens_seen": 568524800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180341023069208,
      "loss": 1.498,
      "theoretical_loss": 3.8616148512524675,
      "tokens_seen": 568590336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180140421263792,
      "loss": 1.51,
      "theoretical_loss": 3.8615210982322288,
      "tokens_seen": 568721408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041799398194583754,
      "loss": 1.59,
      "theoretical_loss": 3.8614273728649446,
      "tokens_seen": 568852480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179739217652959,
      "loss": 1.426,
      "theoretical_loss": 3.8613336751360894,
      "tokens_seen": 568983552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041795386158475427,
      "loss": 1.6531,
      "theoretical_loss": 3.8612400050311493,
      "tokens_seen": 569114624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179338014042127,
      "loss": 1.4923,
      "theoretical_loss": 3.861146362535621,
      "tokens_seen": 569245696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041791374122367105,
      "loss": 1.5301,
      "theoretical_loss": 3.8610527476350116,
      "tokens_seen": 569376768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178936810431294,
      "loss": 1.4552,
      "theoretical_loss": 3.86095916031484,
      "tokens_seen": 569507840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041787362086258777,
      "loss": 1.4689,
      "theoretical_loss": 3.8608656005606354,
      "tokens_seen": 569638912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178535606820462,
      "loss": 1.5132,
      "theoretical_loss": 3.8607720683579396,
      "tokens_seen": 569769984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041783350050150455,
      "loss": 1.4345,
      "theoretical_loss": 3.8606785636923027,
      "tokens_seen": 569901056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178134403209629,
      "loss": 1.5175,
      "theoretical_loss": 3.8605850865492872,
      "tokens_seen": 570032128
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.43047410249710083,
      "objective/train/avg_token_score": 0.02064175345003605,
      "objective/train/docs_used": 332819,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4973698854446411,
      "objective/train/original_loss": 3.5845766067504883,
      "objective/train/theoretical_loss": 3.8604916369144666,
      "objective/train/tokens_used": 570163200,
      "objective/train/unlikelihood_loss": 0.2515372037887573,
      "theoretical_loss": 3.8604916369144666,
      "tokens_seen": 570163200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177933801404213,
      "loss": 1.4534,
      "theoretical_loss": 3.8604916369144666,
      "tokens_seen": 570163200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041777331995987964,
      "loss": 1.6288,
      "theoretical_loss": 3.860398214773425,
      "tokens_seen": 570294272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041775325977933805,
      "loss": 1.4872,
      "theoretical_loss": 3.8603048201117565,
      "tokens_seen": 570425344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177331995987964,
      "loss": 1.4164,
      "theoretical_loss": 3.8602114529150673,
      "tokens_seen": 570556416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177131394182548,
      "loss": 1.525,
      "theoretical_loss": 3.8601181131689737,
      "tokens_seen": 570687488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041769307923771314,
      "loss": 1.438,
      "theoretical_loss": 3.8600248008591045,
      "tokens_seen": 570818560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041767301905717156,
      "loss": 1.4738,
      "theoretical_loss": 3.8599315159710965,
      "tokens_seen": 570949632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176529588766299,
      "loss": 1.6646,
      "theoretical_loss": 3.8598382584905995,
      "tokens_seen": 571080704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176328986960883,
      "loss": 1.5638,
      "theoretical_loss": 3.859745028403273,
      "tokens_seen": 571211776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041761283851554664,
      "loss": 1.6145,
      "theoretical_loss": 3.859651825694787,
      "tokens_seen": 571342848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417592778335005,
      "loss": 1.4472,
      "theoretical_loss": 3.8595586503508246,
      "tokens_seen": 571473920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175727181544634,
      "loss": 1.3499,
      "theoretical_loss": 3.8594655023570765,
      "tokens_seen": 571604992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175526579739218,
      "loss": 1.6236,
      "theoretical_loss": 3.859372381699247,
      "tokens_seen": 571736064
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.31506597995758057,
      "objective/train/avg_token_score": 0.016959313303232193,
      "objective/train/docs_used": 333556,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9501151442527771,
      "objective/train/original_loss": 3.510495901107788,
      "objective/train/theoretical_loss": 3.859325831616836,
      "objective/train/tokens_used": 571801600,
      "objective/train/unlikelihood_loss": 0.19363252818584442,
      "theoretical_loss": 3.859325831616836,
      "tokens_seen": 571801600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041753259779338015,
      "loss": 1.4159,
      "theoretical_loss": 3.859279288363048,
      "tokens_seen": 571867136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175125376128385,
      "loss": 1.6299,
      "theoretical_loss": 3.8591862223342064,
      "tokens_seen": 571998208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174924774322969,
      "loss": 1.5366,
      "theoretical_loss": 3.859093183598455,
      "tokens_seen": 572129280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174724172517553,
      "loss": 1.4375,
      "theoretical_loss": 3.859000172141541,
      "tokens_seen": 572260352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041745235707121365,
      "loss": 1.5723,
      "theoretical_loss": 3.858907187949221,
      "tokens_seen": 572391424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417432296890672,
      "loss": 1.6735,
      "theoretical_loss": 3.858814231007262,
      "tokens_seen": 572522496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041741223671013037,
      "loss": 1.3602,
      "theoretical_loss": 3.858721301301442,
      "tokens_seen": 572653568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173921765295888,
      "loss": 1.4342,
      "theoretical_loss": 3.8586283988175496,
      "tokens_seen": 572784640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041737211634904715,
      "loss": 1.4224,
      "theoretical_loss": 3.8585355235413843,
      "tokens_seen": 572915712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173520561685055,
      "loss": 1.5072,
      "theoretical_loss": 3.858442675458756,
      "tokens_seen": 573046784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173319959879639,
      "loss": 1.5743,
      "theoretical_loss": 3.858349854555486,
      "tokens_seen": 573177856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173119358074223,
      "loss": 1.5621,
      "theoretical_loss": 3.8582570608174045,
      "tokens_seen": 573308928
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.29276636242866516,
      "objective/train/avg_token_score": 0.03536292165517807,
      "objective/train/docs_used": 334721,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1436461210250854,
      "objective/train/original_loss": 4.6639227867126465,
      "objective/train/theoretical_loss": 3.858164294230354,
      "objective/train/tokens_used": 573440000,
      "objective/train/unlikelihood_loss": 0.10083737224340439,
      "theoretical_loss": 3.858164294230354,
      "tokens_seen": 573440000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041729187562688065,
      "loss": 1.4795,
      "theoretical_loss": 3.858164294230354,
      "tokens_seen": 573440000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417271815446339,
      "loss": 1.5866,
      "theoretical_loss": 3.8580715547801865,
      "tokens_seen": 573571072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172517552657974,
      "loss": 1.678,
      "theoretical_loss": 3.8579788424527655,
      "tokens_seen": 573702144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172316950852558,
      "loss": 1.5994,
      "theoretical_loss": 3.857886157233965,
      "tokens_seen": 573833216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041721163490471416,
      "loss": 1.4772,
      "theoretical_loss": 3.857793499109669,
      "tokens_seen": 573964288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171915747241725,
      "loss": 1.5592,
      "theoretical_loss": 3.8577008680657716,
      "tokens_seen": 574095360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171715145436309,
      "loss": 1.5461,
      "theoretical_loss": 3.8576082640881797,
      "tokens_seen": 574226432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041715145436308924,
      "loss": 1.5339,
      "theoretical_loss": 3.8575156871628082,
      "tokens_seen": 574357504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041713139418254766,
      "loss": 1.5063,
      "theoretical_loss": 3.8574231372755836,
      "tokens_seen": 574488576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417111334002006,
      "loss": 1.5664,
      "theoretical_loss": 3.8573306144124437,
      "tokens_seen": 574619648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170912738214644,
      "loss": 1.4587,
      "theoretical_loss": 3.8572381185593354,
      "tokens_seen": 574750720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041707121364092275,
      "loss": 1.5352,
      "theoretical_loss": 3.8571456497022165,
      "tokens_seen": 574881792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041705115346038116,
      "loss": 1.7719,
      "theoretical_loss": 3.8570532078270565,
      "tokens_seen": 575012864
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.410740464925766,
      "objective/train/avg_token_score": 0.012464967556297779,
      "objective/train/docs_used": 335233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4862878322601318,
      "objective/train/original_loss": 3.868055820465088,
      "objective/train/theoretical_loss": 3.857006997003329,
      "objective/train/tokens_used": 575078400,
      "objective/train/unlikelihood_loss": 0.2018526941537857,
      "theoretical_loss": 3.857006997003329,
      "tokens_seen": 575078400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170310932798395,
      "loss": 1.5556,
      "theoretical_loss": 3.8569607929198337,
      "tokens_seen": 575143936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170110330992979,
      "loss": 1.5718,
      "theoretical_loss": 3.856868404966538,
      "tokens_seen": 575275008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041699097291875625,
      "loss": 1.6003,
      "theoretical_loss": 3.856776043953169,
      "tokens_seen": 575406080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169709127382146,
      "loss": 1.6744,
      "theoretical_loss": 3.856683709865737,
      "tokens_seen": 575537152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041695085255767303,
      "loss": 1.4031,
      "theoretical_loss": 3.856591402690264,
      "tokens_seen": 575668224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169307923771314,
      "loss": 1.5336,
      "theoretical_loss": 3.856499122412781,
      "tokens_seen": 575799296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041691073219658975,
      "loss": 1.6005,
      "theoretical_loss": 3.856406869019329,
      "tokens_seen": 575930368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168906720160481,
      "loss": 1.2878,
      "theoretical_loss": 3.856314642495961,
      "tokens_seen": 576061440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041687061183550653,
      "loss": 1.4465,
      "theoretical_loss": 3.8562224428287393,
      "tokens_seen": 576192512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168505516549649,
      "loss": 1.3374,
      "theoretical_loss": 3.856130270003736,
      "tokens_seen": 576323584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041683049147442326,
      "loss": 1.6655,
      "theoretical_loss": 3.8560381240070365,
      "tokens_seen": 576454656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168104312938816,
      "loss": 1.6154,
      "theoretical_loss": 3.855946004824733,
      "tokens_seen": 576585728
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.3080400824546814,
      "objective/train/avg_token_score": 0.017052171751856804,
      "objective/train/docs_used": 335838,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9714576005935669,
      "objective/train/original_loss": 3.790057897567749,
      "objective/train/theoretical_loss": 3.85585391244293,
      "objective/train/tokens_used": 576716800,
      "objective/train/unlikelihood_loss": 0.24715007841587067,
      "theoretical_loss": 3.85585391244293,
      "tokens_seen": 576716800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041679037111334,
      "loss": 1.4714,
      "theoretical_loss": 3.85585391244293,
      "tokens_seen": 576716800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167703109327984,
      "loss": 1.5795,
      "theoretical_loss": 3.8557618468477415,
      "tokens_seen": 576847872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041675025075225676,
      "loss": 1.5285,
      "theoretical_loss": 3.8556698080252936,
      "tokens_seen": 576978944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167301905717151,
      "loss": 1.4149,
      "theoretical_loss": 3.8555777959617203,
      "tokens_seen": 577110016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167101303911735,
      "loss": 1.3468,
      "theoretical_loss": 3.8554858106431675,
      "tokens_seen": 577241088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004166900702106319,
      "loss": 1.5689,
      "theoretical_loss": 3.8553938520557915,
      "tokens_seen": 577372160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041667001003009026,
      "loss": 1.5243,
      "theoretical_loss": 3.8553019201857577,
      "tokens_seen": 577503232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166499498495487,
      "loss": 1.4107,
      "theoretical_loss": 3.8552100150192423,
      "tokens_seen": 577634304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041662988966900704,
      "loss": 1.4474,
      "theoretical_loss": 3.855118136542433,
      "tokens_seen": 577765376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166098294884654,
      "loss": 1.4431,
      "theoretical_loss": 3.855026284741526,
      "tokens_seen": 577896448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165897693079238,
      "loss": 1.4762,
      "theoretical_loss": 3.8549344596027293,
      "tokens_seen": 578027520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165697091273822,
      "loss": 1.3957,
      "theoretical_loss": 3.8548426611122597,
      "tokens_seen": 578158592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041654964894684055,
      "loss": 1.3942,
      "theoretical_loss": 3.854750889256345,
      "tokens_seen": 578289664
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.43218475580215454,
      "objective/train/avg_token_score": 0.007149472367018461,
      "objective/train/docs_used": 337192,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.293778657913208,
      "objective/train/original_loss": 4.05076265335083,
      "objective/train/theoretical_loss": 3.854705013312045,
      "objective/train/tokens_used": 578355200,
      "objective/train/unlikelihood_loss": 0.13084176182746887,
      "theoretical_loss": 3.854705013312045,
      "tokens_seen": 578355200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165295887662989,
      "loss": 1.5097,
      "theoretical_loss": 3.8546591440212237,
      "tokens_seen": 578420736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165095285857573,
      "loss": 1.4131,
      "theoretical_loss": 3.8545674253931432,
      "tokens_seen": 578551808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164894684052157,
      "loss": 1.4509,
      "theoretical_loss": 3.8544757333583632,
      "tokens_seen": 578682880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041646940822467405,
      "loss": 1.3605,
      "theoretical_loss": 3.854384067903151,
      "tokens_seen": 578813952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164493480441324,
      "loss": 1.5583,
      "theoretical_loss": 3.8542924290137863,
      "tokens_seen": 578945024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164292878635908,
      "loss": 1.3599,
      "theoretical_loss": 3.8542008166765576,
      "tokens_seen": 579076096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164092276830492,
      "loss": 1.4714,
      "theoretical_loss": 3.8541092308777647,
      "tokens_seen": 579207168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041638916750250755,
      "loss": 1.5355,
      "theoretical_loss": 3.8540176716037164,
      "tokens_seen": 579338240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163691073219659,
      "loss": 1.2708,
      "theoretical_loss": 3.8539261388407327,
      "tokens_seen": 579469312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163490471414243,
      "loss": 1.5444,
      "theoretical_loss": 3.853834632575143,
      "tokens_seen": 579600384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163289869608827,
      "loss": 1.4964,
      "theoretical_loss": 3.853743152793287,
      "tokens_seen": 579731456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041630892678034105,
      "loss": 1.4589,
      "theoretical_loss": 3.8536516994815146,
      "tokens_seen": 579862528
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.32331377267837524,
      "objective/train/avg_token_score": 0.055630918592214584,
      "objective/train/docs_used": 337878,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.113051414489746,
      "objective/train/original_loss": 4.3019514083862305,
      "objective/train/theoretical_loss": 3.8535602726261864,
      "objective/train/tokens_used": 579993600,
      "objective/train/unlikelihood_loss": 0.16012266278266907,
      "theoretical_loss": 3.8535602726261864,
      "tokens_seen": 579993600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162888665997994,
      "loss": 1.2567,
      "theoretical_loss": 3.8535602726261864,
      "tokens_seen": 579993600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162688064192578,
      "loss": 1.4235,
      "theoretical_loss": 3.8534688722136723,
      "tokens_seen": 580124672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162487462387162,
      "loss": 1.544,
      "theoretical_loss": 3.8533774982303526,
      "tokens_seen": 580255744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041622868605817456,
      "loss": 1.4034,
      "theoretical_loss": 3.8532861506626173,
      "tokens_seen": 580386816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162086258776329,
      "loss": 1.5044,
      "theoretical_loss": 3.8531948294968674,
      "tokens_seen": 580517888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161885656970913,
      "loss": 1.5761,
      "theoretical_loss": 3.853103534719513,
      "tokens_seen": 580648960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041616850551654964,
      "loss": 1.4135,
      "theoretical_loss": 3.8530122663169752,
      "tokens_seen": 580780032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041614844533600806,
      "loss": 1.41,
      "theoretical_loss": 3.852921024275684,
      "tokens_seen": 580911104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161283851554664,
      "loss": 1.3965,
      "theoretical_loss": 3.8528298085820802,
      "tokens_seen": 581042176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161083249749248,
      "loss": 1.4879,
      "theoretical_loss": 3.852738619222615,
      "tokens_seen": 581173248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041608826479438315,
      "loss": 1.4855,
      "theoretical_loss": 3.8526474561837483,
      "tokens_seen": 581304320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041606820461384156,
      "loss": 1.4659,
      "theoretical_loss": 3.852556319451952,
      "tokens_seen": 581435392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160481444332999,
      "loss": 1.5415,
      "theoretical_loss": 3.852465209013705,
      "tokens_seen": 581566464
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.415628045797348,
      "objective/train/avg_token_score": 0.01404473278671503,
      "objective/train/docs_used": 339259,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.375450611114502,
      "objective/train/original_loss": 3.8198952674865723,
      "objective/train/theoretical_loss": 3.8524196636504415,
      "objective/train/tokens_used": 581632000,
      "objective/train/unlikelihood_loss": 0.21155038475990295,
      "theoretical_loss": 3.8524196636504415,
      "tokens_seen": 581632000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160280842527583,
      "loss": 1.3202,
      "theoretical_loss": 3.8523741248555,
      "tokens_seen": 581697536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041600802407221665,
      "loss": 1.3599,
      "theoretical_loss": 3.852283066963836,
      "tokens_seen": 581828608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415987963891675,
      "loss": 1.5542,
      "theoretical_loss": 3.852192035325226,
      "tokens_seen": 581959680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041596790371113343,
      "loss": 1.5125,
      "theoretical_loss": 3.852101029926188,
      "tokens_seen": 582090752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159478435305918,
      "loss": 1.4721,
      "theoretical_loss": 3.852010050753254,
      "tokens_seen": 582221824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041592778335005015,
      "loss": 1.3842,
      "theoretical_loss": 3.8519190977929645,
      "tokens_seen": 582352896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159077231695085,
      "loss": 1.4271,
      "theoretical_loss": 3.85182817103187,
      "tokens_seen": 582483968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041588766298896693,
      "loss": 1.385,
      "theoretical_loss": 3.85173727045653,
      "tokens_seen": 582615040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158676028084253,
      "loss": 1.3607,
      "theoretical_loss": 3.851646396053516,
      "tokens_seen": 582746112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041584754262788366,
      "loss": 1.6908,
      "theoretical_loss": 3.851555547809408,
      "tokens_seen": 582877184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415827482447342,
      "loss": 1.5575,
      "theoretical_loss": 3.8514647257107955,
      "tokens_seen": 583008256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158074222668004,
      "loss": 1.5368,
      "theoretical_loss": 3.8513739297442786,
      "tokens_seen": 583139328
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.37939882278442383,
      "objective/train/avg_token_score": 0.012932621873915195,
      "objective/train/docs_used": 340083,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3577672243118286,
      "objective/train/original_loss": 3.419050931930542,
      "objective/train/theoretical_loss": 3.851283159896468,
      "objective/train/tokens_used": 583270400,
      "objective/train/unlikelihood_loss": 0.294486939907074,
      "theoretical_loss": 3.851283159896468,
      "tokens_seen": 583270400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157873620862588,
      "loss": 1.4448,
      "theoretical_loss": 3.851283159896468,
      "tokens_seen": 583270400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041576730190571716,
      "loss": 1.4166,
      "theoretical_loss": 3.8511924161539826,
      "tokens_seen": 583401472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157472417251755,
      "loss": 1.4534,
      "theoretical_loss": 3.8511016985034523,
      "tokens_seen": 583532544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157271815446339,
      "loss": 1.5211,
      "theoretical_loss": 3.8510110069315173,
      "tokens_seen": 583663616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157071213640923,
      "loss": 1.2879,
      "theoretical_loss": 3.8509203414248256,
      "tokens_seen": 583794688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041568706118355066,
      "loss": 1.4893,
      "theoretical_loss": 3.850829701970037,
      "tokens_seen": 583925760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415667001003009,
      "loss": 1.2667,
      "theoretical_loss": 3.85073908855382,
      "tokens_seen": 584056832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156469408224674,
      "loss": 1.4526,
      "theoretical_loss": 3.8506485011628544,
      "tokens_seen": 584187904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041562688064192575,
      "loss": 1.4288,
      "theoretical_loss": 3.850557939783828,
      "tokens_seen": 584318976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041560682046138417,
      "loss": 1.6138,
      "theoretical_loss": 3.8504674044034393,
      "tokens_seen": 584450048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041558676028084253,
      "loss": 1.633,
      "theoretical_loss": 3.850376895008396,
      "tokens_seen": 584581120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155667001003009,
      "loss": 1.4531,
      "theoretical_loss": 3.850286411585417,
      "tokens_seen": 584712192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041554663991975925,
      "loss": 1.4509,
      "theoretical_loss": 3.8501959541212294,
      "tokens_seen": 584843264
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.31414955854415894,
      "objective/train/avg_token_score": 0.012615097686648369,
      "objective/train/docs_used": 341398,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1601176261901855,
      "objective/train/original_loss": 4.111655235290527,
      "objective/train/theoretical_loss": 3.850150735119537,
      "objective/train/tokens_used": 584908800,
      "objective/train/unlikelihood_loss": 0.15587183833122253,
      "theoretical_loss": 3.850150735119537,
      "tokens_seen": 584908800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041552657973921767,
      "loss": 1.4771,
      "theoretical_loss": 3.8501055226025707,
      "tokens_seen": 584974336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041550651955867603,
      "loss": 1.454,
      "theoretical_loss": 3.850015117016188,
      "tokens_seen": 585105408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154864593781344,
      "loss": 1.5285,
      "theoretical_loss": 3.8499247373488377,
      "tokens_seen": 585236480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041546639919759276,
      "loss": 1.5182,
      "theoretical_loss": 3.8498343835872872,
      "tokens_seen": 585367552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154463390170511,
      "loss": 1.5975,
      "theoretical_loss": 3.8497440557183134,
      "tokens_seen": 585498624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041542627883650954,
      "loss": 1.6371,
      "theoretical_loss": 3.849653753728701,
      "tokens_seen": 585629696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154062186559679,
      "loss": 1.5461,
      "theoretical_loss": 3.8495634776052463,
      "tokens_seen": 585760768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041538615847542626,
      "loss": 1.2369,
      "theoretical_loss": 3.849473227334755,
      "tokens_seen": 585891840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153660982948846,
      "loss": 1.4595,
      "theoretical_loss": 3.849383002904043,
      "tokens_seen": 586022912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041534603811434304,
      "loss": 1.3872,
      "theoretical_loss": 3.8492928042999335,
      "tokens_seen": 586153984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153259779338014,
      "loss": 1.4741,
      "theoretical_loss": 3.849202631509262,
      "tokens_seen": 586285056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041530591775325976,
      "loss": 1.3709,
      "theoretical_loss": 3.8491124845188716,
      "tokens_seen": 586416128
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.35477760434150696,
      "objective/train/avg_token_score": 0.018234847113490105,
      "objective/train/docs_used": 342071,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2699406147003174,
      "objective/train/original_loss": 4.259076118469238,
      "objective/train/theoretical_loss": 3.8490223633156173,
      "objective/train/tokens_used": 586547200,
      "objective/train/unlikelihood_loss": 0.1685103178024292,
      "theoretical_loss": 3.8490223633156173,
      "tokens_seen": 586547200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152858575727181,
      "loss": 1.4075,
      "theoretical_loss": 3.8490223633156173,
      "tokens_seen": 586547200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152657973921765,
      "loss": 1.3445,
      "theoretical_loss": 3.8489322678863624,
      "tokens_seen": 586678272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152457372116349,
      "loss": 1.4237,
      "theoretical_loss": 3.848842198217979,
      "tokens_seen": 586809344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041522567703109327,
      "loss": 1.4361,
      "theoretical_loss": 3.84875215429735,
      "tokens_seen": 586940416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041520561685055163,
      "loss": 1.3083,
      "theoretical_loss": 3.8486621361113684,
      "tokens_seen": 587071488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041518555667001,
      "loss": 1.3499,
      "theoretical_loss": 3.848572143646936,
      "tokens_seen": 587202560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151654964894684,
      "loss": 1.521,
      "theoretical_loss": 3.8484821768909634,
      "tokens_seen": 587333632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151454363089268,
      "loss": 1.5277,
      "theoretical_loss": 3.8483922358303717,
      "tokens_seen": 587464704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151253761283852,
      "loss": 1.4885,
      "theoretical_loss": 3.8483023204520928,
      "tokens_seen": 587595776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041510531594784355,
      "loss": 1.5283,
      "theoretical_loss": 3.848212430743065,
      "tokens_seen": 587726848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041508525576730196,
      "loss": 1.446,
      "theoretical_loss": 3.848122566690239,
      "tokens_seen": 587857920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150651955867603,
      "loss": 1.4859,
      "theoretical_loss": 3.8480327282805744,
      "tokens_seen": 587988992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150451354062187,
      "loss": 1.3935,
      "theoretical_loss": 3.847942915501039,
      "tokens_seen": 588120064
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.4999389052391052,
      "objective/train/avg_token_score": 0.008891026489436626,
      "objective/train/docs_used": 342650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.745088815689087,
      "objective/train/original_loss": 3.872380256652832,
      "objective/train/theoretical_loss": 3.8478980187185003,
      "objective/train/tokens_used": 588185600,
      "objective/train/unlikelihood_loss": 0.1340382695198059,
      "theoretical_loss": 3.8478980187185003,
      "tokens_seen": 588185600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041502507522567705,
      "loss": 1.3831,
      "theoretical_loss": 3.8478531283386124,
      "tokens_seen": 588251136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150050150451354,
      "loss": 1.4038,
      "theoretical_loss": 3.847763366780281,
      "tokens_seen": 588382208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041498495486459383,
      "loss": 1.5453,
      "theoretical_loss": 3.8476736308130426,
      "tokens_seen": 588513280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149648946840522,
      "loss": 1.4038,
      "theoretical_loss": 3.8475839204239044,
      "tokens_seen": 588644352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041494483450351055,
      "loss": 1.3637,
      "theoretical_loss": 3.8474942355998825,
      "tokens_seen": 588775424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149247743229689,
      "loss": 1.4316,
      "theoretical_loss": 3.8474045763280023,
      "tokens_seen": 588906496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041490471414242733,
      "loss": 1.6323,
      "theoretical_loss": 3.8473149425953,
      "tokens_seen": 589037568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148846539618857,
      "loss": 1.4029,
      "theoretical_loss": 3.8472253343888196,
      "tokens_seen": 589168640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041486459378134406,
      "loss": 1.4195,
      "theoretical_loss": 3.8471357516956153,
      "tokens_seen": 589299712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148445336008024,
      "loss": 1.5306,
      "theoretical_loss": 3.8470461945027505,
      "tokens_seen": 589430784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148244734202608,
      "loss": 1.4443,
      "theoretical_loss": 3.8469566627972993,
      "tokens_seen": 589561856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148044132397192,
      "loss": 1.4823,
      "theoretical_loss": 3.8468671565663426,
      "tokens_seen": 589692928
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.5726417303085327,
      "objective/train/avg_token_score": 0.010895898565649986,
      "objective/train/docs_used": 343935,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.0607388019561768,
      "objective/train/original_loss": 3.6878914833068848,
      "objective/train/theoretical_loss": 3.846777675796974,
      "objective/train/tokens_used": 589824000,
      "objective/train/unlikelihood_loss": 0.14677737653255463,
      "theoretical_loss": 3.846777675796974,
      "tokens_seen": 589824000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041478435305917756,
      "loss": 1.3588,
      "theoretical_loss": 3.846777675796974,
      "tokens_seen": 589824000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147642928786359,
      "loss": 1.5026,
      "theoretical_loss": 3.8466882204762936,
      "tokens_seen": 589955072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147442326980943,
      "loss": 1.3238,
      "theoretical_loss": 3.846598790591412,
      "tokens_seen": 590086144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147241725175527,
      "loss": 1.5089,
      "theoretical_loss": 3.8465093861294504,
      "tokens_seen": 590217216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041470411233701106,
      "loss": 1.4212,
      "theoretical_loss": 3.846420007077537,
      "tokens_seen": 590348288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146840521564694,
      "loss": 1.4739,
      "theoretical_loss": 3.846330653422812,
      "tokens_seen": 590479360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146639919759278,
      "loss": 1.501,
      "theoretical_loss": 3.8462413251524215,
      "tokens_seen": 590610432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041464393179538615,
      "loss": 1.5693,
      "theoretical_loss": 3.8461520222535253,
      "tokens_seen": 590741504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041462387161484457,
      "loss": 1.5297,
      "theoretical_loss": 3.846062744713289,
      "tokens_seen": 590872576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041460381143430293,
      "loss": 1.4618,
      "theoretical_loss": 3.8459734925188895,
      "tokens_seen": 591003648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145837512537613,
      "loss": 1.3261,
      "theoretical_loss": 3.8458842656575123,
      "tokens_seen": 591134720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041456369107321965,
      "loss": 1.5224,
      "theoretical_loss": 3.8457950641163516,
      "tokens_seen": 591265792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041454363089267807,
      "loss": 1.3278,
      "theoretical_loss": 3.845705887882612,
      "tokens_seen": 591396864
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.4558895528316498,
      "objective/train/avg_token_score": 0.015658829361200333,
      "objective/train/docs_used": 344744,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6233446598052979,
      "objective/train/original_loss": 3.805819511413574,
      "objective/train/theoretical_loss": 3.845661309252029,
      "objective/train/tokens_used": 591462400,
      "objective/train/unlikelihood_loss": 0.17303699254989624,
      "theoretical_loss": 3.845661309252029,
      "tokens_seen": 591462400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041452357071213643,
      "loss": 1.5114,
      "theoretical_loss": 3.8456167369435073,
      "tokens_seen": 591527936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145035105315948,
      "loss": 1.6073,
      "theoretical_loss": 3.84552761128626,
      "tokens_seen": 591659008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041448345035105316,
      "loss": 1.5835,
      "theoretical_loss": 3.845438510898102,
      "tokens_seen": 591790080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144633901705115,
      "loss": 1.2002,
      "theoretical_loss": 3.8453494357662747,
      "tokens_seen": 591921152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041444332998996994,
      "loss": 1.4201,
      "theoretical_loss": 3.8452603858780288,
      "tokens_seen": 592052224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144232698094283,
      "loss": 1.5521,
      "theoretical_loss": 3.8451713612206246,
      "tokens_seen": 592183296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041440320962888666,
      "loss": 1.5349,
      "theoretical_loss": 3.84508236178133,
      "tokens_seen": 592314368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414383149448345,
      "loss": 1.5031,
      "theoretical_loss": 3.844993387547425,
      "tokens_seen": 592445440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041436308926780344,
      "loss": 1.5198,
      "theoretical_loss": 3.8449044385061955,
      "tokens_seen": 592576512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143430290872618,
      "loss": 1.3538,
      "theoretical_loss": 3.8448155146449396,
      "tokens_seen": 592707584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041432296890672016,
      "loss": 1.5557,
      "theoretical_loss": 3.8447266159509628,
      "tokens_seen": 592838656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143029087261785,
      "loss": 1.4679,
      "theoretical_loss": 3.84463774241158,
      "tokens_seen": 592969728
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.3541666567325592,
      "objective/train/avg_token_score": 0.016156869009137154,
      "objective/train/docs_used": 345976,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.190863847732544,
      "objective/train/original_loss": 3.7492716312408447,
      "objective/train/theoretical_loss": 3.844548894014116,
      "objective/train/tokens_used": 593100800,
      "objective/train/unlikelihood_loss": 0.21690547466278076,
      "theoretical_loss": 3.844548894014116,
      "tokens_seen": 593100800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142828485456369,
      "loss": 1.5316,
      "theoretical_loss": 3.844548894014116,
      "tokens_seen": 593100800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142627883650953,
      "loss": 1.5358,
      "theoretical_loss": 3.844460070745905,
      "tokens_seen": 593231872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041424272818455367,
      "loss": 1.4871,
      "theoretical_loss": 3.8443712725942882,
      "tokens_seen": 593362944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041422266800401203,
      "loss": 1.5228,
      "theoretical_loss": 3.844282499546619,
      "tokens_seen": 593494016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142026078234704,
      "loss": 1.7323,
      "theoretical_loss": 3.8441937515902573,
      "tokens_seen": 593625088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141825476429288,
      "loss": 1.5594,
      "theoretical_loss": 3.8441050287125744,
      "tokens_seen": 593756160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041416248746238717,
      "loss": 1.5317,
      "theoretical_loss": 3.844016330900949,
      "tokens_seen": 593887232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041414242728184553,
      "loss": 1.5874,
      "theoretical_loss": 3.843927658142771,
      "tokens_seen": 594018304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141223671013039,
      "loss": 1.3527,
      "theoretical_loss": 3.843839010425436,
      "tokens_seen": 594149376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141023069207623,
      "loss": 1.5098,
      "theoretical_loss": 3.8437503877363524,
      "tokens_seen": 594280448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140822467402207,
      "loss": 1.3849,
      "theoretical_loss": 3.843661790062935,
      "tokens_seen": 594411520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041406218655967904,
      "loss": 1.3191,
      "theoretical_loss": 3.84357321739261,
      "tokens_seen": 594542592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140421263791374,
      "loss": 1.4866,
      "theoretical_loss": 3.8434846697128098,
      "tokens_seen": 594673664
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.38501954078674316,
      "objective/train/avg_token_score": 0.03767390176653862,
      "objective/train/docs_used": 346603,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2809624671936035,
      "objective/train/original_loss": 3.974229335784912,
      "objective/train/theoretical_loss": 3.843440405240433,
      "objective/train/tokens_used": 594739200,
      "objective/train/unlikelihood_loss": 0.15178601443767548,
      "theoretical_loss": 3.843440405240433,
      "tokens_seen": 594739200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041402206619859576,
      "loss": 1.5946,
      "theoretical_loss": 3.843396147010979,
      "tokens_seen": 594804736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140020060180542,
      "loss": 1.4606,
      "theoretical_loss": 3.8433076492745695,
      "tokens_seen": 594935808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041398194583751254,
      "loss": 1.6322,
      "theoretical_loss": 3.8432191764910426,
      "tokens_seen": 595066880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139618856569709,
      "loss": 1.5872,
      "theoretical_loss": 3.8431307286478678,
      "tokens_seen": 595197952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041394182547642926,
      "loss": 1.4489,
      "theoretical_loss": 3.843042305732526,
      "tokens_seen": 595329024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139217652958877,
      "loss": 1.3802,
      "theoretical_loss": 3.8429539077325043,
      "tokens_seen": 595460096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041390170511534604,
      "loss": 1.435,
      "theoretical_loss": 3.842865534635301,
      "tokens_seen": 595591168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138816449348044,
      "loss": 1.4674,
      "theoretical_loss": 3.842777186428423,
      "tokens_seen": 595722240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041386158475426277,
      "loss": 1.384,
      "theoretical_loss": 3.8426888630993847,
      "tokens_seen": 595853312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041384152457372113,
      "loss": 1.5661,
      "theoretical_loss": 3.8426005646357106,
      "tokens_seen": 595984384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041382146439317954,
      "loss": 1.3068,
      "theoretical_loss": 3.842512291024935,
      "tokens_seen": 596115456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138014042126379,
      "loss": 1.5246,
      "theoretical_loss": 3.842424042254601,
      "tokens_seen": 596246528
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.43139052391052246,
      "objective/train/avg_token_score": 0.019480930641293526,
      "objective/train/docs_used": 348035,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4563043117523193,
      "objective/train/original_loss": 3.788301706314087,
      "objective/train/theoretical_loss": 3.8423358183122582,
      "objective/train/tokens_used": 596377600,
      "objective/train/unlikelihood_loss": 0.16736803948879242,
      "theoretical_loss": 3.8423358183122582,
      "tokens_seen": 596377600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041378134403209627,
      "loss": 1.4645,
      "theoretical_loss": 3.8423358183122582,
      "tokens_seen": 596377600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041376128385155463,
      "loss": 1.3936,
      "theoretical_loss": 3.8422476191854686,
      "tokens_seen": 596508672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041374122367101305,
      "loss": 1.3855,
      "theoretical_loss": 3.8421594448618013,
      "tokens_seen": 596639744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137211634904714,
      "loss": 1.4532,
      "theoretical_loss": 3.842071295328835,
      "tokens_seen": 596770816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041370110330992977,
      "loss": 1.3864,
      "theoretical_loss": 3.8419831705741556,
      "tokens_seen": 596901888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041368104312938813,
      "loss": 1.4822,
      "theoretical_loss": 3.841895070585361,
      "tokens_seen": 597032960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041366098294884655,
      "loss": 1.6476,
      "theoretical_loss": 3.8418069953500558,
      "tokens_seen": 597164032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041364092276830497,
      "loss": 1.649,
      "theoretical_loss": 3.841718944855854,
      "tokens_seen": 597295104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041362086258776333,
      "loss": 1.372,
      "theoretical_loss": 3.8416309190903783,
      "tokens_seen": 597426176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136008024072217,
      "loss": 1.3341,
      "theoretical_loss": 3.8415429180412617,
      "tokens_seen": 597557248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041358074222668005,
      "loss": 1.3834,
      "theoretical_loss": 3.8414549416961443,
      "tokens_seen": 597688320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041356068204613847,
      "loss": 1.5413,
      "theoretical_loss": 3.841366990042676,
      "tokens_seen": 597819392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041354062186559683,
      "loss": 1.4008,
      "theoretical_loss": 3.8412790630685154,
      "tokens_seen": 597950464
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.28977271914482117,
      "objective/train/avg_token_score": 0.01197260245680809,
      "objective/train/docs_used": 348629,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9216458201408386,
      "objective/train/original_loss": 3.9676473140716553,
      "objective/train/theoretical_loss": 3.841235108832321,
      "objective/train/tokens_used": 598016000,
      "objective/train/unlikelihood_loss": 0.14898046851158142,
      "theoretical_loss": 3.841235108832321,
      "tokens_seen": 598016000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135205616850552,
      "loss": 1.4445,
      "theoretical_loss": 3.84119116076133,
      "tokens_seen": 598081536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041350050150451356,
      "loss": 1.3845,
      "theoretical_loss": 3.8411032831087963,
      "tokens_seen": 598212608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134804413239719,
      "loss": 1.4899,
      "theoretical_loss": 3.841015430098599,
      "tokens_seen": 598343680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041346038114343034,
      "loss": 1.4013,
      "theoretical_loss": 3.8409276017184326,
      "tokens_seen": 598474752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134403209628887,
      "loss": 1.4651,
      "theoretical_loss": 3.840839797956,
      "tokens_seen": 598605824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041342026078234706,
      "loss": 1.292,
      "theoretical_loss": 3.8407520187990127,
      "tokens_seen": 598736896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134002006018054,
      "loss": 1.5541,
      "theoretical_loss": 3.8406642642351914,
      "tokens_seen": 598867968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041338014042126384,
      "loss": 1.4107,
      "theoretical_loss": 3.840576534252265,
      "tokens_seen": 598999040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133600802407222,
      "loss": 1.5056,
      "theoretical_loss": 3.840488828837973,
      "tokens_seen": 599130112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041334002006018056,
      "loss": 1.5114,
      "theoretical_loss": 3.840401147980061,
      "tokens_seen": 599261184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133199598796389,
      "loss": 1.4337,
      "theoretical_loss": 3.8403134916662847,
      "tokens_seen": 599392256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132998996990973,
      "loss": 1.3905,
      "theoretical_loss": 3.8402258598844097,
      "tokens_seen": 599523328
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.36742424964904785,
      "objective/train/avg_token_score": 0.029935631901025772,
      "objective/train/docs_used": 349996,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2140491008758545,
      "objective/train/original_loss": 4.204592704772949,
      "objective/train/theoretical_loss": 3.840138252622208,
      "objective/train/tokens_used": 599654400,
      "objective/train/unlikelihood_loss": 0.1415853351354599,
      "theoretical_loss": 3.840138252622208,
      "tokens_seen": 599654400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132798395185557,
      "loss": 1.3492,
      "theoretical_loss": 3.840138252622208,
      "tokens_seen": 599654400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041325977933801407,
      "loss": 1.5691,
      "theoretical_loss": 3.8400506698674635,
      "tokens_seen": 599785472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041323971915747243,
      "loss": 1.4959,
      "theoretical_loss": 3.839963111607965,
      "tokens_seen": 599916544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132196589769308,
      "loss": 1.4411,
      "theoretical_loss": 3.8398755778315135,
      "tokens_seen": 600047616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131995987963892,
      "loss": 1.4184,
      "theoretical_loss": 3.8397880685259165,
      "tokens_seen": 600178688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041317953861584757,
      "loss": 1.4997,
      "theoretical_loss": 3.839700583678991,
      "tokens_seen": 600309760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041315947843530593,
      "loss": 1.4387,
      "theoretical_loss": 3.8396131232785637,
      "tokens_seen": 600440832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131394182547643,
      "loss": 1.5275,
      "theoretical_loss": 3.8395256873124683,
      "tokens_seen": 600571904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131193580742227,
      "loss": 1.4305,
      "theoretical_loss": 3.8394382757685483,
      "tokens_seen": 600702976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130992978936811,
      "loss": 1.57,
      "theoretical_loss": 3.8393508886346552,
      "tokens_seen": 600834048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041307923771313944,
      "loss": 1.3022,
      "theoretical_loss": 3.83926352589865,
      "tokens_seen": 600965120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130591775325978,
      "loss": 1.4727,
      "theoretical_loss": 3.839176187548402,
      "tokens_seen": 601096192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041303911735205616,
      "loss": 1.4899,
      "theoretical_loss": 3.8390888735717885,
      "tokens_seen": 601227264
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.4315738081932068,
      "objective/train/avg_token_score": 0.012260204181075096,
      "objective/train/docs_used": 350686,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4201757907867432,
      "objective/train/original_loss": 3.8342316150665283,
      "objective/train/theoretical_loss": 3.8390452257198096,
      "objective/train/tokens_used": 601292800,
      "objective/train/unlikelihood_loss": 0.11259658634662628,
      "theoretical_loss": 3.8390452257198096,
      "tokens_seen": 601292800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130190571715146,
      "loss": 1.375,
      "theoretical_loss": 3.839001583956697,
      "tokens_seen": 601358336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041299899699097294,
      "loss": 1.4189,
      "theoretical_loss": 3.8389143186910224,
      "tokens_seen": 601489408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129789368104313,
      "loss": 1.5252,
      "theoretical_loss": 3.838827077762668,
      "tokens_seen": 601620480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041295887662988966,
      "loss": 1.5391,
      "theoretical_loss": 3.8387398611595476,
      "tokens_seen": 601751552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129388164493481,
      "loss": 1.6135,
      "theoretical_loss": 3.838652668869581,
      "tokens_seen": 601882624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041291875626880644,
      "loss": 1.4085,
      "theoretical_loss": 3.838565500880699,
      "tokens_seen": 602013696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128986960882648,
      "loss": 1.4676,
      "theoretical_loss": 3.83847835718084,
      "tokens_seen": 602144768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041287863590772317,
      "loss": 1.3773,
      "theoretical_loss": 3.8383912377579508,
      "tokens_seen": 602275840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041285857572718153,
      "loss": 1.6076,
      "theoretical_loss": 3.8383041425999864,
      "tokens_seen": 602406912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041283851554663995,
      "loss": 1.3761,
      "theoretical_loss": 3.8382170716949124,
      "tokens_seen": 602537984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128184553660983,
      "loss": 1.3649,
      "theoretical_loss": 3.8381300250307007,
      "tokens_seen": 602669056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041279839518555667,
      "loss": 1.5752,
      "theoretical_loss": 3.8380430025953327,
      "tokens_seen": 602800128
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.4177052676677704,
      "objective/train/avg_token_score": 0.011318889446556568,
      "objective/train/docs_used": 352047,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3438571691513062,
      "objective/train/original_loss": 3.8496592044830322,
      "objective/train/theoretical_loss": 3.837956004376799,
      "objective/train/tokens_used": 602931200,
      "objective/train/unlikelihood_loss": 0.19700142741203308,
      "theoretical_loss": 3.837956004376799,
      "tokens_seen": 602931200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041277833500501503,
      "loss": 1.6608,
      "theoretical_loss": 3.837956004376799,
      "tokens_seen": 602931200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041275827482447345,
      "loss": 1.441,
      "theoretical_loss": 3.8378690303630973,
      "tokens_seen": 603062272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127382146439318,
      "loss": 1.4288,
      "theoretical_loss": 3.837782080542236,
      "tokens_seen": 603193344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041271815446339017,
      "loss": 1.4916,
      "theoretical_loss": 3.837695154902229,
      "tokens_seen": 603324416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041269809428284854,
      "loss": 1.3033,
      "theoretical_loss": 3.8376082534311022,
      "tokens_seen": 603455488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126780341023069,
      "loss": 1.3031,
      "theoretical_loss": 3.8375213761168867,
      "tokens_seen": 603586560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126579739217653,
      "loss": 1.5044,
      "theoretical_loss": 3.8374345229476248,
      "tokens_seen": 603717632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126379137412237,
      "loss": 1.5035,
      "theoretical_loss": 3.837347693911366,
      "tokens_seen": 603848704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041261785356068204,
      "loss": 1.4344,
      "theoretical_loss": 3.837260888996169,
      "tokens_seen": 603979776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125977933801404,
      "loss": 1.5352,
      "theoretical_loss": 3.8371741081900996,
      "tokens_seen": 604110848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125777331995988,
      "loss": 1.5946,
      "theoretical_loss": 3.837087351481234,
      "tokens_seen": 604241920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125576730190572,
      "loss": 1.4719,
      "theoretical_loss": 3.8370006188576555,
      "tokens_seen": 604372992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041253761283851554,
      "loss": 1.2678,
      "theoretical_loss": 3.836913910307456,
      "tokens_seen": 604504064
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.3923509418964386,
      "objective/train/avg_token_score": 0.010112078860402107,
      "objective/train/docs_used": 352465,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4317771196365356,
      "objective/train/original_loss": 4.348731994628906,
      "objective/train/theoretical_loss": 3.836870565056154,
      "objective/train/tokens_used": 604569600,
      "objective/train/unlikelihood_loss": 0.11279493570327759,
      "theoretical_loss": 3.836870565056154,
      "tokens_seen": 604569600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125175526579739,
      "loss": 1.5122,
      "theoretical_loss": 3.8368272258187366,
      "tokens_seen": 604635136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041249749247743227,
      "loss": 1.2894,
      "theoretical_loss": 3.836740565379606,
      "tokens_seen": 604766208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124774322968907,
      "loss": 1.4996,
      "theoretical_loss": 3.836653928978182,
      "tokens_seen": 604897280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041245737211634904,
      "loss": 1.3454,
      "theoretical_loss": 3.8365673166025913,
      "tokens_seen": 605028352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124373119358074,
      "loss": 1.3027,
      "theoretical_loss": 3.8364807282409674,
      "tokens_seen": 605159424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041241725175526577,
      "loss": 1.5057,
      "theoretical_loss": 3.8363941638814536,
      "tokens_seen": 605290496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123971915747242,
      "loss": 1.3416,
      "theoretical_loss": 3.8363076235122016,
      "tokens_seen": 605421568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041237713139418255,
      "loss": 1.4237,
      "theoretical_loss": 3.83622110712137,
      "tokens_seen": 605552640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123570712136409,
      "loss": 1.3262,
      "theoretical_loss": 3.8361346146971282,
      "tokens_seen": 605683712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041233701103309927,
      "loss": 1.4146,
      "theoretical_loss": 3.836048146227652,
      "tokens_seen": 605814784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041231695085255763,
      "loss": 1.5451,
      "theoretical_loss": 3.8359617017011267,
      "tokens_seen": 605945856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041229689067201605,
      "loss": 1.3069,
      "theoretical_loss": 3.8358752811057455,
      "tokens_seen": 606076928
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.47061339020729065,
      "objective/train/avg_token_score": 0.003165400819852948,
      "objective/train/docs_used": 353634,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.681175708770752,
      "objective/train/original_loss": 3.5376367568969727,
      "objective/train/theoretical_loss": 3.8357888844297094,
      "objective/train/tokens_used": 606208000,
      "objective/train/unlikelihood_loss": 0.1784207969903946,
      "theoretical_loss": 3.8357888844297094,
      "tokens_seen": 606208000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122768304914744,
      "loss": 1.3184,
      "theoretical_loss": 3.8357888844297094,
      "tokens_seen": 606208000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122567703109328,
      "loss": 1.5651,
      "theoretical_loss": 3.8357025116612293,
      "tokens_seen": 606339072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041223671013039114,
      "loss": 1.5169,
      "theoretical_loss": 3.8356161627885235,
      "tokens_seen": 606470144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041221664994984955,
      "loss": 1.2234,
      "theoretical_loss": 3.835529837799818,
      "tokens_seen": 606601216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121965897693079,
      "loss": 1.5053,
      "theoretical_loss": 3.835443536683349,
      "tokens_seen": 606732288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121765295887663,
      "loss": 1.4598,
      "theoretical_loss": 3.8353572594273597,
      "tokens_seen": 606863360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121564694082247,
      "loss": 1.5847,
      "theoretical_loss": 3.835271006020101,
      "tokens_seen": 606994432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121364092276831,
      "loss": 1.3336,
      "theoretical_loss": 3.8351847764498332,
      "tokens_seen": 607125504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121163490471415,
      "loss": 1.3501,
      "theoretical_loss": 3.835098570704826,
      "tokens_seen": 607256576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041209628886659984,
      "loss": 1.4099,
      "theoretical_loss": 3.8350123887733543,
      "tokens_seen": 607387648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004120762286860582,
      "loss": 1.3004,
      "theoretical_loss": 3.8349262306437044,
      "tokens_seen": 607518720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041205616850551656,
      "loss": 1.464,
      "theoretical_loss": 3.8348400963041684,
      "tokens_seen": 607649792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412036108324975,
      "loss": 1.4531,
      "theoretical_loss": 3.834753985743049,
      "tokens_seen": 607780864
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.2646627426147461,
      "objective/train/avg_token_score": 0.01583893597126007,
      "objective/train/docs_used": 354263,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9999536871910095,
      "objective/train/original_loss": 4.469597816467285,
      "objective/train/theoretical_loss": 3.834710939375742,
      "objective/train/tokens_used": 607846400,
      "objective/train/unlikelihood_loss": 0.14511200785636902,
      "theoretical_loss": 3.834710939375742,
      "tokens_seen": 607846400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041201604814443334,
      "loss": 1.4492,
      "theoretical_loss": 3.8346678989486556,
      "tokens_seen": 607911936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119959879638917,
      "loss": 1.3742,
      "theoretical_loss": 3.8345818359093062,
      "tokens_seen": 608043008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041197592778335006,
      "loss": 1.4287,
      "theoretical_loss": 3.834495796613327,
      "tokens_seen": 608174080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119558676028085,
      "loss": 1.4213,
      "theoretical_loss": 3.834409781049053,
      "tokens_seen": 608305152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041193580742226684,
      "loss": 1.5101,
      "theoretical_loss": 3.8343237892048263,
      "tokens_seen": 608436224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119157472417252,
      "loss": 1.4667,
      "theoretical_loss": 3.8342378210689994,
      "tokens_seen": 608567296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041189568706118357,
      "loss": 1.4083,
      "theoretical_loss": 3.83415187662993,
      "tokens_seen": 608698368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041187562688064193,
      "loss": 1.463,
      "theoretical_loss": 3.834065955875987,
      "tokens_seen": 608829440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041185556670010035,
      "loss": 1.2794,
      "theoretical_loss": 3.833980058795545,
      "tokens_seen": 608960512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118355065195587,
      "loss": 1.5537,
      "theoretical_loss": 3.8338941853769892,
      "tokens_seen": 609091584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041181544633901707,
      "loss": 1.5483,
      "theoretical_loss": 3.8338083356087105,
      "tokens_seen": 609222656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041179538615847543,
      "loss": 1.5858,
      "theoretical_loss": 3.8337225094791103,
      "tokens_seen": 609353728
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.4015762507915497,
      "objective/train/avg_token_score": 0.020319471135735512,
      "objective/train/docs_used": 355721,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4089401960372925,
      "objective/train/original_loss": 4.095301628112793,
      "objective/train/theoretical_loss": 3.8336367069765958,
      "objective/train/tokens_used": 609484800,
      "objective/train/unlikelihood_loss": 0.2028389275074005,
      "theoretical_loss": 3.8336367069765958,
      "tokens_seen": 609484800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041177532597793385,
      "loss": 1.273,
      "theoretical_loss": 3.8336367069765958,
      "tokens_seen": 609484800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117552657973922,
      "loss": 1.45,
      "theoretical_loss": 3.8335509280895854,
      "tokens_seen": 609615872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117352056168506,
      "loss": 1.4615,
      "theoretical_loss": 3.833465172806503,
      "tokens_seen": 609746944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041171514543630894,
      "loss": 1.4819,
      "theoretical_loss": 3.833379441115781,
      "tokens_seen": 609878016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116950852557673,
      "loss": 1.3545,
      "theoretical_loss": 3.8332937330058616,
      "tokens_seen": 610009088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116750250752257,
      "loss": 1.4763,
      "theoretical_loss": 3.8332080484651945,
      "tokens_seen": 610140160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116549648946841,
      "loss": 1.5326,
      "theoretical_loss": 3.833122387482236,
      "tokens_seen": 610271232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041163490471414244,
      "loss": 1.3957,
      "theoretical_loss": 3.833036750045452,
      "tokens_seen": 610402304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004116148445336008,
      "loss": 1.4854,
      "theoretical_loss": 3.8329511361433166,
      "tokens_seen": 610533376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115947843530592,
      "loss": 1.4062,
      "theoretical_loss": 3.832865545764311,
      "tokens_seen": 610664448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115747241725176,
      "loss": 1.4865,
      "theoretical_loss": 3.832779978896926,
      "tokens_seen": 610795520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041155466399197594,
      "loss": 1.4416,
      "theoretical_loss": 3.8326944355296595,
      "tokens_seen": 610926592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115346038114343,
      "loss": 1.4191,
      "theoretical_loss": 3.832608915651017,
      "tokens_seen": 611057664
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.4571114480495453,
      "objective/train/avg_token_score": 0.008640832267701626,
      "objective/train/docs_used": 356348,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6731990575790405,
      "objective/train/original_loss": 3.7301838397979736,
      "objective/train/theoretical_loss": 3.83256616451634,
      "objective/train/tokens_used": 611123200,
      "objective/train/unlikelihood_loss": 0.17205381393432617,
      "theoretical_loss": 3.83256616451634,
      "tokens_seen": 611123200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041151454363089267,
      "loss": 1.4817,
      "theoretical_loss": 3.8325234192495126,
      "tokens_seen": 611188736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114944834503511,
      "loss": 1.4969,
      "theoretical_loss": 3.8324379463136697,
      "tokens_seen": 611319808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041147442326980944,
      "loss": 1.4105,
      "theoretical_loss": 3.832352496832018,
      "tokens_seen": 611450880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114543630892678,
      "loss": 1.5492,
      "theoretical_loss": 3.8322670707930957,
      "tokens_seen": 611581952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041143430290872617,
      "loss": 1.5112,
      "theoretical_loss": 3.83218166818545,
      "tokens_seen": 611713024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114142427281846,
      "loss": 1.2652,
      "theoretical_loss": 3.8320962889976347,
      "tokens_seen": 611844096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041139418254764295,
      "loss": 1.58,
      "theoretical_loss": 3.8320109332182124,
      "tokens_seen": 611975168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113741223671013,
      "loss": 1.4196,
      "theoretical_loss": 3.831925600835754,
      "tokens_seen": 612106240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041135406218655967,
      "loss": 1.4769,
      "theoretical_loss": 3.8318402918388386,
      "tokens_seen": 612237312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041133400200601803,
      "loss": 1.506,
      "theoretical_loss": 3.831755006216052,
      "tokens_seen": 612368384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041131394182547645,
      "loss": 1.5132,
      "theoretical_loss": 3.831669743955989,
      "tokens_seen": 612499456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112938816449348,
      "loss": 1.3302,
      "theoretical_loss": 3.8315845050472532,
      "tokens_seen": 612630528
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.4703079164028168,
      "objective/train/avg_token_score": 0.01573745161294937,
      "objective/train/docs_used": 356825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7282345294952393,
      "objective/train/original_loss": 3.63925838470459,
      "objective/train/theoretical_loss": 3.8314992894784536,
      "objective/train/tokens_used": 612761600,
      "objective/train/unlikelihood_loss": 0.17425817251205444,
      "theoretical_loss": 3.8314992894784536,
      "tokens_seen": 612761600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112738214643932,
      "loss": 1.4771,
      "theoretical_loss": 3.8314992894784536,
      "tokens_seen": 612761600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041125376128385154,
      "loss": 1.4603,
      "theoretical_loss": 3.8314140972382105,
      "tokens_seen": 612892672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041123370110330995,
      "loss": 1.2944,
      "theoretical_loss": 3.8313289283151497,
      "tokens_seen": 613023744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112136409227683,
      "loss": 1.3469,
      "theoretical_loss": 3.831243782697906,
      "tokens_seen": 613154816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111935807422267,
      "loss": 1.2951,
      "theoretical_loss": 3.831158660375122,
      "tokens_seen": 613285888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041117352056168504,
      "loss": 1.3934,
      "theoretical_loss": 3.8310735613354483,
      "tokens_seen": 613416960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111534603811434,
      "loss": 1.2533,
      "theoretical_loss": 3.830988485567543,
      "tokens_seen": 613548032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111334002006018,
      "loss": 1.4875,
      "theoretical_loss": 3.830903433060073,
      "tokens_seen": 613679104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111133400200602,
      "loss": 1.4649,
      "theoretical_loss": 3.830818403801713,
      "tokens_seen": 613810176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041109327983951854,
      "loss": 1.2906,
      "theoretical_loss": 3.830733397781144,
      "tokens_seen": 613941248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110732196589769,
      "loss": 1.2456,
      "theoretical_loss": 3.830648414987057,
      "tokens_seen": 614072320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110531594784353,
      "loss": 1.2733,
      "theoretical_loss": 3.830563455408151,
      "tokens_seen": 614203392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110330992978937,
      "loss": 1.3083,
      "theoretical_loss": 3.8304785190331305,
      "tokens_seen": 614334464
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.37115103006362915,
      "objective/train/avg_token_score": 0.012984885834157467,
      "objective/train/docs_used": 357491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3638612031936646,
      "objective/train/original_loss": 3.861525535583496,
      "objective/train/theoretical_loss": 3.830436059543551,
      "objective/train/tokens_used": 614400000,
      "objective/train/unlikelihood_loss": 0.16821658611297607,
      "theoretical_loss": 3.830436059543551,
      "tokens_seen": 614400000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041101303911735205,
      "loss": 1.7196,
      "theoretical_loss": 3.830393605850711,
      "tokens_seen": 614465536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109929789368104,
      "loss": 1.1773,
      "theoretical_loss": 3.8303087158496134,
      "tokens_seen": 614596608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041097291875626877,
      "loss": 1.173,
      "theoretical_loss": 3.8302238490185676,
      "tokens_seen": 614727680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109528585757272,
      "loss": 1.4127,
      "theoretical_loss": 3.830139005346312,
      "tokens_seen": 614858752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041093279839518555,
      "loss": 1.2904,
      "theoretical_loss": 3.8300541848215905,
      "tokens_seen": 614989824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109127382146439,
      "loss": 1.4453,
      "theoretical_loss": 3.8299693874331573,
      "tokens_seen": 615120896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108926780341023,
      "loss": 1.3533,
      "theoretical_loss": 3.8298846131697744,
      "tokens_seen": 615251968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108726178535607,
      "loss": 1.201,
      "theoretical_loss": 3.8297998620202094,
      "tokens_seen": 615383040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041085255767301905,
      "loss": 1.4987,
      "theoretical_loss": 3.8297151339732407,
      "tokens_seen": 615514112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108324974924774,
      "loss": 1.3022,
      "theoretical_loss": 3.8296304290176524,
      "tokens_seen": 615645184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108124373119358,
      "loss": 1.5003,
      "theoretical_loss": 3.8295457471422365,
      "tokens_seen": 615776256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107923771313942,
      "loss": 1.3209,
      "theoretical_loss": 3.8294610883357945,
      "tokens_seen": 615907328
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.2765762507915497,
      "objective/train/avg_token_score": 0.022694801911711693,
      "objective/train/docs_used": 358741,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9784660339355469,
      "objective/train/original_loss": 3.9417734146118164,
      "objective/train/theoretical_loss": 3.829376452587134,
      "objective/train/tokens_used": 616038400,
      "objective/train/unlikelihood_loss": 0.2532103657722473,
      "theoretical_loss": 3.829376452587134,
      "tokens_seen": 616038400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041077231695085256,
      "loss": 1.4609,
      "theoretical_loss": 3.829376452587134,
      "tokens_seen": 616038400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107522567703109,
      "loss": 1.4066,
      "theoretical_loss": 3.8292918398850713,
      "tokens_seen": 616169472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107321965897693,
      "loss": 1.3902,
      "theoretical_loss": 3.82920725021843,
      "tokens_seen": 616300544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041071213640922764,
      "loss": 1.2997,
      "theoretical_loss": 3.829122683576042,
      "tokens_seen": 616431616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041069207622868606,
      "loss": 1.4858,
      "theoretical_loss": 3.8290381399467464,
      "tokens_seen": 616562688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106720160481444,
      "loss": 1.3344,
      "theoretical_loss": 3.8289536193193907,
      "tokens_seen": 616693760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041065195586760284,
      "loss": 1.4025,
      "theoretical_loss": 3.8288691216828297,
      "tokens_seen": 616824832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106318956870612,
      "loss": 1.4296,
      "theoretical_loss": 3.8287846470259264,
      "tokens_seen": 616955904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106118355065196,
      "loss": 1.552,
      "theoretical_loss": 3.828700195337551,
      "tokens_seen": 617086976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410591775325978,
      "loss": 1.3554,
      "theoretical_loss": 3.8286157666065828,
      "tokens_seen": 617218048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041057171514543634,
      "loss": 1.3402,
      "theoretical_loss": 3.8285313608219056,
      "tokens_seen": 617349120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105516549648947,
      "loss": 1.3052,
      "theoretical_loss": 3.8284469779724155,
      "tokens_seen": 617480192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041053159478435307,
      "loss": 1.3052,
      "theoretical_loss": 3.828362618047013,
      "tokens_seen": 617611264
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.46590909361839294,
      "objective/train/avg_token_score": 0.011999816633760929,
      "objective/train/docs_used": 359404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5094083547592163,
      "objective/train/original_loss": 3.980893135070801,
      "objective/train/theoretical_loss": 3.828320446677378,
      "objective/train/tokens_used": 617676800,
      "objective/train/unlikelihood_loss": 0.12393984198570251,
      "theoretical_loss": 3.828320446677378,
      "tokens_seen": 617676800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105115346038115,
      "loss": 1.4529,
      "theoretical_loss": 3.8282782810346063,
      "tokens_seen": 617742336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041049147442326985,
      "loss": 1.3784,
      "theoretical_loss": 3.8281939669241143,
      "tokens_seen": 617873408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104714142427282,
      "loss": 1.3825,
      "theoretical_loss": 3.82810967570446,
      "tokens_seen": 618004480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041045135406218657,
      "loss": 1.3919,
      "theoretical_loss": 3.828025407364577,
      "tokens_seen": 618135552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410431293881645,
      "loss": 1.4041,
      "theoretical_loss": 3.8279411618934045,
      "tokens_seen": 618266624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041041123370110335,
      "loss": 1.4088,
      "theoretical_loss": 3.82785693927989,
      "tokens_seen": 618397696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103911735205617,
      "loss": 1.4916,
      "theoretical_loss": 3.82777273951299,
      "tokens_seen": 618528768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041037111334002007,
      "loss": 1.5682,
      "theoretical_loss": 3.8276885625816663,
      "tokens_seen": 618659840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041035105315947844,
      "loss": 1.538,
      "theoretical_loss": 3.8276044084748904,
      "tokens_seen": 618790912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041033099297893685,
      "loss": 1.3839,
      "theoretical_loss": 3.827520277181641,
      "tokens_seen": 618921984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103109327983952,
      "loss": 1.5539,
      "theoretical_loss": 3.8274361686909035,
      "tokens_seen": 619053056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102908726178536,
      "loss": 1.4889,
      "theoretical_loss": 3.827352082991672,
      "tokens_seen": 619184128
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.3732893466949463,
      "objective/train/avg_token_score": 0.019104547798633575,
      "objective/train/docs_used": 360794,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2005456686019897,
      "objective/train/original_loss": 3.98445725440979,
      "objective/train/theoretical_loss": 3.827268020072948,
      "objective/train/tokens_used": 619315200,
      "objective/train/unlikelihood_loss": 0.14163470268249512,
      "theoretical_loss": 3.827268020072948,
      "tokens_seen": 619315200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041027081243731194,
      "loss": 1.3425,
      "theoretical_loss": 3.827268020072948,
      "tokens_seen": 619315200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041025075225677035,
      "loss": 1.299,
      "theoretical_loss": 3.8271839799237397,
      "tokens_seen": 619446272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102306920762287,
      "loss": 1.6187,
      "theoretical_loss": 3.8270999625330644,
      "tokens_seen": 619577344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102106318956871,
      "loss": 1.3738,
      "theoretical_loss": 3.8270159678899462,
      "tokens_seen": 619708416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041019057171514544,
      "loss": 1.3923,
      "theoretical_loss": 3.8269319959834167,
      "tokens_seen": 619839488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101705115346038,
      "loss": 1.4009,
      "theoretical_loss": 3.8268480468025157,
      "tokens_seen": 619970560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101504513540622,
      "loss": 1.5065,
      "theoretical_loss": 3.82676412033629,
      "tokens_seen": 620101632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101303911735206,
      "loss": 1.3505,
      "theoretical_loss": 3.826680216573795,
      "tokens_seen": 620232704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041011033099297894,
      "loss": 1.3851,
      "theoretical_loss": 3.8265963355040915,
      "tokens_seen": 620363776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100902708124373,
      "loss": 1.7003,
      "theoretical_loss": 3.8265124771162506,
      "tokens_seen": 620494848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100702106318957,
      "loss": 1.255,
      "theoretical_loss": 3.8264286413993496,
      "tokens_seen": 620625920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100501504513541,
      "loss": 1.5705,
      "theoretical_loss": 3.8263448283424726,
      "tokens_seen": 620756992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041003009027081245,
      "loss": 1.3712,
      "theoretical_loss": 3.826261037934713,
      "tokens_seen": 620888064
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.39937683939933777,
      "objective/train/avg_token_score": 0.00654844893142581,
      "objective/train/docs_used": 361341,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4888705015182495,
      "objective/train/original_loss": 4.043774604797363,
      "objective/train/theoretical_loss": 3.8262191512208448,
      "objective/train/tokens_used": 620953600,
      "objective/train/unlikelihood_loss": 0.12301285564899445,
      "theoretical_loss": 3.8262191512208448,
      "tokens_seen": 620953600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100100300902708,
      "loss": 1.3344,
      "theoretical_loss": 3.8261772701651697,
      "tokens_seen": 621019136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040998996990972917,
      "loss": 1.5005,
      "theoretical_loss": 3.8260935250229515,
      "tokens_seen": 621150208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099699097291876,
      "loss": 1.3902,
      "theoretical_loss": 3.8260098024971727,
      "tokens_seen": 621281280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040994984954864595,
      "loss": 1.4164,
      "theoretical_loss": 3.8259261025769566,
      "tokens_seen": 621412352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099297893681043,
      "loss": 1.4026,
      "theoretical_loss": 3.825842425251434,
      "tokens_seen": 621543424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099097291875627,
      "loss": 1.4462,
      "theoretical_loss": 3.8257587705097404,
      "tokens_seen": 621674496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098896690070211,
      "loss": 1.5376,
      "theoretical_loss": 3.8256751383410235,
      "tokens_seen": 621805568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040986960882647945,
      "loss": 1.3613,
      "theoretical_loss": 3.825591528734434,
      "tokens_seen": 621936640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098495486459378,
      "loss": 1.3915,
      "theoretical_loss": 3.8255079416791338,
      "tokens_seen": 622067712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098294884653962,
      "loss": 1.3391,
      "theoretical_loss": 3.8254243771642895,
      "tokens_seen": 622198784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098094282848546,
      "loss": 1.5791,
      "theoretical_loss": 3.8253408351790767,
      "tokens_seen": 622329856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040978936810431296,
      "loss": 1.4436,
      "theoretical_loss": 3.825257315712678,
      "tokens_seen": 622460928
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.2656402587890625,
      "objective/train/avg_token_score": 0.020830322057008743,
      "objective/train/docs_used": 362042,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8504209518432617,
      "objective/train/original_loss": 4.423110008239746,
      "objective/train/theoretical_loss": 3.8251738187542843,
      "objective/train/tokens_used": 622592000,
      "objective/train/unlikelihood_loss": 0.12244540452957153,
      "theoretical_loss": 3.8251738187542843,
      "tokens_seen": 622592000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097693079237713,
      "loss": 1.261,
      "theoretical_loss": 3.8251738187542843,
      "tokens_seen": 622592000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097492477432297,
      "loss": 1.2915,
      "theoretical_loss": 3.8250903442930917,
      "tokens_seen": 622723072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040972918756268804,
      "loss": 1.4155,
      "theoretical_loss": 3.8250068923183065,
      "tokens_seen": 622854144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040970912738214646,
      "loss": 1.3704,
      "theoretical_loss": 3.8249234628191404,
      "tokens_seen": 622985216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096890672016048,
      "loss": 1.4364,
      "theoretical_loss": 3.8248400557848146,
      "tokens_seen": 623116288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096690070210632,
      "loss": 1.2654,
      "theoretical_loss": 3.824756671204555,
      "tokens_seen": 623247360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040964894684052155,
      "loss": 1.1298,
      "theoretical_loss": 3.824673309067598,
      "tokens_seen": 623378432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040962888665997996,
      "loss": 1.4316,
      "theoretical_loss": 3.824589969363184,
      "tokens_seen": 623509504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096088264794383,
      "loss": 1.4605,
      "theoretical_loss": 3.824506652080564,
      "tokens_seen": 623640576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095887662988967,
      "loss": 1.4785,
      "theoretical_loss": 3.8244233572089943,
      "tokens_seen": 623771648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040956870611835505,
      "loss": 1.4477,
      "theoretical_loss": 3.8243400847377407,
      "tokens_seen": 623902720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095486459378134,
      "loss": 1.2426,
      "theoretical_loss": 3.8242568346560737,
      "tokens_seen": 624033792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040952858575727183,
      "loss": 1.4633,
      "theoretical_loss": 3.8241736069532726,
      "tokens_seen": 624164864
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.38581377267837524,
      "objective/train/avg_token_score": 0.025465743616223335,
      "objective/train/docs_used": 363016,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5282154083251953,
      "objective/train/original_loss": 4.084991931915283,
      "objective/train/theoretical_loss": 3.8241320014905993,
      "objective/train/tokens_used": 624230400,
      "objective/train/unlikelihood_loss": 0.14625848829746246,
      "theoretical_loss": 3.8241320014905993,
      "tokens_seen": 624230400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095085255767302,
      "loss": 1.3288,
      "theoretical_loss": 3.824090401618625,
      "tokens_seen": 624295936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040948846539618855,
      "loss": 1.6178,
      "theoretical_loss": 3.8240072186414245,
      "tokens_seen": 624427008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094684052156469,
      "loss": 1.4476,
      "theoretical_loss": 3.8239240580109723,
      "tokens_seen": 624558080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040944834503510533,
      "loss": 1.3995,
      "theoretical_loss": 3.8238409197165772,
      "tokens_seen": 624689152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094282848545637,
      "loss": 1.362,
      "theoretical_loss": 3.8237578037475553,
      "tokens_seen": 624820224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040940822467402206,
      "loss": 1.424,
      "theoretical_loss": 3.823674710093231,
      "tokens_seen": 624951296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093881644934804,
      "loss": 1.4847,
      "theoretical_loss": 3.8235916387429336,
      "tokens_seen": 625082368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093681043129388,
      "loss": 1.348,
      "theoretical_loss": 3.8235085896860017,
      "tokens_seen": 625213440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093480441323972,
      "loss": 1.322,
      "theoretical_loss": 3.8234255629117815,
      "tokens_seen": 625344512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040932798395185556,
      "loss": 1.3699,
      "theoretical_loss": 3.8233425584096254,
      "tokens_seen": 625475584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093079237713139,
      "loss": 1.385,
      "theoretical_loss": 3.8232595761688932,
      "tokens_seen": 625606656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092878635907723,
      "loss": 1.4204,
      "theoretical_loss": 3.8231766161789524,
      "tokens_seen": 625737728
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.4359726309776306,
      "objective/train/avg_token_score": 0.016611715778708458,
      "objective/train/docs_used": 363672,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.319413423538208,
      "objective/train/original_loss": 3.72617506980896,
      "objective/train/theoretical_loss": 3.8230936784291787,
      "objective/train/tokens_used": 625868800,
      "objective/train/unlikelihood_loss": 0.15357278287410736,
      "theoretical_loss": 3.8230936784291787,
      "tokens_seen": 625868800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092678034102307,
      "loss": 1.3302,
      "theoretical_loss": 3.8230936784291787,
      "tokens_seen": 625868800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040924774322968906,
      "loss": 1.4842,
      "theoretical_loss": 3.823010762908953,
      "tokens_seen": 625999872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092276830491474,
      "loss": 1.3199,
      "theoretical_loss": 3.822927869607664,
      "tokens_seen": 626130944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092076228686058,
      "loss": 1.2685,
      "theoretical_loss": 3.822844998514711,
      "tokens_seen": 626262016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040918756268806415,
      "loss": 1.5454,
      "theoretical_loss": 3.8227621496194946,
      "tokens_seen": 626393088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040916750250752257,
      "loss": 1.406,
      "theoretical_loss": 3.822679322911429,
      "tokens_seen": 626524160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409147442326981,
      "loss": 1.1559,
      "theoretical_loss": 3.82259651837993,
      "tokens_seen": 626655232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040912738214643934,
      "loss": 1.356,
      "theoretical_loss": 3.8225137360144252,
      "tokens_seen": 626786304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091073219658977,
      "loss": 1.4281,
      "theoretical_loss": 3.8224309758043464,
      "tokens_seen": 626917376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090872617853561,
      "loss": 1.2494,
      "theoretical_loss": 3.822348237739134,
      "tokens_seen": 627048448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090672016048145,
      "loss": 1.3822,
      "theoretical_loss": 3.822265521808236,
      "tokens_seen": 627179520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040904714142427285,
      "loss": 1.4783,
      "theoretical_loss": 3.8221828280011065,
      "tokens_seen": 627310592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090270812437312,
      "loss": 1.3068,
      "theoretical_loss": 3.822100156307208,
      "tokens_seen": 627441664
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.359665185213089,
      "objective/train/avg_token_score": 0.019122768193483353,
      "objective/train/docs_used": 364826,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3461437225341797,
      "objective/train/original_loss": 3.9998061656951904,
      "objective/train/theoretical_loss": 3.8220588287494284,
      "objective/train/tokens_used": 627507200,
      "objective/train/unlikelihood_loss": 0.16339793801307678,
      "theoretical_loss": 3.8220588287494284,
      "tokens_seen": 627507200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040900702106318957,
      "loss": 1.3026,
      "theoretical_loss": 3.8220175067160085,
      "tokens_seen": 627572736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408986960882648,
      "loss": 1.4798,
      "theoretical_loss": 3.821934879216985,
      "tokens_seen": 627703808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040896690070210635,
      "loss": 1.3787,
      "theoretical_loss": 3.821852273799621,
      "tokens_seen": 627834880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089468405215647,
      "loss": 1.329,
      "theoretical_loss": 3.821769690453407,
      "tokens_seen": 627965952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089267803410231,
      "loss": 1.3329,
      "theoretical_loss": 3.8216871291678416,
      "tokens_seen": 628097024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089067201604815,
      "loss": 1.4236,
      "theoretical_loss": 3.821604589932429,
      "tokens_seen": 628228096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040888665997993985,
      "loss": 1.3524,
      "theoretical_loss": 3.8215220727366823,
      "tokens_seen": 628359168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088665997993982,
      "loss": 1.4028,
      "theoretical_loss": 3.82143957757012,
      "tokens_seen": 628490240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088465396188566,
      "loss": 1.4714,
      "theoretical_loss": 3.8213571044222694,
      "tokens_seen": 628621312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408826479438315,
      "loss": 1.3119,
      "theoretical_loss": 3.8212746532826642,
      "tokens_seen": 628752384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040880641925777336,
      "loss": 1.3945,
      "theoretical_loss": 3.8211922241408454,
      "tokens_seen": 628883456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087863590772317,
      "loss": 1.2536,
      "theoretical_loss": 3.8211098169863607,
      "tokens_seen": 629014528
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.4605938494205475,
      "objective/train/avg_token_score": 0.018184412270784378,
      "objective/train/docs_used": 365375,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5830137729644775,
      "objective/train/original_loss": 3.831573724746704,
      "objective/train/theoretical_loss": 3.8210274318087656,
      "objective/train/tokens_used": 629145600,
      "objective/train/unlikelihood_loss": 0.12006247788667679,
      "theoretical_loss": 3.8210274318087656,
      "tokens_seen": 629145600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087662988966901,
      "loss": 1.4189,
      "theoretical_loss": 3.8210274318087656,
      "tokens_seen": 629145600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040874623871614844,
      "loss": 1.4326,
      "theoretical_loss": 3.8209450685976227,
      "tokens_seen": 629276672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040872617853560686,
      "loss": 1.5258,
      "theoretical_loss": 3.820862727342501,
      "tokens_seen": 629407744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087061183550652,
      "loss": 1.2155,
      "theoretical_loss": 3.8207804080329772,
      "tokens_seen": 629538816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086860581745236,
      "loss": 1.3935,
      "theoretical_loss": 3.8206981106586357,
      "tokens_seen": 629669888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040866599799398195,
      "loss": 1.4249,
      "theoretical_loss": 3.8206158352090664,
      "tokens_seen": 629800960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040864593781344036,
      "loss": 1.4532,
      "theoretical_loss": 3.8205335816738684,
      "tokens_seen": 629932032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086258776328987,
      "loss": 1.4963,
      "theoretical_loss": 3.820451350042646,
      "tokens_seen": 630063104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086058174523571,
      "loss": 1.3688,
      "theoretical_loss": 3.8203691403050106,
      "tokens_seen": 630194176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040858575727181545,
      "loss": 1.2832,
      "theoretical_loss": 3.8202869524505827,
      "tokens_seen": 630325248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085656970912738,
      "loss": 1.5181,
      "theoretical_loss": 3.8202047864689885,
      "tokens_seen": 630456320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040854563691073223,
      "loss": 1.4243,
      "theoretical_loss": 3.820122642349861,
      "tokens_seen": 630587392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085255767301906,
      "loss": 1.4029,
      "theoretical_loss": 3.8200405200828405,
      "tokens_seen": 630718464
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.2529325485229492,
      "objective/train/avg_token_score": 0.023288333788514137,
      "objective/train/docs_used": 366612,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9088571071624756,
      "objective/train/original_loss": 4.29040002822876,
      "objective/train/theoretical_loss": 3.8199994671406348,
      "objective/train/tokens_used": 630784000,
      "objective/train/unlikelihood_loss": 0.16850118339061737,
      "theoretical_loss": 3.8199994671406348,
      "tokens_seen": 630784000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040850551654964895,
      "loss": 1.3071,
      "theoretical_loss": 3.819958419657574,
      "tokens_seen": 630849536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084854563691073,
      "loss": 1.4315,
      "theoretical_loss": 3.819876341063718,
      "tokens_seen": 630980608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040846539618856573,
      "loss": 1.2791,
      "theoretical_loss": 3.8197942842909325,
      "tokens_seen": 631111680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084453360080241,
      "loss": 1.3536,
      "theoretical_loss": 3.8197122493288864,
      "tokens_seen": 631242752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040842527582748246,
      "loss": 1.3218,
      "theoretical_loss": 3.8196302361672556,
      "tokens_seen": 631373824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084052156469408,
      "loss": 1.4322,
      "theoretical_loss": 3.8195482447957225,
      "tokens_seen": 631504896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083851554663992,
      "loss": 1.5428,
      "theoretical_loss": 3.8194662752039776,
      "tokens_seen": 631635968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083650952858576,
      "loss": 1.5233,
      "theoretical_loss": 3.8193843273817167,
      "tokens_seen": 631767040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040834503510531596,
      "loss": 1.3592,
      "theoretical_loss": 3.8193024013186445,
      "tokens_seen": 631898112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083249749247743,
      "loss": 1.3766,
      "theoretical_loss": 3.819220497004471,
      "tokens_seen": 632029184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083049147442327,
      "loss": 1.4008,
      "theoretical_loss": 3.8191386144289137,
      "tokens_seen": 632160256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082848545636911,
      "loss": 1.4924,
      "theoretical_loss": 3.819056753581699,
      "tokens_seen": 632291328
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.4759897291660309,
      "objective/train/avg_token_score": 0.017946870997548103,
      "objective/train/docs_used": 367323,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5524520874023438,
      "objective/train/original_loss": 3.585705041885376,
      "objective/train/theoretical_loss": 3.818974914452557,
      "objective/train/tokens_used": 632422400,
      "objective/train/unlikelihood_loss": 0.13466675579547882,
      "theoretical_loss": 3.818974914452557,
      "tokens_seen": 632422400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040826479438314946,
      "loss": 1.3453,
      "theoretical_loss": 3.818974914452557,
      "tokens_seen": 632422400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082447342026078,
      "loss": 1.3149,
      "theoretical_loss": 3.8188930970312267,
      "tokens_seen": 632553472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082246740220662,
      "loss": 1.3998,
      "theoretical_loss": 3.8188113013074547,
      "tokens_seen": 632684544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040820461384152455,
      "loss": 1.2335,
      "theoretical_loss": 3.8187295272709934,
      "tokens_seen": 632815616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040818455366098297,
      "loss": 1.1875,
      "theoretical_loss": 3.818647774911601,
      "tokens_seen": 632946688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040816449348044133,
      "loss": 1.2951,
      "theoretical_loss": 3.818566044219046,
      "tokens_seen": 633077760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081444332998997,
      "loss": 1.4905,
      "theoretical_loss": 3.8184843351831006,
      "tokens_seen": 633208832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040812437311935805,
      "loss": 1.2673,
      "theoretical_loss": 3.8184026477935458,
      "tokens_seen": 633339904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040810431293881647,
      "loss": 1.338,
      "theoretical_loss": 3.818320982040169,
      "tokens_seen": 633470976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040808425275827483,
      "loss": 1.708,
      "theoretical_loss": 3.8182393379127655,
      "tokens_seen": 633602048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080641925777332,
      "loss": 1.5884,
      "theoretical_loss": 3.8181577154011346,
      "tokens_seen": 633733120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040804413239719156,
      "loss": 1.4255,
      "theoretical_loss": 3.818076114495086,
      "tokens_seen": 633864192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080240722166499,
      "loss": 1.5113,
      "theoretical_loss": 3.817994535184434,
      "tokens_seen": 633995264
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.5092253088951111,
      "objective/train/avg_token_score": 0.017480170354247093,
      "objective/train/docs_used": 367883,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8412432670593262,
      "objective/train/original_loss": 3.7791481018066406,
      "objective/train/theoretical_loss": 3.8179537536242005,
      "objective/train/tokens_used": 634060800,
      "objective/train/unlikelihood_loss": 0.15870891511440277,
      "theoretical_loss": 3.8179537536242005,
      "tokens_seen": 634060800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040800401203610834,
      "loss": 1.3694,
      "theoretical_loss": 3.8179129774590006,
      "tokens_seen": 634126336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079839518555667,
      "loss": 1.5419,
      "theoretical_loss": 3.8178314413086154,
      "tokens_seen": 634257408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040796389167502506,
      "loss": 1.4045,
      "theoretical_loss": 3.817749926723114,
      "tokens_seen": 634388480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079438314944834,
      "loss": 1.3526,
      "theoretical_loss": 3.8176684336923383,
      "tokens_seen": 634519552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040792377131394184,
      "loss": 1.4804,
      "theoretical_loss": 3.8175869622061382,
      "tokens_seen": 634650624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079037111334002,
      "loss": 1.5129,
      "theoretical_loss": 3.817505512254371,
      "tokens_seen": 634781696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040788365095285856,
      "loss": 1.4568,
      "theoretical_loss": 3.817424083826899,
      "tokens_seen": 634912768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078635907723169,
      "loss": 1.3891,
      "theoretical_loss": 3.8173426769135927,
      "tokens_seen": 635043840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078435305917753,
      "loss": 1.3367,
      "theoretical_loss": 3.8172612915043294,
      "tokens_seen": 635174912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078234704112337,
      "loss": 1.4437,
      "theoretical_loss": 3.8171799275889926,
      "tokens_seen": 635305984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040780341023069207,
      "loss": 1.3874,
      "theoretical_loss": 3.8170985851574732,
      "tokens_seen": 635437056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040778335005015043,
      "loss": 1.3605,
      "theoretical_loss": 3.8170172641996682,
      "tokens_seen": 635568128
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.39595553278923035,
      "objective/train/avg_token_score": 0.020078817382454872,
      "objective/train/docs_used": 369288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3277698755264282,
      "objective/train/original_loss": 3.681797742843628,
      "objective/train/theoretical_loss": 3.8169359647054835,
      "objective/train/tokens_used": 635699200,
      "objective/train/unlikelihood_loss": 0.16572554409503937,
      "theoretical_loss": 3.8169359647054835,
      "tokens_seen": 635699200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077632898696088,
      "loss": 1.3921,
      "theoretical_loss": 3.8169359647054835,
      "tokens_seen": 635699200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077432296890672,
      "loss": 1.2567,
      "theoretical_loss": 3.8168546866648283,
      "tokens_seen": 635830272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040772316950852557,
      "loss": 1.3112,
      "theoretical_loss": 3.816773430067622,
      "tokens_seen": 635961344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040770310932798393,
      "loss": 1.5763,
      "theoretical_loss": 3.816692194903789,
      "tokens_seen": 636092416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076830491474423,
      "loss": 1.3896,
      "theoretical_loss": 3.816610981163261,
      "tokens_seen": 636223488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076629889669007,
      "loss": 1.4929,
      "theoretical_loss": 3.8165297888359766,
      "tokens_seen": 636354560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076429287863591,
      "loss": 1.5615,
      "theoretical_loss": 3.816448617911881,
      "tokens_seen": 636485632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076228686058175,
      "loss": 1.5079,
      "theoretical_loss": 3.816367468380926,
      "tokens_seen": 636616704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040760280842527585,
      "loss": 1.2841,
      "theoretical_loss": 3.8162863402330713,
      "tokens_seen": 636747776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075827482447342,
      "loss": 1.4782,
      "theoretical_loss": 3.816205233458281,
      "tokens_seen": 636878848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040756268806419263,
      "loss": 1.1803,
      "theoretical_loss": 3.816124148046529,
      "tokens_seen": 637009920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407542627883651,
      "loss": 1.5352,
      "theoretical_loss": 3.8160430839877937,
      "tokens_seen": 637140992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040752256770310935,
      "loss": 1.3779,
      "theoretical_loss": 3.815962041272061,
      "tokens_seen": 637272064
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.5428274869918823,
      "objective/train/avg_token_score": 0.006741791497915983,
      "objective/train/docs_used": 369983,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.891964077949524,
      "objective/train/original_loss": 3.4886183738708496,
      "objective/train/theoretical_loss": 3.815921527914693,
      "objective/train/tokens_used": 637337600,
      "objective/train/unlikelihood_loss": 0.1536000370979309,
      "theoretical_loss": 3.815921527914693,
      "tokens_seen": 637337600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075025075225677,
      "loss": 1.4306,
      "theoretical_loss": 3.815881019889324,
      "tokens_seen": 637403136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040748244734202613,
      "loss": 1.4211,
      "theoretical_loss": 3.8158000198295814,
      "tokens_seen": 637534208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074623871614845,
      "loss": 1.4194,
      "theoretical_loss": 3.81571904108284,
      "tokens_seen": 637665280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040744232698094286,
      "loss": 1.28,
      "theoretical_loss": 3.8156380836391124,
      "tokens_seen": 637796352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074222668004012,
      "loss": 1.4144,
      "theoretical_loss": 3.815557147488418,
      "tokens_seen": 637927424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074022066198596,
      "loss": 1.4748,
      "theoretical_loss": 3.815476232620784,
      "tokens_seen": 638058496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407382146439318,
      "loss": 1.4531,
      "theoretical_loss": 3.8153953390262427,
      "tokens_seen": 638189568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040736208625877636,
      "loss": 1.3289,
      "theoretical_loss": 3.8153144666948338,
      "tokens_seen": 638320640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073420260782347,
      "loss": 1.4632,
      "theoretical_loss": 3.815233615616605,
      "tokens_seen": 638451712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073219658976931,
      "loss": 1.5634,
      "theoretical_loss": 3.8151527857816077,
      "tokens_seen": 638582784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073019057171515,
      "loss": 1.483,
      "theoretical_loss": 3.815071977179903,
      "tokens_seen": 638713856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040728184553660986,
      "loss": 1.4585,
      "theoretical_loss": 3.814991189801557,
      "tokens_seen": 638844928
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.2507942318916321,
      "objective/train/avg_token_score": 0.031877949833869934,
      "objective/train/docs_used": 370882,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.902974545955658,
      "objective/train/original_loss": 4.440757751464844,
      "objective/train/theoretical_loss": 3.8149104236366433,
      "objective/train/tokens_used": 638976000,
      "objective/train/unlikelihood_loss": 0.10956820845603943,
      "theoretical_loss": 3.8149104236366433,
      "tokens_seen": 638976000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072617853560682,
      "loss": 1.54,
      "theoretical_loss": 3.8149104236366433,
      "tokens_seen": 638976000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072417251755266,
      "loss": 1.4755,
      "theoretical_loss": 3.814829678675242,
      "tokens_seen": 639107072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040722166499498495,
      "loss": 1.2833,
      "theoretical_loss": 3.814748954907439,
      "tokens_seen": 639238144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040720160481444337,
      "loss": 1.3224,
      "theoretical_loss": 3.8146682523233286,
      "tokens_seen": 639369216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040718154463390173,
      "loss": 1.5405,
      "theoretical_loss": 3.8145875709130093,
      "tokens_seen": 639500288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071614844533601,
      "loss": 1.4171,
      "theoretical_loss": 3.814506910666589,
      "tokens_seen": 639631360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040714142427281845,
      "loss": 1.5556,
      "theoretical_loss": 3.81442627157418,
      "tokens_seen": 639762432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040712136409227687,
      "loss": 1.5124,
      "theoretical_loss": 3.8143456536259035,
      "tokens_seen": 639893504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040710130391173523,
      "loss": 1.4711,
      "theoretical_loss": 3.8142650568118848,
      "tokens_seen": 640024576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070812437311936,
      "loss": 1.4587,
      "theoretical_loss": 3.814184481122257,
      "tokens_seen": 640155648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040706118355065196,
      "loss": 1.4772,
      "theoretical_loss": 3.814103926547161,
      "tokens_seen": 640286720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070411233701103,
      "loss": 1.5276,
      "theoretical_loss": 3.8140233930767424,
      "tokens_seen": 640417792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040702106318956874,
      "loss": 1.4636,
      "theoretical_loss": 3.8139428807011537,
      "tokens_seen": 640548864
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.2876954972743988,
      "objective/train/avg_token_score": 0.005228789057582617,
      "objective/train/docs_used": 371543,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9728174805641174,
      "objective/train/original_loss": 3.6427764892578125,
      "objective/train/theoretical_loss": 3.8139026324208456,
      "objective/train/tokens_used": 640614400,
      "objective/train/unlikelihood_loss": 0.19972942769527435,
      "theoretical_loss": 3.8139026324208456,
      "tokens_seen": 640614400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070010030090271,
      "loss": 1.3805,
      "theoretical_loss": 3.8138623894105557,
      "tokens_seen": 640679936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040698094282848546,
      "loss": 1.5623,
      "theoretical_loss": 3.813781919195114,
      "tokens_seen": 640811008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069608826479438,
      "loss": 1.5725,
      "theoretical_loss": 3.813701470045001,
      "tokens_seen": 640942080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040694082246740224,
      "loss": 1.4707,
      "theoretical_loss": 3.8136210419503977,
      "tokens_seen": 641073152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069207622868606,
      "loss": 1.4157,
      "theoretical_loss": 3.8135406349014884,
      "tokens_seen": 641204224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040690070210631896,
      "loss": 1.4061,
      "theoretical_loss": 3.8134602488884664,
      "tokens_seen": 641335296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068806419257773,
      "loss": 1.3694,
      "theoretical_loss": 3.813379883901531,
      "tokens_seen": 641466368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068605817452357,
      "loss": 1.4924,
      "theoretical_loss": 3.8132995399308873,
      "tokens_seen": 641597440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068405215646941,
      "loss": 1.5049,
      "theoretical_loss": 3.8132192169667483,
      "tokens_seen": 641728512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040682046138415247,
      "loss": 1.2999,
      "theoretical_loss": 3.813138914999332,
      "tokens_seen": 641859584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040680040120361083,
      "loss": 1.3115,
      "theoretical_loss": 3.813058634018865,
      "tokens_seen": 641990656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067803410230692,
      "loss": 1.3317,
      "theoretical_loss": 3.812978374015578,
      "tokens_seen": 642121728
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.43664467334747314,
      "objective/train/avg_token_score": 0.01415205281227827,
      "objective/train/docs_used": 373013,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5176647901535034,
      "objective/train/original_loss": 3.9611873626708984,
      "objective/train/theoretical_loss": 3.8128981349797098,
      "objective/train/tokens_used": 642252800,
      "objective/train/unlikelihood_loss": 0.1576123982667923,
      "theoretical_loss": 3.8128981349797098,
      "tokens_seen": 642252800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067602808425276,
      "loss": 1.6215,
      "theoretical_loss": 3.8128981349797098,
      "tokens_seen": 642252800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040674022066198597,
      "loss": 1.4195,
      "theoretical_loss": 3.8128179169015057,
      "tokens_seen": 642383872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040672016048144433,
      "loss": 1.4247,
      "theoretical_loss": 3.8127377197712176,
      "tokens_seen": 642514944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067001003009027,
      "loss": 1.4622,
      "theoretical_loss": 3.812657543579103,
      "tokens_seen": 642646016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066800401203611,
      "loss": 1.3565,
      "theoretical_loss": 3.812577388315426,
      "tokens_seen": 642777088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040665997993981947,
      "loss": 1.5205,
      "theoretical_loss": 3.812497253970459,
      "tokens_seen": 642908160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040663991975927783,
      "loss": 1.5612,
      "theoretical_loss": 3.812417140534478,
      "tokens_seen": 643039232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066198595787362,
      "loss": 1.3668,
      "theoretical_loss": 3.8123370479977687,
      "tokens_seen": 643170304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040659979939819456,
      "loss": 1.4831,
      "theoretical_loss": 3.8122569763506204,
      "tokens_seen": 643301376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000406579739217653,
      "loss": 1.5363,
      "theoretical_loss": 3.812176925583331,
      "tokens_seen": 643432448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040655967903711134,
      "loss": 1.3226,
      "theoretical_loss": 3.8120968956862034,
      "tokens_seen": 643563520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004065396188565697,
      "loss": 1.4465,
      "theoretical_loss": 3.8120168866495483,
      "tokens_seen": 643694592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040651955867602806,
      "loss": 1.5168,
      "theoretical_loss": 3.8119368984636814,
      "tokens_seen": 643825664
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.45222383737564087,
      "objective/train/avg_token_score": 0.015421704389154911,
      "objective/train/docs_used": 373533,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5024516582489014,
      "objective/train/original_loss": 3.635314464569092,
      "objective/train/theoretical_loss": 3.8118969121867696,
      "objective/train/tokens_used": 643891200,
      "objective/train/unlikelihood_loss": 0.1828690469264984,
      "theoretical_loss": 3.8118969121867696,
      "tokens_seen": 643891200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064994984954865,
      "loss": 1.2832,
      "theoretical_loss": 3.8118569311189265,
      "tokens_seen": 643956736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040647943831494484,
      "loss": 1.453,
      "theoretical_loss": 3.811776984605612,
      "tokens_seen": 644087808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064593781344032,
      "loss": 1.4285,
      "theoretical_loss": 3.811697058914075,
      "tokens_seen": 644218880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040643931795386157,
      "loss": 1.4884,
      "theoretical_loss": 3.811617154034657,
      "tokens_seen": 644349952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040641925777331993,
      "loss": 1.4966,
      "theoretical_loss": 3.8115372699577073,
      "tokens_seen": 644481024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040639919759277834,
      "loss": 1.4494,
      "theoretical_loss": 3.8114574066735805,
      "tokens_seen": 644612096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063791374122367,
      "loss": 1.4307,
      "theoretical_loss": 3.8113775641726386,
      "tokens_seen": 644743168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040635907723169507,
      "loss": 1.3986,
      "theoretical_loss": 3.81129774244525,
      "tokens_seen": 644874240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040633901705115343,
      "loss": 1.6454,
      "theoretical_loss": 3.8112179414817877,
      "tokens_seen": 645005312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040631895687061185,
      "loss": 1.6522,
      "theoretical_loss": 3.8111381612726345,
      "tokens_seen": 645136384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062988966900702,
      "loss": 1.6055,
      "theoretical_loss": 3.811058401808177,
      "tokens_seen": 645267456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040627883650952857,
      "loss": 1.479,
      "theoretical_loss": 3.8109786630788083,
      "tokens_seen": 645398528
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.4299853444099426,
      "objective/train/avg_token_score": 0.014909038320183754,
      "objective/train/docs_used": 374553,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5298352241516113,
      "objective/train/original_loss": 4.025496006011963,
      "objective/train/theoretical_loss": 3.8108989450749293,
      "objective/train/tokens_used": 645529600,
      "objective/train/unlikelihood_loss": 0.1589026004076004,
      "theoretical_loss": 3.8108989450749293,
      "tokens_seen": 645529600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040625877632898693,
      "loss": 1.5565,
      "theoretical_loss": 3.8108989450749293,
      "tokens_seen": 645529600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062387161484453,
      "loss": 1.3728,
      "theoretical_loss": 3.810819247786946,
      "tokens_seen": 645660672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062186559679037,
      "loss": 1.518,
      "theoretical_loss": 3.8107395712052714,
      "tokens_seen": 645791744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061985957873621,
      "loss": 1.3697,
      "theoretical_loss": 3.8106599153203256,
      "tokens_seen": 645922816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040617853560682044,
      "loss": 1.53,
      "theoretical_loss": 3.810580280122533,
      "tokens_seen": 646053888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061584754262788,
      "loss": 1.4548,
      "theoretical_loss": 3.8105006656023264,
      "tokens_seen": 646184960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040613841524573727,
      "loss": 1.4705,
      "theoretical_loss": 3.8104210717501434,
      "tokens_seen": 646316032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040611835506519563,
      "loss": 1.5978,
      "theoretical_loss": 3.8103414985564292,
      "tokens_seen": 646447104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000406098294884654,
      "loss": 1.3603,
      "theoretical_loss": 3.810261946011636,
      "tokens_seen": 646578176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040607823470411236,
      "loss": 1.4798,
      "theoretical_loss": 3.810182414106219,
      "tokens_seen": 646709248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060581745235707,
      "loss": 1.4606,
      "theoretical_loss": 3.8101029028306437,
      "tokens_seen": 646840320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040603811434302914,
      "loss": 1.5778,
      "theoretical_loss": 3.8100234121753798,
      "tokens_seen": 646971392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060180541624875,
      "loss": 1.5656,
      "theoretical_loss": 3.8099439421309036,
      "tokens_seen": 647102464
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.45576736330986023,
      "objective/train/avg_token_score": 0.028985600918531418,
      "objective/train/docs_used": 375146,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.656118392944336,
      "objective/train/original_loss": 3.9800736904144287,
      "objective/train/theoretical_loss": 3.8099042148347366,
      "objective/train/tokens_used": 647168000,
      "objective/train/unlikelihood_loss": 0.11932669579982758,
      "theoretical_loss": 3.8099042148347366,
      "tokens_seen": 647168000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040599799398194586,
      "loss": 1.4929,
      "theoretical_loss": 3.809864492687698,
      "tokens_seen": 647233536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004059779338014042,
      "loss": 1.5617,
      "theoretical_loss": 3.8097850638362516,
      "tokens_seen": 647364608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040595787362086264,
      "loss": 1.2915,
      "theoretical_loss": 3.8097056555670608,
      "tokens_seen": 647495680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405937813440321,
      "loss": 1.4292,
      "theoretical_loss": 3.8096262678706267,
      "tokens_seen": 647626752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040591775325977936,
      "loss": 1.5357,
      "theoretical_loss": 3.809546900737457,
      "tokens_seen": 647757824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058976930792377,
      "loss": 1.3051,
      "theoretical_loss": 3.809467554158067,
      "tokens_seen": 647888896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058776328986961,
      "loss": 1.4298,
      "theoretical_loss": 3.8093882281229767,
      "tokens_seen": 648019968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058575727181545,
      "loss": 1.3352,
      "theoretical_loss": 3.809308922622713,
      "tokens_seen": 648151040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040583751253761287,
      "loss": 1.512,
      "theoretical_loss": 3.8092296376478094,
      "tokens_seen": 648282112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040581745235707123,
      "loss": 1.4593,
      "theoretical_loss": 3.809150373188805,
      "tokens_seen": 648413184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057973921765296,
      "loss": 1.3488,
      "theoretical_loss": 3.8090711292362456,
      "tokens_seen": 648544256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405777331995988,
      "loss": 1.4565,
      "theoretical_loss": 3.808991905780683,
      "tokens_seen": 648675328
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.455278605222702,
      "objective/train/avg_token_score": 0.015043376013636589,
      "objective/train/docs_used": 375860,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5879684686660767,
      "objective/train/original_loss": 3.877953290939331,
      "objective/train/theoretical_loss": 3.8089127028126764,
      "objective/train/tokens_used": 648806400,
      "objective/train/unlikelihood_loss": 0.21014168858528137,
      "theoretical_loss": 3.8089127028126764,
      "tokens_seen": 648806400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040575727181544637,
      "loss": 1.4726,
      "theoretical_loss": 3.8089127028126764,
      "tokens_seen": 648806400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040573721163490473,
      "loss": 1.4837,
      "theoretical_loss": 3.8088335203227897,
      "tokens_seen": 648937472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057171514543631,
      "loss": 1.4837,
      "theoretical_loss": 3.8087543583015933,
      "tokens_seen": 649068544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040569709127382146,
      "loss": 1.338,
      "theoretical_loss": 3.808675216739665,
      "tokens_seen": 649199616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056770310932799,
      "loss": 1.4928,
      "theoretical_loss": 3.8085960956275873,
      "tokens_seen": 649330688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040565697091273824,
      "loss": 1.5526,
      "theoretical_loss": 3.8085169949559505,
      "tokens_seen": 649461760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056369107321966,
      "loss": 1.4267,
      "theoretical_loss": 3.8084379147153493,
      "tokens_seen": 649592832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040561685055165496,
      "loss": 1.3143,
      "theoretical_loss": 3.8083588548963867,
      "tokens_seen": 649723904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055967903711134,
      "loss": 1.4323,
      "theoretical_loss": 3.808279815489671,
      "tokens_seen": 649854976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040557673019057174,
      "loss": 1.2942,
      "theoretical_loss": 3.808200796485815,
      "tokens_seen": 649986048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055566700100301,
      "loss": 1.4689,
      "theoretical_loss": 3.8081217978754407,
      "tokens_seen": 650117120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040553660982948846,
      "loss": 1.6256,
      "theoretical_loss": 3.8080428196491747,
      "tokens_seen": 650248192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055165496489469,
      "loss": 1.3445,
      "theoretical_loss": 3.8079638617976492,
      "tokens_seen": 650379264
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.4335899353027344,
      "objective/train/avg_token_score": 0.011522009037435055,
      "objective/train/docs_used": 377374,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4454277753829956,
      "objective/train/original_loss": 3.9384045600891113,
      "objective/train/theoretical_loss": 3.8079243905094886,
      "objective/train/tokens_used": 650444800,
      "objective/train/unlikelihood_loss": 0.13931621611118317,
      "theoretical_loss": 3.8079243905094886,
      "tokens_seen": 650444800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040549648946840524,
      "loss": 1.4248,
      "theoretical_loss": 3.8078849243115043,
      "tokens_seen": 650510336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054764292878636,
      "loss": 1.3241,
      "theoretical_loss": 3.8078060071813846,
      "tokens_seen": 650641408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040545636910732197,
      "loss": 1.6134,
      "theoretical_loss": 3.807727110397942,
      "tokens_seen": 650772480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040543630892678033,
      "loss": 1.505,
      "theoretical_loss": 3.807648233951835,
      "tokens_seen": 650903552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040541624874623874,
      "loss": 1.5242,
      "theoretical_loss": 3.807569377833726,
      "tokens_seen": 651034624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053961885656971,
      "loss": 1.3798,
      "theoretical_loss": 3.8074905420342864,
      "tokens_seen": 651165696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040537612838515547,
      "loss": 1.5358,
      "theoretical_loss": 3.807411726544191,
      "tokens_seen": 651296768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040535606820461383,
      "loss": 1.6171,
      "theoretical_loss": 3.8073329313541233,
      "tokens_seen": 651427840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040533600802407225,
      "loss": 1.5088,
      "theoretical_loss": 3.807254156454772,
      "tokens_seen": 651558912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053159478435306,
      "loss": 1.504,
      "theoretical_loss": 3.8071754018368305,
      "tokens_seen": 651689984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040529588766298897,
      "loss": 1.4844,
      "theoretical_loss": 3.8070966674910007,
      "tokens_seen": 651821056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040527582748244733,
      "loss": 1.308,
      "theoretical_loss": 3.807017953407989,
      "tokens_seen": 651952128
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.30968964099884033,
      "objective/train/avg_token_score": 0.01105769444257021,
      "objective/train/docs_used": 377873,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0822099447250366,
      "objective/train/original_loss": 4.107913494110107,
      "objective/train/theoretical_loss": 3.8069392595785083,
      "objective/train/tokens_used": 652083200,
      "objective/train/unlikelihood_loss": 0.13217970728874207,
      "theoretical_loss": 3.8069392595785083,
      "tokens_seen": 652083200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052557673019057,
      "loss": 1.5052,
      "theoretical_loss": 3.8069392595785083,
      "tokens_seen": 652083200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052357071213641,
      "loss": 1.4459,
      "theoretical_loss": 3.8068605859932783,
      "tokens_seen": 652214272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052156469408225,
      "loss": 1.5091,
      "theoretical_loss": 3.806781932643024,
      "tokens_seen": 652345344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040519558676028084,
      "loss": 1.3071,
      "theoretical_loss": 3.8067032995184764,
      "tokens_seen": 652476416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051755265797392,
      "loss": 1.4541,
      "theoretical_loss": 3.806624686610374,
      "tokens_seen": 652607488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051554663991976,
      "loss": 1.3498,
      "theoretical_loss": 3.8065460939094597,
      "tokens_seen": 652738560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405135406218656,
      "loss": 1.5227,
      "theoretical_loss": 3.806467521406484,
      "tokens_seen": 652869632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040511534603811434,
      "loss": 1.5705,
      "theoretical_loss": 3.8063889690922013,
      "tokens_seen": 653000704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050952858575727,
      "loss": 1.6055,
      "theoretical_loss": 3.806310436957374,
      "tokens_seen": 653131776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040507522567703107,
      "loss": 1.3488,
      "theoretical_loss": 3.8062319249927707,
      "tokens_seen": 653262848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050551654964895,
      "loss": 1.4862,
      "theoretical_loss": 3.806153433189165,
      "tokens_seen": 653393920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040503510531594784,
      "loss": 1.5408,
      "theoretical_loss": 3.8060749615373375,
      "tokens_seen": 653524992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050150451354062,
      "loss": 1.4192,
      "theoretical_loss": 3.8059965100280735,
      "tokens_seen": 653656064
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.4617546498775482,
      "objective/train/avg_token_score": 0.007299356162548065,
      "objective/train/docs_used": 379015,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.598456621170044,
      "objective/train/original_loss": 3.781627655029297,
      "objective/train/theoretical_loss": 3.8059572918240248,
      "objective/train/tokens_used": 653721600,
      "objective/train/unlikelihood_loss": 0.2492816299200058,
      "theoretical_loss": 3.8059572918240248,
      "tokens_seen": 653721600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040499498495486457,
      "loss": 1.5622,
      "theoretical_loss": 3.8059180786521654,
      "tokens_seen": 653787136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404974924774323,
      "loss": 1.5489,
      "theoretical_loss": 3.805839667400412,
      "tokens_seen": 653918208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040495486459378135,
      "loss": 1.4142,
      "theoretical_loss": 3.8057612762636173,
      "tokens_seen": 654049280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049348044132397,
      "loss": 1.4089,
      "theoretical_loss": 3.8056829052325916,
      "tokens_seen": 654180352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040491474423269807,
      "loss": 1.4654,
      "theoretical_loss": 3.8056045542981516,
      "tokens_seen": 654311424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040489468405215643,
      "loss": 1.4081,
      "theoretical_loss": 3.8055262234511193,
      "tokens_seen": 654442496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040487462387161485,
      "loss": 1.4031,
      "theoretical_loss": 3.805447912682324,
      "tokens_seen": 654573568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048545636910732,
      "loss": 1.3817,
      "theoretical_loss": 3.8053696219825994,
      "tokens_seen": 654704640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048345035105316,
      "loss": 1.3633,
      "theoretical_loss": 3.8052913513427864,
      "tokens_seen": 654835712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040481444332998994,
      "loss": 1.4964,
      "theoretical_loss": 3.805213100753731,
      "tokens_seen": 654966784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040479438314944835,
      "loss": 1.3457,
      "theoretical_loss": 3.8051348702062864,
      "tokens_seen": 655097856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047743229689067,
      "loss": 1.3372,
      "theoretical_loss": 3.805056659691311,
      "tokens_seen": 655228928
    },
    {
      "debugging/length": 515.6,
      "debugging/num_segments": 5,
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.3613147735595703,
      "objective/train/avg_token_score": 0.006067417562007904,
      "objective/train/docs_used": 379813,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.339416742324829,
      "objective/train/original_loss": 3.8491370677948,
      "objective/train/theoretical_loss": 3.804978469199669,
      "objective/train/tokens_used": 655360000,
      "objective/train/unlikelihood_loss": 0.20670151710510254,
      "theoretical_loss": 3.804978469199669,
      "tokens_seen": 655360000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047542627883651,
      "loss": 1.4639,
      "theoretical_loss": 3.804978469199669,
      "tokens_seen": 655360000
    }
  ],
  "max_steps": 25177,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 3.3445380096e+17,
  "trial_name": null,
  "trial_params": null
}